SPILLOVERS AND COMPETITION AMONG FOREIGN AND LOCAL … smj2008.pdf · In this study, we examine...

Strategic Management JournalStrat. Mgmt. J., 29: 495–518 (2008)

Published online 22 February 2008 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/smj.674

Received 12 May 2006; Final revision received 22 November 2007

SPILLOVERS AND COMPETITION AMONG FOREIGNAND LOCAL FIRMS IN CHINA

SEA JIN CHANG1* and DEAN XU2**1 School of Business Administration, Korea University, Seoul, Korea2 School of Business, The University of Hong Kong, Pokfulam, Hong Kong

Combining the FDI spillover literature with a competitor analysis framework, we examine therelative size of spillover and competition effects in China between foreign entrants and local firms,among foreign entrants, and among local firms. Our results show that the increased presence offoreign entrants has generally benefited local firms nationally, but has negatively affected thesurvival rates of local firms in regional markets. Surprisingly, foreign entrants are crowded outnot only by their peers, but also by reformed local firms at both the national and regional levels.Copyright 2008 John Wiley & Sons, Ltd.

INTRODUCTION

Foreign direct investment (FDI) is critical to amultinational firm’s global strategy. Research instrategic management has long studied multina-tionals’ strategies and performance in overseasmarkets. It has examined such firms’ entry modedecisions (Kogut, 1983; Chang, 1995), their inter-actions with other multinationals (Henisz andDelios, 2001; Chang and Park, 2005; Miller andEden, 2006), and their ability to cope with thesocioeconomic environments of the host countriesthey enter (Rosenzweig and Singh, 1991; Ghoshaland Westney, 1993; Kostova and Zaheer, 1999).Yet it has typically studied these phenomena onlyfrom the perspective of multinational firms. Exceptfor studies on joint ventures, it has paid little atten-tion to local firms, which compete with multina-tionals in local markets (Meyer, 2004).

Keywords: spillover effects; competition effects; foreigndirect investments; competitor analysis; China*Correspondence to: Sea Jin Chang, School of Business Admin-istration, Korea University, Sungbukku Anamdong 5-1, Seoul,KOREA 136-701. E-mail: [email protected]∗∗ The authors are listed alphabetically; each contributed equally.

In contrast, economists have long consideredhow local firms can benefit from foreign entrants’technology spillovers and thereby improve theirown technological efficiency (Caves, 1974; Blom-strom, 1986; Hejazi and Safarian, 1999; Sinai andMeyer; 2004). Yet for the most part, economistshave also neglected host countries’ competitiveenvironments, in which local firms are integral,and have viewed local firms, especially those inemerging and developing economies, as passiverecipients of technology spillovers.

One possible reason that strategy and economicsresearch has marginalized local firms and charac-terized them as passive may be attributable to oneof FDI theory’s underlying assumptions, namelythat foreign entrants’ advantage over local firmsis sufficiently large to compensate for their liabil-ity of foreignness and to limit competition fromlocal firms (Hymer, 1960; Caves, 1971). Thisassumption may no longer be valid, however, assome local firms in developing countries are suc-cessfully challenging foreign entrants (Dawar andFrost, 1999; Zeng and Williamson, 2003). Forinstance, in the Chinese mobile handset sector, for-eign multinationals such as Motorola and Nokiahad more than 95 percent market share until 1999.

Copyright 2008 John Wiley & Sons, Ltd.

496 S. J. Chang and D. Xu

By 2003, local firms represented by Bird and TCLhad captured over 60 percent market share (Luo,2005). In the Chinese fresh milk industry, sixdomestic firms have shut out multinationals suchas Kraft and Danone by securing exclusive accessto high-quality dairy farms and tightly controllingdistribution channels. Anecdotal accounts furthersuggest that competition is intensifying not onlybetween foreign and local firms but also amongforeign entrants and among local firms themselves,a possibility overlooked by prior studies.

In this study, we examine these multiple formsof competition in an emerging market context.Our point of departure is the literature on FDIspillovers to local firms, which we believe hasgenerally neglected the possibility of such com-petition. In order to overcome this shortcoming,it is necessary to adopt a theoretical frameworkthat explains competition among firms. We there-fore draw insights from the literature on competitoranalysis and competitive dynamics to guide ourstudy. Specifically, Chen (1996) proposes two fac-tors—market commonality and resource similar-ity—as antecedents of interfirm rivalry. AlthoughChen uses this framework to analyze pair-wiserivalry, tension, and competitive interaction, weextend it to the study of competition among var-ious groups of firms, so that we can account forboth the positive and negative effects between for-eign and local firms, and among foreign and localfirms themselves.

The empirical analysis in this study is based onpanel data for more than 200,000 firms in Chinafrom 1998 to 2005. China is the largest emerg-ing market and the largest annual recipient of FDIin the world. It is an ideal setting to observespillovers and competition between and among for-eign entrants and local incumbents because it com-prises vast regional markets, and economic andsocial reforms have created competing firms withheterogeneous resources. Our findings provide arelatively complete picture of the mutual benefitsand costs of competition among foreign and localfirms in China.

SPILLOVERS AND COMPETITIONAMONG FOREIGN AND LOCAL FIRMS

FDI and the spillover effects

Spillovers denote the transfer of resources, suchas technological knowledge, between organizations

(e.g., from foreign entrants to local firms) with-out a contractual relationship (Meyer, 2004). Caves(1974) suggests that foreign entrants can increasethe speed of technology spillovers to host indus-tries by demonstrating their technological supe-riority, and by competing and transacting withdomestic firms. He tested for these effects witha sample of 22 Australian manufacturing sectors,and found a positive relationship between localfirms’ labor productivity (value-added per worker)and the share of foreign firms’ employment in anindustry. This finding provided early evidence forFDI’s positive impact on domestic firms. Severalsubsequent studies corroborated Caves’s results.For instance, Globerman (1979) found a positiverelationship in Canadian industries between laborproductivity and several measures for foreign pres-ence, including the proportion of value-added pro-duced in foreign-owned plants. Blomstrom andPersson (1983) and Blomstrom (1986) found a pos-itive relationship for Mexican industries betweenforeign participation in an industry and the relativechanges in labor productivity.

Since the 1990s, however, research on this topichas yielded a confusing jumble of results. On theone hand, some of this work has confirmed thefindings of earlier studies. For example, Hejaziand Safarian (1999) found that FDI stocks posi-tively influenced the total factor productivity of 22OECD (Organisation for Economic Co-operationand Development) countries. Driffield and Munday(2000) observed a positive relationship betweenindustry comparative advantage, reflected in theexport/import ratio, and foreign firms’ employmentshares in U.K. manufacturing industries. Liu et al.(2000) confirmed a positive relationship betweenlabor productivity and two measures of foreignpresence—the share of employment attributableto foreign firms and the share of foreign capitalstock in U.K. manufacturing industries. Buckley,Clegg, and Wang (2002) concluded that foreignfirms that invested in China could generate tech-nological and international market access spilloverbenefits for local firms. Sinai and Meyer (2004)found that the share of foreign employment in anindustry strongly predicted spillovers in Estonia.

In contrast, other studies have found no evi-dence, or contradictory evidence, of technologyspillovers from multinationals to local firms. Had-dad and Harrison (1993), perhaps the first studyto employ a firm-level dataset, found foreign firmsdid not affect the productivity of local firms in

Copyright 2008 John Wiley & Sons, Ltd. Strat. Mgmt. J., 29: 495–518 (2008)DOI: 10.1002/smj

Spillovers and Competition in China 497

the Moroccan manufacturing sector. Using a largesample of Venezuelan plants, Aitken and Harri-son (1999) found the level of foreign ownershipdepressed local firms’ productivity, except for jointventures. Konings (2001), again with firm-leveldata, found FDI had a negative impact on localfirms’ productivity in Bulgaria and Romania. Fur-thermore, this negative effect might also be presentin developed markets. De Backer and Sleuwaegen(2003), for example, found new foreign entrantsand the total number of foreign firms in the Bel-gian manufacturing sectors decreased the foundingrate and increased the exit rate of domestic firms.

Underlying these mixed results are major con-ceptual and technical problems. First, this researchhas studied only the impact of foreign entrants onlocal incumbents. It has neglected not only thecompetition that local firms might pose to for-eign entrants, but also the possibility that spilloversand competition can occur among local firms andamong foreign entrants. Foreign entrants mightpossess strong technological and brand advan-tages, but they also face liabilities of foreignnessbecause of cultural differences and underdevelopedsocial, economic, and political infrastructures inhost countries (Hymer, 1960; Khanna et al., 2005).In addition, local firms may possess local knowl-edge that is not immediately accessible to foreignfirms, or may be able to benefit from govern-ment policies and social networks that are locallyembedded (Makino and Delios, 1996; Lu and Xu,2006).

Second, most of this research has used the indus-try, not the firm, as the unit of analysis. Thus, it hasnot gauged whether individual firms’ productivityimproved. Further, because many inefficient localfirms might have exited an industry, the productiv-ity of surviving local firms, reflected in industry-level productivity, might be biased upward.

Most significantly, this research has lacked atheoretical framework to explain the negative FDIspillover effects (or competition effects) that havebeen observed in both developed and developingmarkets. The positive FDI spillover effects are welldocumented, and their mechanisms are relativelyeasy to understand. For example, spillovers canoccur when a technological gap exists between for-eign and local firms (Gerschenkron, 1962; Meyer,2004). This gap should not be so large that recip-ients of spillovers lack the capacity to absorbthe technology received (Kokko, 1994; Kokko,Tasini, and Zejan, 1996). We believe, however,

that spillover theory is not complete if it does notaccount for competition effects, and that it is neces-sary to identify the contexts in which competitioneffects may dominate positive spillover effects, orvice versa. A competition-based theory could helpidentify such contexts and may offer a possibleexplanation for negative spillovers.

Competitor analysis and the competitioneffects

Competition has been one of the central themesof the strategy field. Originally, competitive anal-ysis was conducted at the industry level, withinsights drawn from industrial organization eco-nomics (Bain, 1956; Porter, 1980). Rivalry insidean industry was recognized as a key compo-nent of the five forces that determine firm per-formance (Porter, 1980). Quickly, however, strat-egy researchers started looking for more preciseways to identify competitors at a subindustrylevel, namely the strategic group. This approachviews firms within a strategic group as a focalcompany’s most direct competitors (McGee andThomas, 1986; Barney and Hoskisson, 1990).

As the strategy field became increasingly inter-ested in firm-specific factors that contribute tocompetitive advantage, there arose a need to ana-lyze individual competitors of a focal company,and to predict the rivalry and interaction betweena pair of competitors. Chen (1996) integrates therelated literatures in competitor analysis and inter-firm rivalry (Chen and MacMillan, 1992; Fer-rier, 2001) to derive a framework for this pur-pose. He synthesizes the market-focused perspec-tive (Porter, 1980; Gimeno and Woo, 1994) andthe resource-based view (Barney, 1986; Conner,1994) to propose that two constructs, market com-monality and resource similarity, are antecedentsof interfirm rivalry. These constructs influence thedrivers of competitive behavior, awareness, moti-vation, and capability, which in turn influencethe likelihood of competitive attack and responsebetween two rivals (Chen, 1996; Yu and Cannella,2007; Chen, Su, and Tsai, 2007). More specifically,he predicts that greater market commonality andincreased resource similarity will lead to mutualforbearance, but provoke retaliation once an attackhas been launched.

Although this framework was developed to ana-lyze the competitive dynamics between two indi-vidual firms, it can be extended beyond the dyad



level. Market commonality refers to the extent ofmarket overlap among competitors, and the con-cept of market is broad enough to encompassgeographic market, market segment, and brand;resource similarity indicates the extent to whichcompetitors possess comparable types and levelsof strategic endowments (Chen, 1996). Companiestargeting the same, single geographic market havechosen the same niche based on location (Porter,1980, 1998). Baum and Mezias (1992) demon-strated that similar organizations located in thesame region competed more intensely and reducedeach other’s chance of survival. On the other hand,companies that have similar ‘sticky’ and immo-bile resources may be constrained to develop sim-ilar capabilities and take similar strategic positions(Collis, 1991; Teece, Pisano, and Shuen, 1997),thus becoming direct competitors.

We apply this framework to examine group-levelinfluences.1 While this scaling-up makes it difficultto observe details of interfirm rivalry and interac-tions, it increases the model’s predictability in away. Chen’s (1996) framework contains a mutualforbearance hypothesis and an escalating retalia-tion hypothesis: a pair of direct competitors mightrefrain from acting against each other, and willlikely retaliate when acted upon. But the frame-work does not predict when an initial attack willbe launched. The prospect that an attack wouldalmost certainly invite retaliation from direct com-petitors perhaps suggests that mutual forbearanceis more likely to be observed in a duopoly or anoligopoly situation. When there are large groupsof firms in an industry, however, it is much morelikely that some firms will initiate an attack inorder to achieve first-mover advantage (Lieber-man and Montgomery, 1988) with the hope thattheir competitors will not respond to this (Chenand MacMillan, 1992). Furthermore, some firmsmight act first to gain such first-mover advantagebecause they think their competitors might even-tually initiate attacks. Thus, with large groups offirms, initial attacks would be more frequent andmore certain. Once an attack has been launched,firms in the same market and with similar resource

1 Scaling up the competitor analysis framework from a dyadiclevel to groups of firms is equivalent to changing an individual-level theory to a group-level theory, according to Klein, Danse-reau, and Hall’s (1994) terminology. Our underlying assumptionis that foreign entrants and local firms differ markedly from eachother in their capabilities and strategies, and that they impactother firms as distinctive groups.

types will react, consistent with Chen’s (1996)prediction of escalating retaliation. Consequently,industry rivalry will become more intense. Marketcommonality and resource similarity should thusbe positively related to the level of competitionfor groups of firms. The value of Chen’s frame-work, therefore, can be enhanced by applying it toexamine competition among large groups of firmsalong some market and resource dimensions.

Spillovers and competition: the hypotheses

The combination of the FDI literature with thecompetitor analysis framework should provide amore complete picture that includes both the pos-itive effects of spillovers and the negative effectsof competition, with respect to both foreign andlocal firms, and better predict when one effect willdominate the other. We define spillover effects asthe positive influences caused by the presence ofa group of firms on members of another group,which enhance the latter’s chances of survival.Competition effects, on the other hand, are the neg-ative influences caused by the presence of a groupof firms on members of another group, whichdecrease the latter’s chances of survival. In fact,spillover and competition effects are two sides ofthe same coin. A firm is simultaneously a sourceof knowledge spillovers and a source of competi-tion to other firms in the same industry. Thus, whencompetition is moderate, spillover effects are morelikely to dominate; otherwise competition effectswill prevail. We rely on the dimensions of marketcommonality and resource similarity to gauge thelevel of competition among firms, and hence therelative size of spillover and competition effects.

Market commonality

The distinction between regional and national mar-kets significantly determines market commonality.Firms in the same region typically share a commongeographic market. They are also subject to thesame set of subnational institutional environments,which constrain their choices and force them todevelop similar market strategies (DiMaggio andPowell, 1983; Oliver, 1991). Thus, firms may havemore market commonality with firms in the sameregional market than they do with those situated indifferent regions. A regional market can be basedon physical access, such as the Manhattan hotelindustry (Baum and Mezias, 1992); on unique



sociocultural characteristics, such as Quebec inCanada; or on administrative divisions, such asthe Chinese provinces (Vanhonacker, 1997). Whena multinational firm enters an overseas market,it must choose a specific location for its foreignsubsidiary. Traditionally, researchers have exam-ined the location strategies of multinational firmsas a function of ownership and location advan-tages (Dunning, 1981; Buckley and Ghauri, 2004),and have not paid much attention to the marketcompetition factors in the host country. Recently,however, researchers have started to consider thechallenges that local firms exert on foreign entrantsin regional markets. Miller and Eden (2006), forinstance, found that U.S. subsidiaries of foreigncommercial banks had a higher exit rate in regionswith higher local density.

A domestic incumbent is likely to see the entryof foreign firms as a serious threat, especially whenit is locally bound (Dawar and Frost, 1999). Multi-national firms typically enjoy technological supe-riority and strong brand loyalty. They are highlyvisible, and can secure preferential treatment fromlocal governments due to their strong bargainingpower (Moran, 1985; Kim, 1988). As a result, theirentry into a host market will often bring shockand disequilibrium to a local industry, heighten-ing the awareness of local firms (Chen, 1996), andthe entrants will be seen as aggressive intruders.Under such circumstances, forbearance is unlikely,and attacks and counterattacks will be frequent,causing industry rivalry to intensify.

The resultant competitive tension (Chen et al.,2007) can be particularly salient in emerging mar-kets, which often comprise many regional mar-kets that have significantly different income levelsand customer demands. National markets typicallyhave not been established for most industries insuch countries, and in the absence of national eco-nomic drivers, such as a national distribution sys-tem, most of the domestic firms compete region-ally (Prahalad and Lieberthal, 1998; Peng, Tan,and Tong, 2004). Correspondingly, the competitioneffects between foreign entrants and local incum-bents will also be stronger in a regional marketthan it is in the national market, because, whenfaced with foreign entrants, the regionally boundincumbents have no other means but to respond tothe intruders in their regions by direct retaliation.

The spillover effect, on the other hand, canbe a more universal phenomenon. Although theagglomeration literature may suggest that firms can

benefit from being proximate to each other withincertain geographic boundaries (Krugman, 1991;Saxenian, 1994; Chang and Park, 2005), thereare many reasons to believe that spillover effectscan go beyond narrowly defined local boundariesand become more pronounced at the country level(Keller, 2002). First, through one major mecha-nism of FDI spillovers—the demonstration effect(Caves, 1974)—spillovers can take place acrossregions within a nation. This effect occurs whenthe introduction of new foreign technologies andproducts inspires domestic entrepreneurs and inno-vators to develop these goods for their home mar-kets. National boundaries, not regional ones, pre-sumably limit domestic entrepreneurs’ access toinformation about foreign technologies and prod-ucts.

Second, the spillover of advanced knowledgeand technologies across regions may be facilitatedby the higher mobility of better educated employ-ees, who are less bound to local job markets (Ahn,de la Rica, and Ugidos, 1999). For example, localfirms that do not compete directly with foreignfirms in their own regional markets can learn fromforeign or other local firms by hiring managers andengineers from them.

Third, local governments in many countries haveencouraged greenfield investments and ‘friendly’mergers and acquisitions as a way to reach eco-nomic growth targets. In China, for instance, animportant promotion criterion for local governmentofficials is the amount of investment they attract tothe region. Knowledge spillovers are facilitated bycross-regional investments and can easily be usednationwide.

Summarizing the above arguments, we expectthe strong presence of foreign firms or local firmsin a regional market exerts competition effectson both their peers and other firms that com-pete directly with them in the same regional mar-ket. The presence of those firms in a nationalmarket does not, however, pose direct competi-tion to firms in other regions, and thus allowsthe positive spillover effects to dominate nation-ally.

Hypothesis 1. Spillover effects are more evidentin a national market, and competition effectsare more evident in a regional market, amongdifferent groups of firms.



Resource similarity

Foreign entrants and local firms have differentresource profiles. In analyzing how local compa-nies in emerging markets can compete with multi-national ‘giants,’ Dawar and Frost (1999) note thatit is important to consider the type of strategicassets each has accumulated: those that are trans-ferable abroad versus those that are customized toa specific home market. Relative to local competi-tors, multinational firms possess proprietary assets,often in the form of advanced technologies, brandnames, and managerial know-how that they cantransfer to their foreign subsidiaries (Buckley andCasson, 1976; Dunning, 1988; Hitt et al., 2000).They also have abundant capital and experiencedexpatriate managers who can be assigned to manysites worldwide.

In contrast, domestic incumbents typically enjoylocally embedded advantages such as marketingand distribution channels, access to information,and network connections (Beamish, 1988; Lu andXu, 2006). Their managers have usually beeneducated and trained domestically and thoroughlyknow their home market. Multinational firms valuethe assets and resources that local firms possess,and local firms lack the transferable assets thatmultinational firms possess.

Yet although multinational and local firmseach possess distinct types of organizational,financial, and human resources, there are variationswithin each group. Some multinational firmshave acquired host-country-specific assets becausethey are culturally and ethnically proximate to ahost country, have prior operational experience,or have internalized certain local knowledgethrough their joint venture partners (Makino andDelios, 1996; Luo, 1997). Conversely, somelocal firms have developed transferable assets byrestructuring, innovating, and internationalizing,and they employ these assets to compete withforeign entrants (Dawar and Frost, 1999; Zeng andWilliamson, 2003).

Thus, in terms of strategic resources, differenttypes of foreign and local firms may be more orless dissimilar to each other. Among foreign andlocal firms, foreign firms that have no local tiesand conventional local firms that have no trans-ferable assets lie at the extremes of the transfer-able assets—locally embedded resources dimen-sion—and thus have the most dissimilar resource

profiles. Given their distinct competitive advan-tages, each group is likely to take very differentstrategic positions in an industry: multinationalfirms probably enjoy greater operating efficiencyand market power, and conventional local firmscapitalize on their better regional networks andknowledge. Such differences make it less likelythese groups will compete directly, and thus facili-tate cooperation and knowledge spillovers betweenthem. On the other hand, more localized foreignfirms that have acquired country-specific assetswill compete more directly with local firms, andthreaten the latter’s survival. Their entries intolocal markets will more likely be seen as aggres-sive attacks, and local firms are more likely toretaliate. Similarly, reformed or internationalizedlocal firms that have developed assets comparableto those of multinationals will more likely chal-lenge foreign entrants, compete directly with them,and crowd them out of the country.2

Hypothesis 2. Spillover effects are more evi-dent among groups of firms with dissimilarresource profiles, and competition effects aremore evident among groups of firms with similarresource profiles.

EMPIRICAL CONTEXT ANDRESEARCH DESIGN

Research setting

China provides an ideal setting for our research.First, it has recently liberalized its economy. Asa consequence, there are many local firms withvarying resource profiles and capabilities. Second,the Chinese government has encouraged FDI inorder to modernize its backward industries. Thereare numerous multinationals that operate in a fullrange of industries in China. Third, competitionbetween foreign entrants and local firms, amongforeign firms, and among local firms has intensi-fied in recent years; thus providing an interesting

2 One might argue that spillover effects could also be greateramong groups of firms with similar resource profiles, becausesimilar resource profiles might suggest higher absorptive capac-ity among these groups. Nonetheless, the absorptive capacityargument and its counterargument, the technological gap hypoth-esis, are more associated with the level of resources than theyare with the type of resources (Cantwell, 1989; Meyer, 2004).Because our concept of resource similarity is based on the typeof resource, we make no inferences from the absorptive capacityargument.



setting in which to examine the mutual influencesamong the players. Finally, because China com-prises many large, heterogeneous regional mar-kets, it is an interesting setting in which toobserve spillover and competition effects along ageographic-market dimension.

The most important event contributing toChina’s progress in the past quarter century is the‘reform and open-door policy,’ which the Commu-nist government adopted in 1978. Through reform,many traditional state-owned enterprises (SOEs)that had monopolies were dismantled and trans-formed into shareholding or limited liability enter-prises with a modern ownership structure, or pri-vatized. The number of collectives also decreased.Overall, the number of ‘conventional local firms,’namely SOEs and collectives, declined, and thenumber of ‘reformed local firms,’ including pri-vate firms, shareholding firms, and limited liabilityfirms, increased rapidly. As a group, ‘reformedlocal firms’ comprise the modernized and restruc-tured Chinese companies.

In the meantime, the open-door policy led toincreasing amounts of FDI, first in the coastalregions, and subsequently in the inland, rural areas.The first foreign investments established in thespecial economic zones (SEZs) were a showcaseof modern technology and management practices,which quickly disseminated to other regions—aperfect example of the demonstration effect.

An important dimension of competition in Chinais the geographic market. China is a collectionof regional markets with enormous differencesin income levels, consumer tastes, and subcul-tures. Roughly speaking, each Chinese provinceis a regional market; over time, China developeda system of de facto ‘federalism’ and regionaldecentralization at the provincial level (Jin, Qian,and Weingast, 2005). Differences in governmentalpolicies between provinces effectively segmentedregional markets (Vanhonacker, 1997). The oper-ations of multinational firms in China have beenbound by these regional boundaries (Prahalad andLieberthal, 1998).

Another important aspect of competition inChina is the different resource endowments firmspossess (Peng and Luo, 2000; Peng et al., 2004;Ralston et al., 2006). Foreign firms, like multina-tional firms in other emerging markets, have themost liquid and transferable assets. In contrast,the SOEs have monopolies over state-controlled

resources such as land and oil. They are embed-ded in a bureaucratic network because their man-agers are often former party or governmental offi-cials. Other local firms, which lack the advan-tages possessed by multinationals and SOEs, havehad to develop their own transferable assets inorder to compete. Some of them, such as Lenovo,Huawei, and Baidu, have accumulated enoughskills and resources to achieve competitive advan-tages over their multinational competitors, conven-tional SOEs, and collectives.

Foreign entrants in China can be classifiedinto two large categories. Much inward FDI inChina comes from Hong Kong, Macau, and Tai-wan (HMT). Because of the unique historicaland social linkages between these regions andmainland China, multinational firms from theseGreater China areas (hereafter HMT firms) sharemany Chinese characteristics and differ substan-tially from the non-Chinese multinationals (here-after non-HMT foreign firms). Because of theircultural origin, HMT firms enjoy some access tolocal knowledge and locally embedded resources,and therefore may constitute a serious threat tolocal firms. On the other hand, although non-HMTforeign firms are from a wide range of coun-tries and may each have distinctive resources oftheir own, they share some common characteris-tics. Most possess transferable proprietary assetsthat local firms cannot access, but lack the locallyembedded resource advantages enjoyed by indige-nous firms and, to some extent, by the HMT firms.In this sense, they can be considered a distinctivegroup.

Although spillover effects have benefited Chi-nese companies since the early 1980s, the largescale entry of multinational firms with superiorcapabilities has often been interpreted as a collec-tive attack on local incumbents that merits retalia-tion. In the meantime, although both non-HMT for-eign firms and HMT firms might generate spillovereffects for local firms, because HMT firms haveresources that are more similar to local firms thannon-HMT firms do, their spillover effects maybe substantially offset by the stronger competi-tion effects they exert on local firms. Conversely,among local firms, effective competitive responsesto entry by multinationals is more likely to comefrom reformed local firms than it is from conven-tional local firms because the former have donemore to imitate the assets, resources, ownership



structures, and incentive schemes of multinationalentrants.

Sample

This study used the Annual Industrial SurveyDatabase (1998–2005) of the Chinese NationalBureau of Statistics (NBS). The NBS collectsfinancial information on industrial firms and pub-lishes aggregated information in the official ChinaStatistics Yearbooks. By law, all firms in China arerequired to cooperate with the survey and to submitfinancial information. Prior to 1998, the NBS col-lected information only from firms above the town-ship administrative level,3 so this database did notinclude information from private firms. In 1998,the NBS expanded its database coverage to include1) all SOEs, and 2) all non-SOE firms, includ-ing foreign firms, with annual sales of at least5 million RMB (roughly, $600,000 US, accordingto the official 2005 exchange rate) in the year priorto the survey. Foreign firms are legally defined byvarious relevant laws. In cases for which there areequity joint ventures or foreign-invested sharehold-ing enterprises (including those from Hong Kong,Macao, and Taiwan), the laws normally requirethe foreign or HMT partner to hold at least a 25percent share of the registered capital.4 The num-ber of firms included in this database ranged from162,033 to 280,188. These numbers are consis-tent with those reported in the official Yearbookspublished by the NBS. Chow (1993) confirmedthat the NBS statistics are largely accurate andinternally consistent for empirical analysis. Sev-eral prior studies have used the NBS databases,including Pan, Li, and Tse (1999), Buckley et al.(2002), and Park, Li, and Tse (2006).

The annual survey database contains key finan-cial indicators as well as demographic informationsuch as the firm’s name, manager’s name, year

3 Administratively, all non-private Chinese firms correspond to alevel in the governmental hierarchy. Some large SOEs functionat the level of central ministries (their managers are ministry-level officials), and other firms correspond to regional or sub-regional hierarchies. These levels of hierarchy include: central,provincial, regional, county, district/township, and street/village.4 Huang (2003 : 37–41) notes cases of ‘round-trip foreign directinvestments,’ for which Chinese firms or individuals investedin China under disguised foreign identities either to laundermoney or to exploit subsidies to foreign firms. We exploredthe possible bias introduced by these firms by dropping foreigndirect investments from Hong Kong and the Cayman Islands.Our results do not vary.

of establishment, address, ownership, and level ofgovernmental control. For this study, we assem-bled a panel database by matching yearly data witha unique company identifier. There were manycases in which the same firms with exactly thesame names and the same addresses had differentfirm identifiers in different years because signif-icant changes in ownership, such as joint ven-tures and mergers and acquisitions, occurred. Wetherefore developed a detailed software algorithmto assess whether firms’ demographic informationmatched with the same firms’ observations acrossyears. For each year, 11–19 percent of the samplefirms exited the database, and 11–22 percent ofthe firms that appeared in the database were new.5

Figure 1 shows the percentages of firms as clas-sified by ownership type. It shows that the pro-portion of conventional local firms (i.e., SOEs andcollectively-owned firms) declined from more than70 percent in 1998 to less than 16 percent in 2005.These firms were replaced by reformed local firms(i.e., private firms, shareholding firms, and lim-ited liability companies). At the same time, theshare of multinational firms increased from 16 per-cent to more than 21 percent. For the early yearsof our sample, slightly more than half of thesemultinationals were from Hong Kong, Macao, and

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1998 1999 2000 2001 2002 2003 2004 2005

non-HMK Foreign firms

HK, Macao,Taiwanese firms

Reformed locals

Conventional locals

Number offirms 165,186 162,107 163,000 169,344 182,094 196,904 280,100 273,513

Figure 1. Distribution of firms according to ownershiptype

5 The percentages of firms entered were based on 1998–2005,except for 2003–2004. Between 2003 and 2004, the numberof firms in the annual survey increased sharply due to theChinese government’s effort to identify potential tax payers. Asa consequence, 127,738 firms were added in 2004.



Taiwan. By 2005, the proportions of non-HMT for-eign firms and HMT firms were roughly equal.

Measurements

In order to avoid survivor bias, we used firm sur-vival as the criterion for assessing the effects ofspillovers and competition. Several works haveexamined the survival/exit of incumbent domes-tic businesses (Duhaime and Grant, 1984; Chang,1996; Sharma and Kesner, 1996). Shaver, Mitchell,and Yeung (1997) demonstrated that foreign en-trants enhanced their survival by learning from ear-lier entrants’ success/failure. We treat Survival as adichotomous variable that notes survival/exit. Exitoccurs when a valid observation at time t does notshow up in the database at time t + 1, after it hasmatched with various demographic information inseveral instances. Firms that exited might haveclosed down or been sold, or their sales might havedipped below 5 million RMB in the case of non-SOEs. In order to ensure that we did not falselyidentify exit in cases for which firms used incon-sistent identifiers, we used the algorithm describedabove. For some questionable cases, we checkedfirms’ identities more closely by reading their self-descriptions.

Prior work on FDI spillovers has used sev-eral variables to reflect foreign firms’ presencein an industry. For instance, Aitken and Harri-son (1999) used the employment-weighted per-centage of equity owned by foreign firms. Javor-cik (2004) measured foreign firms’ presence asthe share of firm’s total equity owned by for-eign investors. Most studies, however, have usedforeign employment shares to predict positivespillover effects (Caves, 1974; Driffield and Mun-day, 2000; Liu et al., 2000; Buckley et al., 2002;Kosova, 2005). We used the employment shares ofvarious groups of firms to measure their presence.These groups include non-HMT foreign firms,HMT firms, reformed locals (private, sharehold-ing, and limited liability companies), and conven-tional locals (SOEs and collectives). We definedan industry with the three-digit Chinese StandardIndustry Classification and measured the aggre-gated shares of employment for three types offirms at time t: National non-HMT Foreign Firms’Employment, National HMT Firms’ Employment,and National Reformed Local Firms’ Employment.The National Conventional Local Firms’ Employ-

ment was treated as the base group in all our anal-yses. We also measured Regional non-HMT For-eign Firms’ Employment, Regional HMT Firms’Employment, and Regional Reformed Local Firm’sEmployment with the share of employment of eachtype of firm in an industry in a region. We definedregions using two-digit area codes for provinces,autonomous regions, and centrally administeredcities (Beijing, Chongqing, Shanghai, and Tianjin).

We used these measures to gauge the net ofspillover and competition effects6 as determinedby the market commonality and resource similar-ity dimensions. Consistent with our market com-monality hypothesis (Hypothesis 1), we expect theregional market-level employment share variablesto have a more negative effect than the correspond-ing national market-level variables do. Applyingour resource similarity hypothesis (Hypothesis 2)to the Chinese context, we expect to see, for exam-ple, that HMT firms’ employment affects local firmsurvival more negatively than non-HMT foreignfirms’ employment does, and that reformed localfirms’ employment affects foreign firm survivalmore negatively than conventional local firms’employment does.

We included several firm characteristics as con-trol variables. We measured Firm Size with a firm’ssales at time t, which we log-transformed. We mea-sured Firm Age from the years of establishment totime t. We measured Profitability with return onassets. We measured the degree of decentralizedcontrol for local firms with Decentralized Con-trol, which ranges from 1 (central government) and2 (province) to 10 (privately owned) (Park et al.,2006). We also included a dichotomous variable,Reformed Local Firms, to differentiate reformedlocal firms (coded 1) from conventional locals(coded 0). For foreign firms, we used a dichoto-mous variable, HMT Firms, to separate HMT firms(coded 1) from non-HMT foreign firms. We alsomeasured ownership control by denoting whetherforeign operations were created from scratch orwere formed by either acquiring or creating jointventures with local firms. We measured ForeignJV & Acquisition as a dichotomous variable, for

6 Kosova (2005) modeled competition effects (which she referredto as crowding-out effects) as a function of year-to-year salesgrowth by foreign firms. Conversely, she used foreign employ-ment share to predict positive spillover effects. We do not believethat any measure can easily separate spillover effects from com-petition effects because multinationals can be both the source ofspillovers and competitors that drive other firms out of business.



which we defined foreign greenfield investment asthe baseline.

Methods

We assembled a panel dataset consisting ofboth foreign and local firms in China during1998–2005. Because we wanted to observe firms’survival chances, we needed to have a one-yearlag, and thus lost observations for 2005. Usinga one-year lag between explanatory variables anddependent variables also helped us control forreverse causality. Thus, we had seven years offirm-level data points, for a total of 1,318,823 firm-year observations. We had 35,648 observationswith missing information on key variables andunrealistic numbers that were possibly due to dataentry errors. After removing those observations,we had 1,273,175 valid firm-year observations,which consisted of 1,032,106 local firm-years,including both conventional local firms andreformed local firms, and 241,069 foreign firm-years, including both non-HMT foreign firms andHMT firms.

For both samples, we employed a discrete timelogit specification (Allison, 1984) to estimate mod-els with survival/exit as the dependent variable.We used virtually the same set of independentvariables for both samples, except for some con-trol variables (i.e., Level of Control and ReformedLocal Firms for local firm analysis, and ForeignJV and Acquisition and HMT Firms for foreignfirm analysis). The unit of analysis in these mod-els is the firm-year. Each calendar year, startingfrom 1998 to a firm’s year of exit or the censoringyear (i.e., 2005), constitutes a valid observation.The estimated coefficients provide the changes inthe log-odds of survival for each one-unit increasein the independent variables. Our inclusion of theindustry, area, and year fixed effects allows thebaseline hazard rate (i.e., the intercept) to varyacross possible combinations of industry, region,and year. By design, both the dependent variable(a zero/one dichotomous variable that indicatessurvival/exit of each firm at time t + 1) and inde-pendent variables (measured at time t) are time-varying. We used full samples of local and foreignfirms, including those that exited at time t + 1, inthe survival models, which means these modelsare not subject to survivor bias. We also calcu-lated clustered robust standard errors by allowing

‘clustering’ of error terms to correlate among thesame firms.

RESULTS

Table 1 shows descriptive statistics and correla-tions, and Tables 2 and 3 show results for localfirm survival and foreign firm survival, respec-tively. The descriptive statistics suggest that em-ployment shares defined at the national market andregional market are highly correlated. In Tables 2and 3, we entered employment share variablesdefined at the national market level and at theregional market level sequentially in a set of threemodels to detect changes in signs when we addednew sets of variables. In all models, the coeffi-cients of the conventional local firms in nationaland regional markets were set to zero as the refer-ence group.7

Hypothesis 1 predicts a positive spillover effecton all local firms, including conventional localfirms and reformed local firms, from non-HMTforeign firms, HMT firms, and the reformed locals,in the national market. In Table 2, national non-HMT foreign firm share and national reformedlocal firm share are positively significant for localfirm survival in Model 1, suggesting spillovereffects from non-HMT foreign firms and reformedlocal firms to other local firms in the nationalmarket. Hypothesis 1 also predicts a stronger com-petition effect in regional markets. In Model 2,the regional employment shares of HMT firmsand reformed local firms are significantly nega-tive for the survival of local firms in their ownregional markets. When we include the employ-ment shares for both national and regional mar-kets, Model 3 confirms Hypothesis 1: competitioneffects are more likely to outweigh spillover effectsin a regional market than they are in a nationalmarket.

Hypothesis 2 explores whether resource simi-larity among different types of firms causes moresevere competition and leads to lower survivalrates. We examine Hypothesis 2 by comparing

7 The employment shares of these four types of firms are summedto be one. Thus, we need to choose one type of firm as areference group (e.g., conventional local firms), and interpretcoefficients with respect to those of the reference group. Doingso is equivalent to setting the coefficients of the conventionallocal firms to zero and interpreting other groups’ coefficients asthe differences from those of conventional local firms.



Table 1. Descriptive statistics

a) local firms (N = 1, 032, 106)

Mean Std. Dev Mini Max

(1) Survival (t + 1) (= 1) 0.85 0.36 0.00 1.00(2) National non-HMT foreign firm’s employment (t) 0.08 0.07 0.00 0.59(3) National HMT firm’s employment (t) 0.10 0.10 0.00 0.73(4) National reformed local firm’s employment (t) 0.38 0.18 0.00 1.00(5) Regional non-HMT foreign firm’s employment (t) 0.08 0.11 0.00 1.00(6) Regional HMT firm’s employment (t) 0.09 0.13 0.00 1.00(7) Regional reformed local firm’s employment (t) 0.39 0.25 0.00 1.00(8) Firm size (t) 9.41 1.54 0.00 18.54(9) Firm age (t) 15.55 15.01 0.00 184.00

(10) Firm ROA (t) 5.99 20.75 −199.50 989.10(11) Level of control (t) 6.31 2.21 0.00 10.00(12) Reformed local firm (t) (vs. conventional local firms) 0.48 0.50 0.00 1.00

Correlation matrix

1 2 3 4 5 6 7 8 9 10 11

2 −0.02∗∗∗

3 −0.03∗∗∗ 0.73∗∗∗

4 0.00 0.02∗∗∗ −0.12∗∗∗

5 −0.02∗∗∗ 0.59∗∗∗ 0.45∗∗∗ 0.02∗∗∗

6 −0.02∗∗∗ 0.40∗∗∗ 0.52∗∗∗ −0.04∗∗∗ 0.30∗∗∗

7 0.01∗∗∗ 0.11∗∗∗ 0.01∗∗∗ 0.71∗∗∗ −0.09∗∗∗ −0.17∗∗∗

8 0.17∗∗∗ 0.02∗∗∗ −0.01∗∗∗ 0.18∗∗∗ 0.03∗∗∗ 0.04∗∗∗ 0.18∗∗∗

9 0.05∗∗∗ −0.16∗∗∗ −0.16∗∗∗ −0.15∗∗∗ −0.12∗∗∗ −0.15∗∗∗ −0.17∗∗∗ 0.0010 0.01∗∗∗ 0.03∗∗∗ 0.03∗∗∗ 0.04∗∗∗ 0.03∗∗∗ 0.00 0.04∗∗∗ 0.15∗∗∗ −0.14∗∗∗

11 0.00 0.22∗∗∗ 0.20∗∗∗ 0.33∗∗∗ 0.14∗∗∗ 0.18∗∗∗ 0.34∗∗∗ 0.06∗∗∗ −0.44∗∗∗ 0.16∗∗∗

12 0.04∗∗∗ 0.24∗∗∗ 0.18∗∗∗ 0.40∗∗∗ 0.14∗∗∗ 0.13∗∗∗ 0.44∗∗∗ 0.21∗∗∗ −0.32∗∗∗ 0.05∗∗∗ 0.51∗∗∗

b) foreign firms (N = 241, 069)

Mean Std. Dev Mini Max

(1) Survival (t + 1) (= 1) 0.90 0.30 0.00 1.00(2) National non-HMT foreign firm’s employment (t) 0.14 0.09 0.00 0.59(3) National HMT firm’s employment (t) 0.19 0.13 0.00 1.00(4) National reformed local firm’s employment (t) 0.35 0.16 0.00 0.84(5) Regional non-HMT foreign firm’s employment (t) 0.17 0.16 0.00 1.00(6) Regional HMT firm’s employment (t) 0.25 0.22 0.00 1.00(7) Regional reformed local Firm’s employment (t) 0.30 0.21 0.00 1.00(8) Firm size (t) 10.19 1.29 0.00 18.09(9) Firm age (t) 7.18 5.77 0.00 126.00

(10) Firm ROA (t) 4.29 15.51 −197.80 968.17(11) Foreign joint ventures & acquisition (vs. greenfield invst) (t) 0.47 0.50 0.00 1.00(12) HK, Macao, Taiwanese firm (t) (vs. non-HMT foreign firm) 0.56 0.50 0.00 1.00

the coefficients associated with the employmentshares of various groups of firms in Table 2. Asdiscussed earlier, HMT firms’ resource profilesare more similar to those of local firms thannon-HMT foreign firms are. Any spillover effectsfrom HMT firms are therefore likely to be offset

by the strong competition effects they exhibit onlocal firm survival. Indeed, HMT firms’ employ-ment share in a national market is insignificant,but non-HMT foreign firms’ employment share ispositively significant. Conversely, regional HMTfirms’ share negatively affects local firm survival,



Correlation matrix

1 2 3 4 5 6 7 8 9 10 11

2 0.003 −0.01∗∗∗ 0.62∗∗∗

4 0.00 −0.23∗∗∗ −0.39∗∗∗

5 0.02∗∗∗ 0.56∗∗∗ 0.29∗∗∗ −0.10∗∗∗

6 0.01† 0.26∗∗∗ 0.52∗∗∗ −0.24∗∗∗ −0.11∗∗∗

7 −0.02∗∗∗ −0.15∗∗∗ −0.26∗∗∗ 0.67∗∗∗ −0.24∗∗∗ −0.48∗∗∗

8 0.21∗∗∗ 0.06∗∗∗ −0.05∗∗∗ 0.08∗∗∗ 0.09∗∗∗ −0.01∗∗ 0.02∗∗∗

9 0.01∗∗ −0.03∗∗∗ −0.01∗∗∗ 0.08∗∗∗ −0.03∗∗∗ 0.02∗∗∗ 0.03∗∗∗ 0.11∗∗∗

10 0.07∗∗∗ 0.02∗∗∗ −0.02∗∗∗ 0.05∗∗∗ 0.03∗∗∗ −0.07∗∗∗ 0.07∗∗∗ 0.19∗∗∗ −0.01∗∗∗

11 −0.04∗∗∗ 0.16∗∗∗ 0.06∗∗∗ 0.37∗∗∗ 0.07∗∗∗ −0.01∗∗∗ 0.29∗∗∗ −0.12∗∗∗ −0.18∗∗∗ 0.04∗∗∗

12 −0.03∗∗∗ −0.03∗∗∗ 0.11∗∗∗ −0.08∗∗∗ −0.22∗∗∗ 0.36∗∗∗ −0.13∗∗∗ −0.09∗∗∗ 0.07∗∗∗ −0.06∗∗∗ −0.04∗∗∗

Note; ∗∗∗ : p < 0.001, ∗∗ : p < 0.01, ∗ : p < 0.05, † : p < 0.10

but the effect of regional non-HMT foreign firms’share is insignificant. Overall, the results suggestthat HMT firms pose greater threats to local firmsthan non-HMT foreign firms do, lending supportfor Hypothesis 2.

The effects of reformed local firms on other localfirms present somewhat conflicting results. Thecoefficients for employment shares of reformedlocal firms in the national market have positivelysignificant effects on the survival of local firms, butthose for regional markets have negatively signifi-cant effects. The positive impact of reformed localfirms on other local firms in the national market,although consistent with Hypothesis 1, seems con-tradictory to Hypothesis 2, which predicts moresevere competition among similar types of firms.On the other hand, the reformed local firms’ sharein regional markets is negatively significant, con-sistent with both Hypotheses 1 and 2, suggestingthat reformed local firms tend to crowd out otherlocal firms in the same regional market.

In order to resolve these conflicting results, weran separate models for reformed local firms andconventional local firms in Models 4 and 5, respec-tively. In general, conventional local firms enjoymore positive spillover effects than reformed localfirms do, from the presence of both reformed localfirms and foreign firms. The positive spillovereffects from non-HMT firms at the national level,observed in Model 3 with a sample of all localfirms, seem to be driven by conventional localfirms. Conventional local firms are being crowdedout only by the presence of reformed local firmsin the same regional market. On the other hand,reformed local firms, whose resource profiles aresomewhat similar to foreign firms, are crowdedout by non-HMT foreign firms and HMT firms

at the national level as well as at the regionallevel. These additional models provide strong sup-port for Hypothesis 2, which proposes spillovereffects are greater among groups with more hetero-geneous resource profiles and competition effectsare greater among groups with more homogenousresource profiles.8

Various firm-level controls suggest that the largerand the older local firms are, the more likely theyare to survive. Our models show profitability attime t is negatively associated with survival, whichmight reflect the government’s unwillingness todivest unprofitable SOEs. The results show thatthe decentralization of government control has noeffect on local firms’ survival. Between two typesof local firms, reformed local firms have highersurvival rates than conventional local firms do.

Table 3 shows regression results for the survivalof foreign entrants, including non-HMT foreignfirms and HMT firms. The employment shares ofnon-HMT foreign firms, HMT firms, and reformedlocal firms in the national market were negativelysignificant for foreign firm survival in Model 1,indicating competition was so severe that foreignfirms were crowded out of the national marketboth by other foreign firms and by reformed localfirms. Model 2 shows that employment shares ofthese three groups in regional markets were neg-atively significant for foreign firm survival, whichsuggests foreign firms were also crowded out of

8 While the purpose of this study is to examine intergroup influ-ences, the positive intragroup effect of reformed local firms atthe national level remains a puzzle. A possible explanation is thata large presence of reformed local firms, as a new organizationalform, causes strong legitimating effects on themselves—an insti-tutional process identified in previous research (Li, Yang, andYue, 2007).



Tabl

e2.

Surv

ival

oflo

cal

firm

sat

time

t+1

duri

ng19

98–

2005

Var

iabl

eA

lllo

cal

firm

sR

efor

med

loca

lfir

ms

Con

vent

iona

llo

cal

firm

s

(1)

(2)

(3)

(4)

(5)

Em

ploy

men

tsh

are

offir

ms

Nat

iona

lno

n-H

MT

fore

ign

0.51

0.54

−0.5

51.

06in

natio

nal

mar

kets

firm

shar

e(t)

(0.1

7)∗∗

(0.1

7)∗∗

(0.3

5)∗

(0.2

4)∗∗

∗

Nat

iona

lH

MT

firm

shar

e(t

)−0

.12

−0.0

5−0

.86

0.27

(0.1

5)(0

.15)

(0.2

2)∗∗

∗(0

.21)

Nat

iona

lre

form

edlo

cal

firm

0.44

0.50

0.27

0.50

shar

e(t

)(0

.07)

∗∗∗

(0.0

7)∗∗

∗(0

.11)

∗(0

.09)

∗∗∗

Nat

iona

lco

nven

tiona

llo

cal

firm

shar

e(t

)0

00

0

Em

ploy

men

tsh

are

offir

ms

Reg

iona

lno

n-H

MT

fore

ign

−0.0

3−0

.05

−0.1

90.

06in

regi

onal

mar

kets

firm

shar

e(t)

(0.0

4)(0

.04)

(0.0

6)∗∗

(0.0

6)R

egio

nal

HM

Tfir

msh

are

(t)

−0.1

1−0

.13

−0.3

7−0

.43

×10

−2

(0.0

4)∗∗

(0.0

4)∗∗

(0.0

6)∗∗

∗(0

.05)

Reg

iona

lre

form

edlo

cal

firm

−0.0

5−0

.08

−0.1

2−0

.06

shar

e(t

)(0

.02)

∗(0

.02)

∗∗∗

(0.0

4)∗∗

(0.0

3)∗

Reg

iona

lco

nven

tiona

llo

cal

firm

shar

e(t

)0

00

0

Firm

char

acte

rist

ics

Firm

size

(t)

0.34

0.34

0.34

0.51

0.31

(0.2

4×

10−2

)∗∗∗

(0.2

4×

10−2

)∗∗∗

(0.2

4×

10−2

)∗∗∗

(0.5

2×

10−2

)∗∗∗

(0.2

8×

10−2

)∗∗∗

Firm

age(

t)0.

010.

010.

010.

010.

01(0

.26

×10

−3)∗∗

∗(0

.26

×10

−3)∗∗

∗(0

.26

×10

−3)∗∗

∗(0

.46

×10

−3)∗∗

∗(0

.31

×10

−3)∗∗

∗

Firm

profi

tabi

lity

(t)

−0.1

3×

10−2

−0.1

3×

10−2

−0.1

3×

10−2

−0.1

6×

10−2

−0.8

5×

10−3

(0.1

5×

10−3

)∗∗∗

(0.1

5×

10−3

)∗∗∗

(0.1

5×

10−3

)∗∗∗

(0.2

7×

10−3

)∗∗∗

(0.1

8×

10−3

)∗∗∗



Tabl

e2.

(Con

tinu

ed)

Var

iabl

eA

lllo

cal

firm

sR

efor

med

loca

lfir

ms

Con

vent

iona

llo

cal

firm

s

(1)

(2)

(3)

(4)

(5)

Lev

elof

cont

rol

0.47

×10

−30.

65×

10−3

0.59

×10

−30.

03−0

.05

(0.1

9×

10−2

)(0

.19

×10

−2)

(0.1

9×

10−2

)(0

.27

×10

−2)

(0.2

8×

10−2

)∗∗∗

Ref

orm

edlo

cal

firm

s(v

s.0.

300.

300.

30co

nven

tiona

llo

cals

)(0

.01)

∗∗∗

(0.0

1)∗∗

∗(0

.01)

∗∗∗

Inte

rcep

t−1

.43

−1.3

5−1

.44

−3.4

1−0

.74

(0.0

4)∗∗

∗(0

.04)

∗∗∗

(0.0

4)∗∗

∗(0

.08)

∗∗∗

(0.0

5)∗∗

∗

N1,

032,

106

1,03

2,10

61,

032,

106

493,

287

538,

817

Wal

dch

i-sq

(d.f

.)42

327.

742

235.

442

358.

619

194.

042

358.

6(2

43)∗∗

∗(2

43)∗∗

∗(2

46)∗∗

∗(2

44)∗∗

∗(2

46)∗∗

∗

Not

e;∗∗

∗:

p<

0.00

1,∗∗

:p

<0.

01,

∗:

p<

0.05

,†

:p

<0.

10.

198

indu

stry

dum

mie

s,30

regi

ondu

mm

ies,

and

6ye

ardu

mm

ies

are

not

show

n.N

umbe

rsin

pare

nthe

ses

are

robu

stst

anda

rder

rors

adju

sted

for

clus

teri

ngon

indi

vidu

alfir

ms.

Inca

seal

lfir

ms

inou

rsa

mpl

eex

itan

indu

stry

ina

year

,th

ein

dust

rydu

mm

ype

rfec

tlyex

plai

nsth

eex

itlik

elih

ood

and

ther

efor

eth

ose

obse

rvat

ions

are

drop

ped

inth

ees

timat

ion.

Whe

nw

ebr

eak

upsa

mpl

esto

conv

entio

nal

and

refo

rmed

loca

lfir

ms,

we

tend

tofin

dm

ore

such

case

s.T

hus,

the

num

ber

ofob

serv

atio

nsin

Mod

els

4an

d5

dono

tad

dup

toth

atof

am

odel

usin

gal

llo

cal

firm

s.



Tabl

e3.

Surv

ival

offo

reig

nfir

ms

attim

et+

1du

ring

1998

–20

05

Var

iabl

eA

llfo

reig

nfir

ms

Non

-HM

Tfo

reig

nfir

ms

HM

Tfo

reig

nfir

ms

(1)

(2)

(3)

(4)

(5)

Em

ploy

men

tsh

are

Nat

iona

lno

n-H

MT

−1.6

6−1

.61

−1.3

0−1

.59

offir

ms

info

reig

nfir

msh

are(

t)(0

.32)

∗∗∗

(0.3

3)∗∗

∗(0

.51)

∗(0

.44)

∗∗∗

natio

nal

mar

kets

Nat

iona

lH

MT

firm

−1.1

9−0

.74

−1.0

4−0

.72

shar

e(t

)(0

.29)

∗∗∗

(0.2

9)∗

(0.4

7)∗

(0.3

7)†

Nat

iona

lre

form

edlo

cal

−1.1

8−0

.78

−0.7

6−0

.71

firm

shar

e(t

)(0

.16)

∗∗∗

(0.1

7)∗∗

∗(0

.27)

∗∗(0

.22)

∗∗

Nat

iona

lco

nven

tiona

llo

cal

firm

shar

e(t

)0

00

0

Em

ploy

men

tsh

are

Reg

iona

lno

n-H

MT

−0.3

2−0

.23

−0.1

5−0

.26

offir

ms

info

reig

nfir

msh

are(

t)(0

.08)

∗∗∗

(0.0

8)∗∗

(0.1

1)(0

.12)

∗

regi

onal

mar

kets

Reg

iona

lH

MT

firm

−0.6

8−0

.61

−0.4

8−0

.68

shar

e(t

)(0

.08)

∗∗∗

(0.0

8)∗∗

∗(0

.14)

∗∗(0

.10)

∗∗∗

Reg

iona

lre

form

edlo

cal

−0.5

5−0

.48

−0.2

1−0

.65

firm

shar

e(t

)(0

.07)

∗∗∗

(0.0

7)∗∗

∗(0

.11)

†(0

.10)

∗∗∗

Reg

iona

lco

nven

tiona

llo

cal

firm

shar

e(t

)0

00

0

Firm

Firm

size

(t)

0.69

0.70

0.70

0.71

0.70

char

acte

rist

ics

(0.0

1)∗∗

∗(0

.01)

∗∗∗

(0.0

1)∗∗

∗(0

.01)

∗∗∗

(0.0

1)∗∗

∗

Firm

age(

t)−0

.01

−0.0

1−0

.01

−0.0

1−0

.01

(0.1

3×

10−2

)∗∗∗

(0.1

3×

10−2

)∗∗∗

(0.1

3×

10−2

)∗∗∗

(0.2

0×

10−2

)∗∗∗

(0.1

7×

10−2

)∗∗∗



Tabl

e3.

(Con

tinu

ed)

Var

iabl

eA

llfo

reig

nfir

ms

Non

-HM

Tfo

reig

nfir

ms

HM

Tfo

reig

nfir

ms

(1)

(2)

(3)

(4)

(5)

Firm

profi

tabi

lity

(t)

0.01

0.01

0.01

0.01

0.01

(0.7

3×

10−3

)∗∗∗

(0.7

3×

10−3

)∗∗∗

(0.7

3×

10−3

)∗∗∗

(0.1

1×

10−2

)∗∗∗

(0.1

0×

10−2

)∗∗∗

Fore

ign

JV&

acqu

isiti

on−0

.13

−0.1

3−0

.13

−0.1

7−0

.09

(vs.

gree

nfiel

din

vest

men

t)(0

.02)

∗∗∗

(0.0

2)∗∗

∗(0

.02)

∗∗∗

(0.0

3)∗∗

∗(0

.02)

∗∗∗

HK

,M

acao

,Ta

iwan

ese

−0.1

7−0

.16

−0.1

6fir

ms

(vs.

othe

rfo

reig

nfir

ms)

(0.0

2)∗∗

∗(0

.02)

∗∗∗

(0.0

2)∗∗

∗

Inte

rcep

t11

.76

11.5

311

.91

−4.6

511

.60

(1.0

6)∗∗

∗(1

.08)

∗∗∗

(1.0

5)∗∗

∗(1

.03)

∗∗∗

(0.3

0)∗∗

∗

N24

1,06

924

1,06

924

1,06

910

6,96

113

4,02

4W

ald

chi-

sq(d

.f.)

9,03

8.7

9,04

6.8

9,09

7.0

5561

.171

29.3

(229

)∗∗∗

(229

)∗∗∗

(232

)∗∗∗

(224

)∗∗∗

(224

)∗∗∗

Not

e;∗∗

∗:

p<

0.00

1,∗∗

:p

<0.

01,

∗:

p<

0.05

,†

:p

<0.

10.

182

indu

stry

dum

mie

s,30

regi

ondu

mm

ies,

and

6ye

ardu

mm

ies

are

not

show

n.N

umbe

rsin

pare

nthe

ses

are

robu

stst

anda

rder

rors

adju

sted

for

clus

teri

ngon

indi

vidu

alfir

ms.

Inca

seal

lfir

ms

inou

rsa

mpl

eex

itan

indu

stry

ina

year

,th

ein

dust

rydu

mm

ype

rfec

tlyex

plai

nsth

eex

itlik

elih

ood

and

ther

efor

eth

ose

obse

rvat

ions

are

drop

ped

inth

ees

timat

ion.

Whe

nw

ebr

eak

upsa

mpl

esto

HM

Tan

dno

n-H

MT

fore

ign

firm

s,w

ete

ndto

find

mor

esu

chca

ses.

Thu

s,th

enu

mbe

rof

obse

rvat

ions

inM

odel

s4

and

5do

not

add

upto

that

ofa

mod

elus

ing

all

fore

ign

firm

s.



Chinese regional markets. When we include theemployment shares for both national and regionalmarkets in Model 3, Hypothesis 1—which arguesthat competition effects are more likely to out-weigh spillover effects in regional markets than innational markets—is not supported in the foreignfirm sample, because foreign firms face intensecompetition in both national and regional markets.This result suggests that in China, foreign firms donot benefit from other firms’ spillovers. Most for-eign entrants in China might rely on the proprietaryassets of their own multinational operations. What-ever knowledge spillovers there are from othercompanies in China might be completely offset bycompetition from these same companies.

We test Hypothesis 2 in the foreign firm sampleby comparing the coefficients of reformed localfirms and those of conventional local firms. Thenegative signs for national reformed local firmshare and regional reformed local firm share indi-cate that, compared to conventional local firms,whose coefficients are set to zero, reformed localfirms impose stronger competition effects on for-eign entrants in both the national and regionalmarkets. The negative signs of non-HMT foreignfirm and HMT firm groups at both the nationaland regional levels in Model 3 are also consistentwith Hypothesis 2, suggesting there is intense com-petition among foreign firms because these firms’resource profiles are similar. Overall, foreign firmsappear more likely to be crowded out by HMTfirms and by reformed local firms that emulate theirresources profiles than they are by conventionallocals that rely on completely different resourceendowments. Hypothesis 2 is thus supported in theforeign firm sample.

We perform separate analyses for non-HMTforeign firms and HMT firms, in Models 4 and 5of Table 3. Both non-HMT foreign firms and HMTfirms show somewhat similar patterns, althoughthe significance levels for negative competitioneffects at the regional level seem stronger for HMTfirms than they do for non-HMT firms.

Our various firm-level controls suggest that thelarger, the younger, and the more profitable foreignfirms are, the more likely they are to survive.Foreign firms that were formed by joint ventures oracquisitions have lower survival rates than foreigngreenfield investments do. Between the two typesof foreign firms, HMT firms have lower survivalrates than non-HMT foreign firms do.

In order to check the robustness of our results,we implemented a longer time lag for measur-ing the presence of foreign and local firms indetermining their likelihood of survival. A longertime lag can make spillover effects more pro-nounced and negatively bias competition effects,because more recent entries exert immediate com-petition effects, and spillover effects may takelonger to materialize. To illustrate this potentialbias with a longer time lag, we ran separateregressions in Tables 4 and 5 with two sets ofvariables: employment shares measured at timet − 1 (two-year lag), which reflects the accumu-lated entries (presence) of different types of firmsup until time t − 1; and the change of employ-ment shares from time t − 1 to time t, whichreflects changes (increase/decrease) in employmentshares of different types of firm from time t − 1 totime t. Note that there were fewer observationsin the models in Tables 4 and 5 because we lostobservations for one year for each additional timelag. In general, employment shares at time t − 1tend to be positive and significant, showing morespillover effects than they do competition effects.For instance, the regional employment share vari-ables of HMT firms and reformed local firms inlocal firm regressions (see Model 2 of Table 4)suggest positive spillover effects, contrary to theresults based on a one-year lag (see Table 2). Onthe other hand, the changes of employment sharesfrom time t − 1 to time t tend to be negative.For instance, the change of HMT firms’ share ina national market is strongly negative, and thechanges in HMT firms’ share and reformed localfirms’ share in a regional market are negative andsignificant, suggesting that the increased presenceof HMT firms and reformed local firms from timet − 1 to time t exhibits strong competitive pres-sures.

Similarly, the models for the foreign firmssample in Table 5, using a two-year lag, showagain that the employment shares with a two-year lag tend to soften the negative competitioneffects—the regional non-HMT foreign firms andregional reformed local firms shares, that werenegative and significant in the one-year lag modelin Table 3, were insignificant. The changes of theemployment shares from time t − 1 to time t reflectnegative competition effects. To sum up, a longertime lag tends to create upward bias for spillovereffects and downward bias for competition effects.



Table 4. Survival of local firms at time t + 1 during 1998–2005 with a two-year lag

Variable (1) (2) (3)

Employment share of firms National non-HMT foreign 0.50 0.45in National markets at firm share(t − 1) (0.23)∗ (0.24)†time t − 1 National HMT firm share 0.44 0.34

(t − 1) (0.24)∗ (0.20)†National reformed local firm 0.86 0.76

share (t − 1) (0.09)∗∗∗ (0.09)∗∗∗

National conventional localFirm share (t − 1)

0 0

Employment share of firms Regional non-HMT foreign 0.01 −0.001in Regional markets at firm share(t − 1) (0.04) (0.04)time t − 1 Regional HMT firm share 0.14 0.12

(t − 1) (0.04)∗∗ (0.04)∗∗

Regional reformed local firm 0.17 0.13share (t − 1) (0.02)∗∗∗ (0.02)∗∗∗

Regional conventional localfirm share (t − 1)

0 0

Change of employment National non-HMT foreign 0.15 0.09share of firms in national firm share(�t) (0.20) (0.20)markets between time National HMT firm share (�t) −0.97 −0.93t − 1 and t (0.18)∗∗∗ (0.18)∗∗∗

National reformed local firm 0.14 0.16share (�t) (0.09) (0.09)†

National conventional localfirm share (�t)

0 0

Change of employment Regional non-HMT foreign −0.01 −0.01share of firms in regional firm share(�t) (0.06) (0.06)markets between time Regional HMT firm share (�t) −0.17 −0.10t − 1 and t (0.06)∗∗ (0.07)

Regional reformed local firm −0.08 −0.09share (�t) (0.02)∗∗ (0.03)∗∗

Regional conventional localfirm share (�t)

0 0

Firm characteristics Firm size(t − 1) 0.30 0.30 0.30(0.25 × 10−2)∗∗∗ (0.25 × 10−2)∗∗∗ (0.25 × 10−2)∗∗∗

Firm age(t − 1) 0.01 0.01 0.01(0.26 × 10−3)∗∗∗ (0.26 × 10−3)∗∗∗ (0.26 × 10−3)∗∗∗

Firm profitability (t − 1) 0.61 × 10−2 0.61 × 10−2 0.61 × 10−2

(0.18 × 10−3)∗∗∗ (0.18 × 10−3)∗∗∗ (0.18 × 10−3)∗∗∗

Level of control 0.04 0.04 0.04(0.19 × 10−2) (0.19 × 10−2) (0.19 × 10−2)

Reformed local firms 0.33 0.32 0.32(vs. conventional locals) (0.01)∗∗∗ (0.01)∗∗∗ (0.01)∗∗∗

Intercept −2.00 −1.65 −1.95(0.26)∗∗∗ (0.03)∗∗∗ (0.05)∗∗∗

N 899,947 899,947 899,947Wald chi-sq (d.f.) 27574.2 37305.2 37391.8

(244)∗∗∗ (244)∗∗∗ (250)∗∗∗

Note; ∗∗∗ : p < 0.001, ∗∗ : p < 0.01, ∗ : p < 0.05, † : p < 0.10. 196 industry dummies, 30 region dummies, and 6 year dummies arenot shown. Numbers in parentheses are robust standard errors adjusted for clustering on individual firms.

DISCUSSION AND CONCLUSION

This study combines the FDI spillover literatureand a competitor analysis framework to examine

spillovers and competition among foreign entrantsand local firms. Our findings suggest that bothspillover and competition effects from variousgroups of firms affect firms in other groups in



Table 5. Survival of foreign firms at time t + 1 during 1998–2005 with a two-year lag

Variable (1) (2) (3)

Employment share of firms National non-HMT foreign −1.99 −2.21in National markets at firm share(t − 1) (0.50)∗∗∗ (0.47)∗∗∗

time t − 1 National HMT firm share −1.09 −1.09(t − 1) (0.40)∗∗ (0.40)∗∗

National reformed local firm −0.62 −0.60share (t − 1) (0.21)∗∗ (0.23)∗∗

National conventional localFirm share (t − 1)

0 0

Employment share of firms Regional non-HMT foreign 0.12 0.22in Regional markets at firm share(t − 1) (0.09) (0.09)∗

time t − 1 Regional HMT firm share −0.15 −0.07(t − 1) (0.09)† (0.09)

Regional reformed local firm −0.06 −0.03 ×10−2

share (t − 1) (0.08) (0.08)Regional conventional local

firm share (t − 1)0 0

Change of employment National non-HMT foreign −2.04 −2.17share of firms in national firm share(�t) (0.42)∗∗∗ (0.44)∗∗∗

markets between time National HMT firm share (�t) −1.86 −1.85t − 1 and t (0.36)∗∗∗ (0.38)∗∗∗

National reformed local firm −0.83 −0.86share (�t) (0.25)∗∗∗ (0.26)∗∗∗

National conventional localFirm share (�t)

0 0

Change of employment Regional non-HMT foreign −0.08 0.07share of firms in regional firm share(�t) (0.13) (0.13)markets between time Regional HMT firm share (�t) −0.21 −0.06t − 1 and t (0.13)† (0.13)

Regional reformed local firm −0.02 0.07share (�t) (0.10) (0.10)

Regional conventional localfirm share (�t)

0 0

Firm characteristics Firm size(t − 1) 0.44 0.44 0.44(0.01)∗∗∗ (0.01)∗∗∗ (0.01)∗∗∗

Firm age(t − 1) −0.01 −0.01 −0.01(0.12 × 10−2)∗∗∗ (0.12 × 10−2)∗∗∗ (0.12 × 10−2)∗∗∗

Firm profitability (t − 1) 0.02 0.02 0.02(0.52 × 10−3)∗∗∗ (0.52 × 10−3)∗∗∗ (0.52 × 10−3)∗∗∗

Foreign JV & acquisition (vs. −0.20 −0.20 −0.20greenfield investment) (0.02)∗∗∗ (0.02)∗∗∗ (0.02)∗∗∗

HK, Macao, Taiwanese firms −0.19 −0.18 −0.18(vs. foreign firms) (0.02)∗∗∗ (0.02)∗∗∗ (0.02)∗∗∗

Intercept −0.93 −1.48 −0.96(4.86) (5.05)† (5.05)

N 215,251 215,251 215,251Wald chi-sq (d.f.) 8286.9 8266.1 8306.6

(232)∗∗∗ (232)∗∗∗ (238)∗∗∗

Note; ∗∗∗ : p < 0.001, ∗∗ : p < 0.01, ∗ : p < 0.05, † : p < 0.10. 182 industry dummies, 30 region dummies, and 6 year dummies arenot shown. Numbers in parentheses are robust standard errors adjusted for clustering on individual firms.

China. They also indicate that competition effectsare more likely to outweigh spillover effects inregional markets than they are in national mar-kets. Further, competition effects are more likely

to outweigh spillover effects among firms of sim-ilar resource types than they are among firms ofdistinct resource profiles. This study has severalimportant theoretical contributions.



First, this study contributes to the FDI spilloverliterature. In contrast to prior work that has gener-ally assumed spillovers go only from foreign firmsto local firms, we model spillovers as occurringbetween foreign firms and local firms, among for-eign firms, and among local firms. Further, unlikeresearch that has focused only on the conditionsthat contribute to spillovers, we introduce compe-tition effects by incorporating a framework that isbuilt on the market and resource configurations ofrelevant competitors. Thus, we have added a strate-gic dimension to the spillover literature and gener-ated new insights into the net effect of spilloversand competition among firms. By doing so, thisstudy recasts the existing spillover literature into amore general framework that accommodates bothpositive and negative effects among different typesof foreign and local firms.

Second, this study also contributes to competi-tor analysis research in several ways. Although inthe past this line of inquiry has done an excel-lent job in examining rivalry between firms, it hasfocused only on competition and largely ignoredthe positive side of interfirm relations, such asindustry clustering, network externalities, and pos-sible spillover effects (Porter, 1998; Chung andAlcacer, 2002; Chang and Park, 2005). By com-bining this research with the FDI literature, thisstudy helps bring a more balanced view in studyinginterfirm effects. Moreover, the empirical contextof the competitor analysis research has often beena single domestic industry (Baum and Korn, 1996,1999; Chen et al., 2007). In contrast, this studyconsiders multiple industries and an internationalsetting, and it therefore widens the application ofthe competitor analysis framework. This study alsoexplicitly considers the rivalry between multina-tional firms and local firms, which has not yetbeen studied in extant research. Another contri-bution to the competitor analysis literature is thereconceptualization of market commonality, whichhas been narrowly defined in terms of multi-marketcontact between large firms such as airlines andautomakers (Smith et al., 1991; Gimeno, 1999; Yuand Cannella, 2007). By applying this concept interms of firms that compete in the same regionalmarkets, this study has improved its applicability.

Third, this study highlights a major opportunityfor strategy research. We conceptualize local firmsnot just as recipients of technology spillovers,or merely as subject to competition; instead, wesee them as competent learning organizations with

heterogeneous resource profiles. We have thus shedsome light on how local firms have begun tocompete against powerful multinational firms. Inparticular, we show that the presence of reformedlocal firms has had a strong competition effect onall foreign firms. In contrast, conventional localfirms do not exhibit such an effect, because theyrely on exclusive networks or policy benefits, anddo not compete directly with foreign entrants. Thisgeneral finding should direct future research toexplore further the strategies and performance ofdifferent groups of local companies that competewith multinational firms.

Finally, this study helps resolve some conflict-ing findings from extant research. It is challeng-ing to separate spillover effects from competitioneffects empirically, but our theoretical frameworkand research design identified contexts in whichone effect might be stronger or weaker than theother. Foreign firms’ employment share in nationalmarkets have positive spillover effects on localfirm survival, confirming previous research thatshowed a similar positive net effect at the countrylevel (Buckley et al., 2002; Caves, 1974; Driffieldand Munday, 2000; Liu et al., 2000; Sinai andMeyer, 2004). At the same time, we find bothforeign and local firms crowd out each other aswell as their own peers. Our results for HMT firmshare in regional markets point to the moderat-ing effect of the geographic market dimension indetermining the relative size of FDI’s spilloverand competition effects, which existing researchhas overlooked. Further, the strong competitioneffect imposed by reformed local firms on for-eign entrants has not been found in earlier studies,especially in the context of emerging markets. Pre-viously, only two FDI studies, both with samplesfrom the United Kingdom, have considered thepossibility of ‘reverse spillovers’ of local industriesonto foreign-owned sectors (Driffield and Mun-day, 2000; Liu et al., 2000). Our findings suggestforeign entrants face a strong liability of foreign-ness despite their superior technological capabili-ties. These findings, although perplexing from theperspective of traditional FDI literature, are con-sistent with our theoretical framework.

This study has several limitations and provokesquestions that merit further research. First, weexamined a context in which the net of spilloverand competition effects may be positive or neg-ative, instead of trying to separate these effects.Future studies should find better measures and



more appropriate empirical contexts that may helpseparate these two effects. Second, our sampleconsisted of relatively large firms by Chinese stan-dards. If small local firms did not have enoughresources and market power to compete againstforeign firms and reformed local firms, compe-tition effects could have been greater if we hadincluded these firms. Third, the degree of spillovereffects might be related to the share of foreignownership in a firm because foreign multination-als are more likely to transfer advanced technol-ogy to their wholly owned subsidiaries than theyare to joint ventures in order to avoid involuntaryspillovers. Fourth, the degree of spillover effectsmight be contingent upon the recipients’ absorp-tive capacity. Future studies should consider howthe degree of spillover varies according to a firm’sownership structure, technological sophistication,and the recipients’ absorptive capacity.

Some of our measures are specific to the empir-ical domain of this study, but our theories andhypotheses are generalizable to other contexts.Although our proxy of market commonality withprovincial versus national markets might be China-specific, we believe the idea of regionalized com-petition and nationwide spillovers is applicable toother countries. It is reasonable to expect, however,that the lower the regional barriers in a country are,the smaller the difference in spillover and com-petition effects will be between the national andregional levels. Similarly, although our classifi-cation of firms is based on our empirical settingand may be somewhat simplistic, it is likely that,regardless of the context, some multinational firmsare more localized than others, and the resourceendowments of some local firms resemble those offoreign firms. Overall, China is not fundamentallydifferent from other emerging markets, except thatits markets are perhaps bigger and more heteroge-neous. Future research can validate our findings inother contexts by defining markets and classifyingresource profiles in other ways.

This study has several practical implications.In contrast to their earlier obsession with inwardFDI and GDP growth, Chinese policy makers andthe public are currently concerned with how FDImight negatively affect domestic companies’ sur-vival (Huang, 2003). Our results depict FDI asa double-edged sword: Chinese firms have ben-efited from, and been negatively affected by, for-eign entrants, and this impact can be delineatedlargely in terms of geographic location and the

resource profiles of both foreign and local firms.The biggest beneficiaries of spillover effects mightbe local firms located in inland areas, which canbenefit from spillovers from both foreign multina-tionals and reformed local firms that compete witheach other in the coastal regions, while avoidingdirect competition with either group. In the mean-time, host-country governments may improve localfirms’ survival chances by restructuring them insuch a way that they can develop their own dis-tinctive resources. Governments can also facilitatespillovers among local firms by encouraging merg-ers and acquisitions across regions. On the otherhand, foreign firms’ survival is negatively affectedby competition from both other multinationals andstrong local competitors within the same region, aswell as from the rest of the nation. This result issimilar to Shaver and Flyer’s (2000) finding thatfirms with strong competitive advantages avoidagglomeration because they would stand to losemore than they would gain. Thus, multinationalfirms need to choose their competitors and partnersmore carefully. Because most Chinese firms, espe-cially private firms, shareholding firms, and lim-ited liability firms, can compete vigorously againstforeign entrants and crowd them out of the domes-tic market, it might be risky for multinationals toallow further spillovers of advanced technologiesto these reformed local firms.

Our approach is useful for evaluating the rela-tive performance of foreign and local firms. Ourfindings suggest that emerging markets are imper-fectly competitive, and strong local firms eventu-ally emerge. The more heterogeneous local mar-kets and local resources are, the more chances localcontenders have to survive and succeed. Over-all, this study enhances our understanding of howlocal firms arise and compete with multinationalsin emerging markets.

ACKNOWLEDGEMENTS

We appreciate helpful comments and suggestionsfrom two anonymous reviewers; Theresa Cho;Renata Kosova; John Lafkas; and seminar par-ticipants at Peking University, Korea University,INSEAD, the Wharton School, and Harvard Busi-ness School. We thank Yuping Zeng for her excel-lent research assistance. This paper was largelycompleted while Dean Xu was at Guanghua Schoolof Management, Peking University. Sea Jin Chang



appreciates the research grant from the KoreaResearch Foundation, KRF-2007-342-B00016.

REFERENCES

Ahn N, de la Rica S, Ugidos A. 1999. Willingness tomove for work and unemployment duration in Spain.Economica 66: 335–357.

Aitken BJ, Harrison AE. 1999. Do domestic firmsbenefit from direct foreign investment? Evidence fromVenezuela. The American Economic Review 89(3):605–618.

Allison PD. 1984. Event History Analysis: Regression forLongitudinal Event Data . Sage: Beverly Hills, CA.

Bain J. 1956. Barriers to New Competition . HarvardUniversity Press: Cambridge, MA.

Barney JB. 1986. Strategic factor markets: expectations,luck, and business strategy. Management Science 32:1230–1241.

Barney JB, Hoskisson RE. 1990. Strategic groups:untested assertions and research proposals. Manage-rial and Decision Economics 11: 187–198.

Baum JAC, Korn HJ. 1996. Competitive dynamics ofinterfirm rivalry. Academy of Management Journal39(2): 255–291.

Baum JAC, Korn HJ. 1999. Dynamics of dyadic compet-itive interaction. Strategic Management Journal 20(3):251–278.

Baum JAC, Mezias SJ. 1992. Localized competition andorganizational failure in the Manhattan hotel industry:1898–1990. Administrative Science Quarterly 37:580–604.

Beamish PW. 1988. Multinational Joint Ventures inDeveloping Countries . Routledge: London, UK.

Blomstrom M. 1986. Foreign investment and productiveefficiency: the case of Mexico. The Journal ofIndustrial Economics 35(1): 97–110.

Blomstrom M, Persson H. 1983. Foreign investment andspillover efficiency in an underdeveloped economy:evidence from the Mexican manufacturing industry.World Development 11: 493–501.

Buckley PJ, Casson MC. 1976. The Future of theMultinational Enterprise. Macmillan Press: London,UK.

Buckley PJ, Clegg J, Wang CQ. 2002. The impactof inward FDI on the performance of Chinesemanufacturing firms. Journal of International BusinessStudies 33(4): 637–655.

Buckley PJ, Ghauri PN. 2004. Globalisation, economicgeography, and the strategy of multinational enter-prises. Journal of International Business Studies 35(2):81–98.

Cantwell J. 1989. Technical Innovation and MultinationalCorporations . Basil Blackwell: Oxford, UK.

Caves RE. 1971. International corporations: the industrialeconomics of foreign investment. Economica 38:1–27.

Caves RE. 1974. Multinational firms, competition, andproductivity in host-country markets. Economica 41:176–193.

Chang SJ. 1995. International expansion strategy ofJapanese firms: capability building through sequentialentry. Academy of Management Journal 38(2):383–407.

Chang SJ. 1996. An evolutionary perspective ondiversification and corporate restructuring: entry, exitand economic performance during 1981–89. StrategicManagement Journal 17(8): 587–611.

Chang SJ, Park S. 2005. Types of firms generating net-work externalities and MNCs’ co-location decisions.Strategic Management Journal 26(7): 595–615.

Chen MJ. 1996. Competitor analysis and interfirmrivalry: toward a theoretical integration. Academy ofManagement Review 21(1): 100–134.

Chen MJ, MacMillan IC. 1992. Nonresponse and delayedresponse to competitive moves: the roles of competitordependence and action irreversibility. Academy ofManagement Journal 35(3): 539–570.

Chen MJ, Su KH, Tsai W. 2007. Competitive ten-sion: the awareness-motivation-capability perspective.Academy of Management Journal 50(1): 101–118.

Chow GC. 1993. Capital formation and economic growthin China. Quarterly Journal of Economics 108(3):809–842.

Chung W, Alcacer J. 2002. Knowledge seeking andlocation choice of foreign direct investment inthe United States. Management Science 48(12):1534–1554.

Collis DJ. 1991. A resource-based analysis of globalcompetition: the case of the bearings industry.Strategic Management Journal , Summer Special Issue12: 49–68.

Conner KR. 1994. The resource-based challenge to theindustry-structure perspective In Best Paper Proceed-ings of the 54th Annual Academy of Management Con-ference, Moore D (ed).: 17–21.

Dawar N, Frost T. 1999. Competing with giants: survivalstrategies for local companies in emerging markets.Harvard Business Review 77(2): 119–129.

De Backer K, Sleuwaegen L. 2003. Does foreign directinvestment crowd out domestic entrepreneurship?Review of Industrial Organization 22(1): 67–84.

DiMaggio PJ, Powell WW. 1983. The iron cage revisited:institutional isomorphism and collective rationality inorganizational fields. American Sociological Review48: 147–160.

Driffield N, Munday M. 2000. Industrial performance,agglomeration, and foreign manufacturing investmentin the UK. Journal of International Business Studies31(1): 21–37.

Duhaime IM, Grant JM. 1984. Factors influencingdivestment decision-making: evidence from a fieldstudy. Strategic Management Journal 5(4): 301–318.

Dunning JH. 1981. International Production and theMultinational Enterprise. Allen and Unwin: London,UK.

Dunning JH. 1988. The eclectic paradigm of internationalproduction: a restatement and some possible exten-sions. Journal of International Business Studies 19(1):1–31.

Ferrier WJ. 2001. Navigating the competitive land-scape: the drivers and consequences of competitive



aggressiveness. Academy of Management Journal 44:858–877.

Gimeno J. 1999. Reciprocal threats in multimarketrivalry: staking out ‘spheres of influence’ in theU.S. airlines industry. Strategic Management Journal20(2): 101–128.

Gimeno J, Woo CY. 1994. Multipoint competition,market rivalry and firm performance: a test of thecomplete mediation model. In Best Paper Proceedingsof the 54th Annual Academy of ManagementConference, Moore D (ed).: 32–36.

Gerschenkron A. 1962. Economic Backwardness inHistorical Perspective: A Book of Essays . BelknapPress of Harvard University Press: Cambridge, MA.

Ghoshal S, Westney ED. 1993. Introduction and over-view. In Organization Theory and the MultinationalCorporation , Ghoshal S, Westney ED (eds). StMartin’s Press: New York; 1–23.

Globerman S. 1979. Foreign direct investment and‘spillover’ efficiency benefits in Canadian manufac-turing industries. The Canadian Journal of Economics12(1): 42–56.

Haddad M, Harrison A. 1993. Are there positivespillovers from direct foreign investment? Evidencefrom panel data for Morocco. Journal of DevelopmentEconomics 42(1): 51–74.

Hejazi W, Safarian AE. 1999. Trade, foreign directinvestment, and R&D spillovers. Journal of Interna-tional Business Studies 30(3): 491–511.

Henisz WJ, Delios A. 2001. Uncertainty, imitation, andplant location: Japanese multinational corporations,1990–1996. Administrative Science Quarterly 46(3):443–475.

Hitt MA, Dacin MT, Levitas E, Arregle JL, Borza A.2000. Partner selection in emerging and developedmarket contexts: resource-based and organizationallearning perspectives. Academy of ManagementJournal 43(3): 449–467.

Huang Y. 2003. Selling China: Foreign Direct InvestmentDuring the Reform Era . Cambridge University Press:Cambridge, UK.

Hymer SH. 1960. The international operation of nationalfirms: a study of direct foreign investment. Ph.D. diss.Massachusetts Institute of Technology, Boston, MA.

Javorcik BS. 2004. Does foreign direct investmentincrease the productivity of domestic firms? Insearch of spillovers through backward linkages. TheAmerican Economic Review 94(3): 605–627.

Jin HH, Qian YY, Weingast BR. 2005. Regional decen-tralization and fiscal incentives: Federalism, Chi-nese style. Journal of Public Economics 89(9–10):1719–1742.

Keller W. 2002. Geographical localization and interna-tional technology diffusion. The American EconomicReview 92(1): 120–142.

Khanna T, Palepu KG, Sinha J, Klump A, Kaji N,Sanchez L,Yacoub, M. 2005. Strategies that fitemerging markets. Harvard Business Review 83(6):63–76.

Kim WC. 1988. The effects of competition and corporatepolitical responsiveness on multinational bargainingpower. Strategic Management Journal 9(3): 289–295.

Klein KJ, Dansereau F, Hall RJ. 1994. Levels issues intheory development, data collection, and analysis.Academy of Management Review 19(2): 195–229.

Kogut B. 1983. Foreign direct investment as a sequentialprocess. In The Multinational Corporation in the1980s , Kindleberger C, Audretsch D (eds). MITPress: Cambridge, MA; 38–56.

Kokko A. 1994. Technology, market characteristics, andspillovers. Journal of Development Economics 43:279–293.

Kokko A, Tasini R, Zejan M. 1996. Local technologicalcapability and productivity spillover from FDI inthe Uruguayan manufacturing sector, Journal ofDevelopment Studies 32: 602–611.

Konings J. 2001. The effects of foreign direct investmenton domestic firms: evidence from firm-level panel datain emerging economics. Economics of Transition 9(3):619–633.

Kosova R. 2005. Do foreign firms crowd out domesticfirms? Evidence from the Czech Republic. Workingpaper, George Washington University, Washington,D.C.

Kostova T, Zaheer S. 1999. Organizational legitimacyunder conditions of complexity: the case of themultinational enterprise. Academy of ManagementReview 24(1): 64–81.

Krugman P. 1991. Increasing returns and economic geog-raphy. Journal of Political Economy 99: 483–499.

Li J, Yang JY, Yue DR. 2007. Identity, community, andaudience: how wholly owned foreign subsidiaries gainlegitimacy in China. Academy of Management Journal50(1): 175–190.

Lieberman MB, Montgomery DB. 1988. First-moveradvantages. Strategic Management Journal , SummerSpecial Issue 9: 41–58.

Liu XM, Siler P, Wang CQ, Wei YQ. 2000. Productivityspillovers from foreign direct investment: evidencefrom UK industrial level panel data. Journal ofInternational Business Studies 31(3): 407–425.

Lu JW, Xu D. 2006. Growth and survival of internationaljoint ventures: an external-internal legitimacy perspec-tive. Journal of Management 32(3): 427–448.

Luo Y. 1997. Partner selection and venturing success:the case of joint ventures with firms in thePeople’s Republic of China. Organization Science8(6): 648–662.

Luo Y. 2005. Selling China: foreign direct investmentduring the reform era. (Book Review). Academy ofManagement Review 30(1): 200–203.

Makino S, Delios A. 1996. Local knowledge transfer andperformance: implications for alliance formation inAsia. Journal of International Business Studies 27(5):905–927.

McGee J, Thomas H. 1986. Strategic groups: theory,research and taxonomy. Strategic ManagementJournal 7(2): 141–160.

Meyer KE. 2004. Perspectives on multinational enter-prises in emerging economies. Journal of InternationalBusiness Studies 35: 259–276.

Miller SR, Eden L. 2006. Local density and foreignsubsidiary performance. Academy of ManagementJournal 49(2): 341–355.



Moran TH. 1985. Multinational Corporations: ThePolitical Economy of Foreign Direct Investment .Lexington Books: Lexington, MA.

Oliver C. 1991. Strategic responses to institutionalprocesses. Academy of Management Review 16(1):145–179.

Pan Y, Li S, Tse DK. 1999. The impact of order andmode of market entry on profitability and marketshare. Journal of International Business Studies 30(1):81–103.

Park SH, Li S, Tse DK. 2006. Market liberalizationand firm performance during China’s economictransition. Journal of International Business Studies37: 127–147.

Peng MW, Luo YD. 2000. Managerial ties and firmperformance in a transition economy: the nature ofa micro-macro link. Academy of Management Journal43(3): 486–501.

Peng MW, Tan J, Tong TW. 2004. Ownership types andstrategic groups in an emerging economy. Journal ofManagement Studies 41(7): 1105–1129.

Porter ME. 1980. Competitive Strategy: Techniques forAnalyzing Industries and Competitors . Free Press:New York.

Porter ME. 1998. Clustering and the new economics ofcompetition. Harvard Business Review 76(6): 77–90.

Prahalad CK, Lieberthal K. 1998. The end of corporateimperialism. Harvard Business Review 76(4): 68–79.

Ralston DA, Terpstra-Tong J, Terpstra RH, Wang X,Egri C. 2006. Today’s state-owned enterprises ofChina: are they dying dinosaurs or dynamic dynamos?Strategic Management Journal 27(9): 825–843.

Rosenzweig PM, Singh JV. 1991. Organizational envi-ronments and the multinational enterprise. Academyof Management Review 16: 340–361.

Saxenian A. 1994. Regional Advantage. Harvard Univer-sity Press: Cambridge, MA.

Sharma A, Kesner IF. 1996. Diversifying entry: some exante explanations for postentry survival and growth.Academy of Management Journal 39(3): 635–677.

Shaver JM, Flyer F. 2000. Agglomeration economies,firm heterogeneity, and foreign direct investment in theUnited States. Strategic Management Journal 21(12):1175–1193.

Shaver JM, Mitchell W, Yeung B. 1997. The effect ofown-firm and other-firm experience on foreign directinvestment survival in the United States, 1987–92.Strategic Management Journal 18(10): 811–824.

Sinai E, Meyer KE. 2004. Identifying spillovers oftechnology transfer: the case of Estonia. Journal ofComparative Economics 32(3): 445–466.

Smith KG, Grimm CM, Gannon MJ, Chen MJ. 1991.Organizational information processing, competitiveresponses, and performance in the U.S. domesticairline industry. Academy of Management Journal 34:1–26.

Teece DJ, Pisano G, Shuen A. 1997. Dynamic capabili-ties and strategic management. Strategic ManagementJournal 18(7): 509–533.

Vanhonacker W. 1997. Entering China: an unconven-tional approach. Harvard Business Review 75(2):130–136.

Yu T, Cannella AA. 2007. Rivalry between multinationalenterprises: an event history approach. Academy ofManagement Journal 50(3): 663–684.

Zeng M, Williamson PJ. 2003. The hidden dragons.Harvard Business Review 81(10): 92–99.


SPILLOVERS AND COMPETITION AMONG FOREIGN AND LOCAL … smj2008.pdf · In this study, we examine...

Documents

Transcript of SPILLOVERS AND COMPETITION AMONG FOREIGN AND LOCAL … smj2008.pdf · In this study, we examine...