The impact of entry mode choice on foreign affiliate performance: the case of foreign MNEs in South Korea.
* This study examines the performance impacts of entry mode choice based on the perceptions of managers of a large sample of foreign MNEs in South Korea. Using an extended transaction costs model, an evaluation of performance relative to predicted entry mode is carried out.
* We find that affiliate performance is a multidimensional and complex phenomenon, which may be properly explained by multiple factors, including the coordination and control of the affiliate by the MNE, that go beyond initial entry mode choices.
* Foreign MNEs following the entry mode predicted by the extended transaction costs model showed poorer performance than non-followers in respect of non-financial performance. However, no significant differences were found for financial performance. The findings appear to reinforce a recent call to re-examine transaction costs theory and lend further support for incorporating additional potentially important determinants of affiliate performance.
Entry Mode Choice * Foreign Affiliate Performance * Transaction Costs Theory * Cultural Distance * Institutional Factors * Multinational Enterprises * South Korea
International entry mode selection has been one of the most extensively researched topics in international business. Substantial prior research has been undertaken in the field to explain why firms select a particular mode of entry into international markets. While developed home and host countries were almost an exclusive domain of interest in earlier entry mode studies (e.g., Gatignon/Anderson 1988, Kogut/Singh 1988, Hennart 1991, Kim/Hwang 1992), more recently transition host economies, with their growing economic significance, have attracted increasing attention from researchers in the field (e.g., Tse/Pan/Au 1997, Brouthers/Brouthers 2001, Chen/Hu 2002). The purpose of this study is to build on existing research by examining the extent to which the foreign affiliates of MNEs in South Korea (Korea hereafter) make theoretically optimal entry mode decisions and whether or not this entry mode choice is related to enhanced performance compared to alternatives.
A variety of entry mode choices are available to firms expanding internationally, ranging from exporting, contractual arrangements (e.g., licensing), through to FDI (either joint venture or wholly owned subsidiary). This variety of choices is well reflected in entry mode studies that are broadly classified into three main categories depending on their focus of interest, i.e., whether studies examine: i) the choice between equity and non-equity modes (e.g., Buckley/Casson 1976, Rugman 1981, cited in Rugman/Verbeke 2003); ii) the choice between shared and full control modes (e.g., Hennart 1982, Padmanabhan/Cho 1996, Brouthers/Brouthers 2001); or iii) the choice between greenfield operations and acquisitions (e.g., Kogut/Singh 1988, Chang/Rosenzweig 2001, Harzing 2002). The difference in focus often makes a simple or outright comparison of the results not only difficult and complex but also problematic (Kogut/Singh 1988).
Entry mode studies have also been extended to explore the performance implications of entry mode choice (Root 1987, Woodcock et al. 1994, Li 1995). The underlying position of these studies is that entry mode choice for the international affiliate has a critical impact on its performance as the parent strategically selects a mode that generates better performance than alternatives. However, contrary to extensive research found in the area of international entry mode selection, there has been a relative dearth of studies linking entry mode to performance (Woodcock et al. 1994, Brouthers et al. 1999). In addition, the majority of extant performance studies involving entry mode choice do not go beyond simple comparisons of performance between alternative modes of entry, albeit with some recent exceptions (e.g., Shaver 1998, Brouthers et al. 1999, Brouthers 2002, Chen/Hu 2002) that attempt to account for 'endogenous' entry mode selection.
It has been suggested that researchers should explore more frequently the characteristics of each investment decision to better understand the link between entry mode and performance (Shaver 1998). More importantly, it appears that more research effort needs to be directed at examining if theoretically-optimal entry mode decisions lead to superior performance compared to alternatives. Research in this direction not only enhances our understanding of MNE entry mode behaviour but also provides useful insights for both researchers and managers by evaluating the normative power of existing theories and by showing how managers can make entry decisions more likely to generate success internationally.
Our research objectives in this study are two-fold. First, we attempt to examine whether firms follow theoretical predictions in making their entry mode decision, i.e., the choice between full and shared control modes of entry. We then explore whether theoretically-optimal entry mode choice generates better performance than that of entry mode selected otherwise. Following the suggestions of recent research (e.g., Delios/Beamish 1999, Brouthers 2002, Chen/Hu 2002), the present study applies a theoretical framework that extends traditional transaction costs theory to explain the entry mode behaviour of foreign firms in Korea and its influence on their Korean affiliate performance. Similar to Brouthers and Brouthers (2000) and Brouthers (2002), the proposed framework also includes cultural and institutional variables that are addressed in the literature.
We tested our research questions by using primary data collected from a large sample of 228 (further reduced to 222 firms for performance analysis) foreign firms in Korea. Subjective financial and non-financial measures were used to capture the complex and multidimensional phenomenon of performance. Following Brouthers (2002), we conducted a two-stage statistical analysis; first, logistic regression analysis to categorize the sample into two groups, i.e., fit and no-fit groups and then multiple regression to examine the linkage between entry mode fit and affiliate performance, while controlling for various firm, industry and country factors.
The structure of this paper is as follows. The next section reviews the literature on entry mode choice and performance, from which hypotheses are developed. We then address research methodology, followed by a discussion of regression results. Finally, the paper concludes with a note on the limitations of the study and the implications of the findings for future research.
Literature Review and Hypotheses
Transaction costs theory has been frequently applied to explain a firm's international entry mode choice in foreign markets. The position of earlier studies, Buckley and Casson (1976) in particular, is that firms would be better off by internalizing their international operations rather than coordinating them through the market mechanism (Coase 1937, Williamson 1991) in the presence of market imperfections. This is because the "transaction costs associated with managing an internal market across borders and the related requirement to decentralize many value-added activities" (cited in Rugman/Verbeke 2003, p. 127) are lower than those of using external markets due to the existence of market imperfections, especially arising from the difficulty of pricing proprietary assets, such as knowledge and ideas (Buckley/Casson 1976). The fundamental tenet of the theory can, thus, be succinctly put as follows: Firms choose an entry mode that minimizes the transaction costs associated with their international operations (Hennart 1988, Kogut/Singh 1988, Brouthers/Brouthers 2001). In other words, managers make a choice between equity and non-equity modes of entry by comparing the costs of internal coordination and control to those associated with finding, negotiating with, and monitoring an intermediary or contractor in an external market.
These transaction costs arguments have been successfully applied to investigate the MNE's preference for a particular ownership structure for its foreign affiliate among equity modes (Anderson/Gatignon 1986, Gatignon/Anderson 1988, Hennart 1988, 1991, Kogut/Singh 1988, Kim/Hwang 1992). In a similar vein, the choice between full and shared ownership depends on the relative costs and benefits of the two alternative ownership structures, i.e., joint venture vs. wholly owned subsidiary.
The transfer of tacit and intangible knowledge can be costly in an imperfect market because of the difficulty of the external market to determine the optimal pricing of these proprietary assets (for more detailed discussion, see Rugman/Verbeke 2003). This difficulty predisposes firms to choose internalized and hierarchical organizational structure over contractual arrangements for their foreign operation (Williamson 1985, Hennart 1988). There is a general consensus in the literature that firms equipped with high levels of proprietary assets prefer a wholly-owned subsidiary (WOS) to a joint venture (JV) among available FDI entry strategies (Guillen 2003). Not only the level but also the complexity and tacitness of proprietary assets are positively associated with high control modes of entry. Hill et al. (1990) provided theoretical explanations for the relationship using a unifying conceptual framework, especially the dissemination risk construct underpinned by transaction costs theory. According to Hill et al. (1990), firms engage in a low dissemination risk and cost mode (i.e., WOS over JV) as they do not want their proprietary assets, intangible and tacit assets in particular, to be disseminated to, or dissipated by, their local partner(s). By completely internalizing their operation through WOS, they are able to transfer their firm-specific advantages to their affiliates while preventing them from being unduly exploited by local partners in the host country. Through this fully internalized control, they can thus minimize transaction costs incurred to transfer their firm-specific proprietary assets to the foreign market.
Although firms still face constraints of various forms, both internal and external, (e.g., administrative heritage in Bartlett/Ghoshal 1989) in transferring knowledge to their overseas subsidiaries (Rugman/Verbeke 2003), they can transfer their firm-specific assets across borders more easily through internalization than through the market mechanism (Buckley/Casson 1976). In the context of equity ownership, MNEs prefer to choose WOS over JVs when they transfer high levels of proprietary assets. This is because the transfer of these often highly unstructured and tacit assets can be better managed through a high control mode than through a low control mode. The use of high control modes is particularly suggested in emerging host markets that have weak intellectual property (IP) protection mechanisms (Chen/Hu 2002). Through WOS, MNEs are able to transfer their proprietary assets more appropriately, while protecting them from undue exploitation by their JV partners. The following hypothesis is thus proposed:
Hypothesis 1. The greater the degree of proprietary assets (technology) transferred from parent to affiliate, the greater the association with a high control entry mode.
The nature of industry a firm operates in moderates the impact the transfer of proprietary assets such as technology has on entry mode choice. Kogut and Singh (1988) found that MNEs tend to choose JVs over WOS when they invest overseas in R&D-intensive industries. This is particularly the case when investing in host countries with sophisticated and competitive high-technology industries because foreign MNEs, by partnering with local firms, are also able to tap into the complementary technologies and know-how of the host country. In addition, foreign MNEs investing in high-technology industries are often pressured by host governments to form JVs, so that local firms can benefit from access to the state-of-the art technology or know-how of their foreign partners (Child/Markoczy 1993). In vibrant economies like Korea, in particular, the MNE's bargaining power with the host country government is usually weakened (Tsang 2005), with the firm's ability to opt for a WOS being curtailed accordingly. Based on the above discussion, we hypothesize:
Hypothesis la. The greater the increases in technology transfer for a firm involved in high-technology industries, the greater the association with shared control mode of entry.
Internal transaction costs increase with geographical distance and cultural dissimilarity between the home country and the host country of an affiliate. The increase in transaction costs is due to added misunderstandings and growing expenditures that result from a greater need for continuous checking on performance (Rugman/Verbeke 2003).
The concept of cultural distance was empirically tested in Kogut and Singh's (1988) seminal work of 228 foreign ventures in the U.S. It has since been frequently used in several subsequent studies to explain the entry mode behaviour of MNEs, with some studies (e.g., Kogut/Singh 1988, Erramilli/Rao 1993, Agarwal 1994, Hennart/Larimo 1998) supporting a negative relationship between cultural distance and control. The preference for a JV over a WOS in culturally distant markets can be explained by transaction costs arguments that firms can minimize transaction costs associated with the "liability of foreignness" (Hennart 1991) through partnership with a local partner(s) equipped with knowledge about local market conditions and political situations (Anand/Delios 1997). The partnership with local partners allows a foreign MNE to better manage its local employees and relationships with local suppliers, distributors and governments. Furthermore, parent managers are usually reluctant to commit their resources to highly uncertain and unfamiliar situations and thus select a low resource commitment mode.
In contrast, Chen and Hu (2002) recently found that foreign MNEs from culturally distant markets show a greater tendency to select a high control mode (i.e., WOS) in China than MNEs from culturally similar home countries. This positive relationship between cultural distance and control is also supported in prior studies. For example, Padmanabhan and Cho (1996) and Anand and Delios (1997) both found that greater cultural distance between Japan and the host country of Japanese MNEs is significantly related to their choice of a full control mode (i.e., WOS). This contradictory relationship may be explained by the same transaction costs theory that firms engage in a high control mode in culturally distant countries because the cost of finding, negotiating with, and monitoring a local JV partner(s) is greater than the cost incurred in a WOS (Brouthers/Brouthers 2001). The costs of using an external partner become very high due to bounded rationality and opportunism arising from information asymmetry that is exacerbated in an international context, especially in highly distant and dissimilar host markets. The uncertainty and unfamiliarity arising from greater cultural distance makes it difficult not only to address all the contingencies and potentialities in their JV contracts due to greater bounded rationality, but also to enforce the contracts because of greater opportunistic behaviour by local partners resulting from information asymmetry (Williamson 1985, Hill 1990, Taylor et al. 1998, Chang/Taylor 1999).
Notwithstanding the contradictory empirical results in the literature, we can argue that foreign MNEs from culturally distant markets may well prefer a WOS over a JV in host countries like Korea with a relatively less sophisticated FDI regime and its short history. It may be difficult for foreign MNEs to find a suitable local partner(s), let alone draw up a complicated and comprehensive JV contract (or agreement) with the local partner(s) in the host country. The costs associated with a JV become greater for firms from culturally dissimilar markets because of heightened inability to anticipate the behaviour of local partners embedded in a vastly different local culture. The following hypothesis is proposed:
Hypothesis 2. The greater the cultural distance between the home country of an investing MNE and the host country, the greater the association with a high control entry mode.
Despite its usefulness as one of the most frequently cited variables, the national cultural distance construct has been often subjected to severe criticisms for its flawed conceptual and methodological properties (Shenkar 2001). While the construct provides a simple, quantitative, and tangible measure (e.g., Kogut and Singh's composite index of national cultural distance), it largely fails to capture the intricate, complex and fuzzy nature of national culture (Shenkar 2001) and neglects the regulatory dimensions of institutions within the country (Harzing 2003).
There has been an increasing call to incorporate equally important home and host country characteristics (e.g., country risk, host country restrictions on foreign direct investment, and availability of local partners) in the explanation of entry mode choices by MNEs. Institutional theory has been suggested to supplement the heavy reliance on transaction costs theory (North 1990, Brouthers/Brouthers 2000, Xu/Shenkar 2002, Harzing 2003, Xu/Pan/Beamish 2004). The usefulness of institutional theory is supported by the argument that the legitimacy of a firm is endowed by institutions where transactions of the firm take place. A firm's choice of entry mode can thus be constrained by the institutional context in which the firm operates (Beamish/Delios 1997, Brouthers/Brouthers 2000). For example, a host country's institutions such as regulations on foreign ownership may restrict their access to local assets to protect domestic industries, thus affecting the choice of organizational forms in the local market (Harzing 2003). A firm's legitimacy in the restrictive host market becomes another important determinant in addition to transaction costs-based efficiency.
Despite increasing global liberalization and deregulation, some host countries still restrict majority or full foreign ownership in fledgling and/or strategic industries to protect their indigenous firms (Xu/Shenkar 2002). Restrictions on foreign ownership can be direct and/or indirect, as reflected in the following: Outright prohibition of foreign ownership, a ceiling on the extent of ownership, red tape, prior government approvals, and deprivation of financial and tax privileges awarded to domestic firms. Firms tend to enter a market with a mode of entry that requires small resource commitment where legal restrictions exist (Delios/Beamish 1999). From the above discussion, we propose the following hypothesis:
Hypothesis 3. The greater the level of restrictions on full foreign ownership the greater the association with a shared control mode of entry.
The underlying position of transaction costs theory is that the modes selected following the predictions of transaction costs theory will provide firms with the most efficient and least costly organizational design and lead to better performance (Williamson 1985, Chen/Hu 2002). However, other scholars (e.g., Ghoshal/Moran 1996, Zajac/Olsen 1993) suggest an equally plausible argument that the transaction costs-based mode does not necessarily lead to superior performance at the level of the affiliate. According to this view, cost minimization or efficiency maximization may not lead to the best performing mode because organizational performance is a complex, multidimensional phenomenon comprising both efficiency and value enhancement of an operation. Brouthers (2002) warns against the single-minded focus on the efficiency maximization and cost minimization tenets of transaction costs theory. He argues that an organization's entry mode will achieve optimal performance when other contextual variables are factored in to its decision, such as cultural and institutional variables. Distance between the home and host country, either cultural or institutional, may negatively affect affiliate performance due to increasing tensions arising from misunderstanding and miscommunication between parties across the borders. Institutional context variables are also believed to affect a firm's performance as institutionally-sanctioned organizational form and activities enhance the firm's legitimacy through its external isomorphism and ultimately its survival and success within a particular institutional context. Following the general assertion expressed in the literature (e.g., Brouthers (2002) in particular), the following hypothesis is proposed:
Hypothesis 4. Firms selecting entry modes consistent with the predictions of transaction costs theory and other contextual (i.e., cultural and institutional) variables will perform better than firms selecting entry modes otherwise.
Sample Selection and Data Collection
According to the Directory of Foreign Investment Companies (MOCIE 2002), there were 3542 foreign manufacturing ventures registered with the Ministry of Commerce, Industry, and Energy (MOCIE) as at 30 June 2002. The directory is updated annually by the Division of Import in the Ministry and contains information including company name, parent company, investment amount, founding year, and postal/email addresses of each registered company. With the support of an umbrella organization of the Ministry, Invest Korea (formerly, Korea Investor Support Centre, KISC), the Korea Institute for Industrial Economics and Trade (KIET) initially emailed the entire population of foreign manufacturing companies letters of request to participate as a panel company for the institute's database construction, i.e., Industry Statistics Database, which was to be used for the institute's surveys. Out of the total population, 1500 firms agreed to participate and as part of their agreement, each provided the name, phone number and email address of its planning/coordination department head or director. This database of 1500 firms comprised our survey sample. The then-current email addresses of our prospective respondents were used to minimize the rates of undeliverable requests that are often found to be high in electronic surveys (Sills/Song 2002).
With the support of the KIET, we then emailed cover letters with a link to the Web-based questionnaire to the department heads/directors of the survey sample, who were appropriately positioned to answer the survey questions. While the use of the entire population may remedy non-coverage error (Sills/Song 2002), we decided to focus on these 1500 firms to obtain quality responses and to minimize the occurrence of potential problems arising from electronically sending our questionnaire unsolicited or without advance warning (Kaye/Johnson 1999). This two-step approach is in line with Anderson and Gansneder (1995). To induce more cooperation and participation from firms that had not replied to the first email-out, two telephone follow-ups were made to non-response firms. This effort was further bolstered by specifying in the survey questionnaire the sources of support obtained, such as the Ministry, local government departments and relevant industry associations, which helped increase the credibility of the survey. In addition, a further mention was made in the questionnaire of the contributions of reliable information provided by each firm to the Ministry's future policy-making in areas of industry, technology and trade/energy. Another important feature of the on-line survey adopted in this study is a mechanism that enabled each respondent to engage in real-time interaction with the survey team to discuss concerns relating to the survey (e.g., clarity, security). To ensure that only the selected sample could complete the questionnaire, we required our respondents to enter their survey identification number and password. Further concerns (e.g., confidentiality, response duplication, security) were addressed by a password method of entry, personalized email messages, and further assurances of confidentiality in the questionnaire.
Electronic surveys are increasingly gaining in popularity as a powerful tool in survey research (Schaefer/Dillman 1998, Kaye/Johnson 1999, Cook/Heath/Thompson 2000, Sills/Song 2002). Despite their attractiveness in many respects (e.g., cost, ease of delivery), they are not without limitations such as sampling, technical issues, response rate, and security concerns. In order to prevent these potential problems occurring, we took various measures in survey design and implementation, including the use of pre-contacts, personalized contacts, password protection, and security and confidentiality assurances. Although there is still room for improvement (e.g., use of a mixed-mode strategy), the present survey with its careful survey design and implementation appears to be a sensible means of collecting meaningful data given it targets a specific Internet experienced sample. Furthermore, secondary sources such as company annual reports, documents, and websites were also used to complement the primary data collected through the survey. The primary data do not include all the necessary information in areas like ownership and lines of business. The actual ownership percentage of each firm was collected from the Ministry's directory and company annual reports. Where a firm's main line of business was not clear (e.g., 'other manufacturing'), its annual report and website were examined to determine the firm's grouping, i.e., 'high-tech' vs. 'low-tech' industry. Secondary sources are thus used as a complement to the primary data collected through the survey, given that not all firms (e.g., unlisted firms, in particular) disclose their information through secondary sources.
Through these efforts, we obtained responses from 262 firms, with an approximate response rate of 17 percent, out of which the responses of 34 firms were deleted because they do not satisfy the 10 percent ownership rule of the OECD, albeit labelled as FDI under the Foreign Investment Promotion Act (FIPA) of Korea. This elimination process resulted in a final sample of 228 firms for the first-stage logistic regression analysis. The sample used in the second-stage multiple regression analysis of performance was further reduced to 222 firms that were all at least three years old at the time of the survey, which is in line with prior studies (e.g., Chen/Hu 2002, Pangarkar/Lim 2003). According to Newbould et al. (1978), a three-year requirement is necessary because it would be difficult to assess the affiliate performance, especially its financial performance, in the early stages of foreign entry. This criterion is slightly more stringent than that of Nitsch et al. (1996) including subsidiaries that are at least two years old. The response sample distribution was found to be similar to that of the entire population in respect of industry classification, location, and parent nationality.
We developed a draft questionnaire in English that was translated into Korean by a bilingual author. The Korean version of the draft was then cross-checked by two other Korean researchers and academics in the field and adjusted accordingly. Designed for a larger project, the questionnaire contains a large number of questions, including entry mode, source of material and sales, R&D activities, technology transfer, relationship with local suppliers/distributors, and performance, some of which will be used in further studies.
Following Shaver (1998) and Brouthers (2002), we adopted a two-stage approach to our analysis to test whether firms selecting theoretically and contextually predicted entry modes perform any differently than firms whose entry mode choices were not predicted by the model. First, we conducted a binomial logistic regression analysis to examine the relationship between independent variables (including control variables) and the dichotomous dependent variable, i.e., entry mode (JV vs. WOS). The binomial regression model is chosen for its robustness and the categorical nature of the dependent variable (Hair et al. 1998) where JV is treated as a baseline case with a value of zero. A positive regression coefficient means that the independent variable increases the probability of choosing an alternative mode (i.e., WOS) while reducing the probability of choosing the baseline mode (i.e., JV). In contrast, a negative coefficient decreases the predicted probability of selecting a WOS over a JV. Second, from the logistic regression, two groups were derived, i.e., 'fit' and 'no-fit' groups. The 'fit' group denotes those firms that selected a mode predicted by the extended transaction cost model, while the 'no-fit' group refers to firms that selected a mode different to the one prescribed by the model (Brouthers 2002). We then compared the two groups in performance by using univariate analysis (two-sample t-test) and then multiple regression analysis to isolate the contribution of mode fit to affiliate performance, while controlling for other variables that are conjectured to impact on performance.
Measurement of Variables
Two dependent variables are included in this study. First, our dependent variable for testing entry mode choice is a dummy variable that assumes one if the equity of the affiliate is wholly owned and zero if it is jointly owned. A shared-equity affiliate is one which is equal to or greater than 10 percent but less than 95 percent foreign-owned, while a wholly owned affiliate is 95-100 percent foreign-owned. In line with Curhan, Davidson and Suri (1977), the usual 95 percent cutoff is also adopted in this study. While recognizing alternative types of entry mode (e.g., contractual arrangements), our focus is limited to FDI entry modes (Zhao/Luo/Suh 2004), with particular attention to ownership modes, i.e., shared vs. full ownership. A continuous dependent variable is also used to explore if relationships between the variables are significantly different to the ones obtained when a dichotomous variable is used. However, as the focus of this study is on testing whether a firm's observed group membership is consistent with the one prescribed by the extended transaction cost model, a dichotomous classification, albeit somewhat limiting and arbitrary, is ultimately used in the final logistic regression model. This dichotomous classification has also been adopted in several prior studies (e.g., Padmanabhan/Cho 1996, Brouthers/Brouthers 2001, Cho/Padmanabhan 2005).
For our second dependent variable, performance, we adopted a multidimensional approach (Barkema/Vermeulen 1997) by including the following five items: Sales growth, market share, operating profit, employee retention, and productivity, These multiple measures of performance (both financial and non-financial) were obtained from affiliate respondents based on their perceptions of performance at the affiliate level. While recognizing the shortcomings of subjective measures, we believe their perceptions do actually matter in determining the actions that will be taken by the affiliate. The use of subjective measures can also be justified by the high correlation between objective and subjective measures of performance found in prior studies (e.g., Dess/Robinson 1984, Geringer/Hebert 1991). Multiple measures are used to avoid the problems associated with depending on narrowly defined criteria (e.g., profitability) because financial performance may not be the only determinant of the effectiveness of foreign entry (Kitching 1974, Anderson 1990, Geringer/Hebert 1991, Kim/Hwang 1992, Luo 1997, Pangarkar/Lim 2003), let alone the difficulty of assessing it. All performance variables were measured on a seven-point Likert scale where 1 denotes "very low", while 7 indicates "very high" compared to the domestic industry average. While a scaling method of this kind is somewhat limited in addressing the complexities and intricacies of performance, it has been commonly used in prior studies adopting cross-sectional survey methods for data collection. An attempt is made in this study to reduce the inherent limitations of the quantitative survey method by including multiple measures of performance.
The degree of transfer is measured on a seven-point Likert-type scale, ranging from 1 ("very little") to 7 ("very significant"). Firms not engaging in technology transfer are given a corresponding scale, which is consistent with Johnston's (1972) suggestion, to minimize a potentially significant reduction of the sample size. Compared to other proxies used in prior studies such as R&D or advertising intensity (Kogut/Singh 1988, Padmanabhan/Cho 1996), this measure is a more accurate representation of a particular investment's expropriation risks in a host country (Zhao et al. 2004). This is particularly true because the parent's R&D or advertising intensity itself may not necessarily translate into the actual degrees of risk an MNE faces.
National cultural distance was used to capture the cultural context in which a firm operates. The national cultural dimensions put forward by Hofstede (1980), CCC (1987), and Hofstede and Bond (1988) are used to compute the aggregate cultural distance index. Given its relevance in the Confucian Korean culture, we expand Kogut and Singh's (1988) formula by including the short-term/long-term orientation dimension (also referred to as 'Confucian Dynamism') in the calculation of the aggregate index (see Kogut/Singh (1988) for more detail). Our sample contains countries for which the scores of some of Hofstede's five dimensions are unavailable. The missing cases for the original four value dimensions are treated as "missing" rather than being replaced with the score of a country closest to the country concerned (e.g., Pan 1996, Chen/Hu 2002). This is due to the potentially substantial differences between the two allegedly "close" countries (Hofstede 2001, Harzing 2003). This treatment can be justified by a negligible number (2) of missing cases. However, due to several missing cases for the fifth time-orientation dimension, a replacement method is adopted. Consistent with Barkema and Vermeulen (1997), Read's (1993) scores of "marginal propensity to save" are used to compute a composite cultural distance index.
The degree of host country restrictions was captured by a dummy variable that assumes one if the affiliate was founded after the mid-1980s (non-restrictive host country policy), and zero otherwise (restrictive host country policy). Realizing the need for foreign capital in the country's continued economic growth, the Korean government significantly relaxed restrictions on foreign ownership from the mid-1980s, especially in manufacturing sectors and accelerated its deregulation and liberalization of various industries in the 1990s (Rho 2002). While some restrictions still existed in the form of indirect discrimination (e.g., red tape, prior governmental approval) against foreign investors, the country significantly lifted its historically stringent restrictions on full foreign ownership since the mid-1980s in manufacturing and ultimately extended this open policy to the service sector following the 1997 economic crisis (KISC 2003).
Lastly, entry mode fit used as the independent variable in the second-stage multiple regression is separately explained in the results section. Additional control variables used in the second-stage multiple regression models (i.e., entry mode, location, and investment motives) are also addressed in the results section.
We include three control variables for the logistic regression model: Establishment mode, size, and industry. A dummy variable is used to capture the establishment mode, which is equal to one if the affiliate is established through acquisition and zero if it is set up through a greenfield operation. We used the number of affiliate employees as a measure of affiliate size. As the skewness and kurtosis of the size variable are significant, data transformation using natural logs was made to satisfy the normality assumption. While industry is controlled for in our study because our sample is limited to manufacturing firms, we use a sub-sectoral dummy to control for any exogenous sub-industry effect by grouping the firms into two categories using the two-digit Korean SIC: Low-technology vs. high-technology industry, following Chen and Hu (2002). The dichotomous variable is equal to one when the affiliate is in the high-technology industry group and zero when it is in the low-technology industry group.
Results and Discussion
Before running the logistic regression, a correlation analysis (Table 1) was conducted to gain a preliminary understanding of the relationship between entry mode and the independent variables including the control variables, and to detect any multicollinearity among the independent/control variables used in both logistic and multiple regressions.
As shown in Table 1, the correlation analysis shows some statistically significant relationships with entry mode; however, no particularly strong correlation has been detected among the independent/control variables, posing no multicollinearity problem in the regression models (Aczel 1996, Hair et al. 2006). Additional diagnostics of multicollinearity (e.g., variance inflation factor (VIF) and tolerance) are discussed in the multiple regression results section.
The results of the logistic regression analysis are provided in Table 2. The total number of observations is 215 due to missing cases for one or more variables. The regression model in Table 2 shows significant explanatory power with the chi-square of 61.989 and the p-score of 0.000. In addition, the Cox and Snell R-square and Negelkerke R-square measures also demonstrate the good fit of the model, with a high level of explanatory power. The model explains 70.7 percent of the entry mode actually selected. This figure compares favourably with the classification ratios found in prior studies such as Brouthers and Brouthers (2001) and Padmanabhan and Cho (1996) although they are slightly lower than those of Barkema and Vermeulen (1997), Brouthers (2002), and Harzing (2002). More importantly, the classification accuracy of the model is much greater than the chance rate (50.19 percent), with an improvement rate of 40.86 percent far surpassing the generally accepted minimum rate of 25 percent. The model thus provides very meaningful information for identifying group membership.
While the degree of technology transfer is positively associated with the choice of JVs over WOS, and thus contrary to Hypothesis l, this is not statistically significant. Interestingly, this association is counter to the majority of prior findings and transaction costs arguments (Hill et al. 1990, Guillen 2003). According to the literature, firms equipped with high levels of proprietary assets prefer a WOS to a JV to prevent their assets from being unduly exploited by their local partner(s). This line of argument has been reinforced by transaction costs theorists who postulate that firms, through direct control, can minimize risks and costs associated with the transfer of their assets, intangible and tacit assets in particular.
Similar to the cultural distance/experience interaction variable (Cho/Padmanabhan 2005), the technology transfer/industry interaction variable measures the 'compensating' effect of a firm's industry (high tech vs. low-tech) on the relationship between technology transfer and entry mode. The "moderated" technology transfer variable (technology transfer/industry) fails to support Hypothesis 1 a at p < 0.05. The lack of a moderating effect is further evidenced in additional logistic regressions conducted on two industry sub-samples, i.e., high vs. low-tech industries (Table 3). However, as shown in Table 2, the moderately significant relationship with shared modes of entry clearly warrants closer examination.
Table 3 shows that technology transfer is not significantly related to ownership mode. However, technology transfer is moderately associated with shared control modes of entry for the high-tech industries, although no statistical significance is demonstrated for the low-tech industries.
Table 2 indicates that cultural distance has a significant effect on entry mode decisions in Korea, as predicted, thus supporting Hypothesis 2. More specifically, foreign MNEs in Korea favour a high control entry mode i.e., WOS over JVs, the greater the cultural distance between home and host countries. This finding is consistent with that of Padamanabhan and Cho (1996) and Anand and Delios (1997), and Chen and Hu (2002). Given Korea's relatively short history of massive inward foreign investment, it may have been difficult for foreign MNEs to find suitable local partners, let alone draw up a complicated and comprehensive JV contract with local firms a view also supported by the findings of a recent KPMG survey (2001) in which finding a JV partner(s) was the most difficult obstacle to foreign MNEs' entry into the country. This problem is often exacerbated by the degree of cultural distance between the MNE's home market and Korea because the foreign MNE's learning is further disturbed by the psychic distance felt by the MNEs in making its entry into the host country (Johanson/Vahlne 1977).
As also indicated in Table 2, foreign MNEs under a non-restrictive host country FDI regime were found to choose a high control mode of entry (WOS), whereas foreign MNEs chose JVs over WOS when restrictions on full ownership existed. However, the results do not support Hypothesis 3 at a conventional level of p < 0.05, albeit in line with those of Gomes-Casseres (1990) and Padmanabhan and Cho (1996).
As previously mentioned, we have also used a continuous dependent variable of ownership to test if the results remain the same. The results show no significant difference to the ones obtained from the use of a dichotomous ownership variable (shared vs. full) (1).
Prior to running multiple regression for our second-stage analysis, we conducted preliminary and exploratory tests: Descriptive statistics, correlation, and univariate analysis (Two-sample T-test). Descriptive statistics show that none of the variables used in the multiple regression have significant skewness or kurtosis, except for affiliate size with 4.604 and 28.942 each. After the log-transformation, the variable shows much smaller skewness and kurtosis than that of the non-transformed variable, with P-P normality plots also approximating normal distribution. As noted earlier, there is no particularly strong correlation that warrants concern about multicollinearity (Hair et al. 2006). As indicated in Table 5, this is also supported by the VIFs ranging from 1.019 to 1.221 and the Tolerance levels ranging from 0.819 to 0.982, showing that neither of the variables is correlated significantly with any of the remaining variables (Weisberg 1985, Maddala 1992).
The two-sample t-test in Table 4 displays no significant differences between the fit (i.e., correctly classified) and no-fit groups (i.e., not-correctly classified) in all performance areas. However, these univariate results fail to show the unique impact of entry mode fit on performance, independent of other potentially powerful explanatory variables.
To isolate its contribution to explaining affiliate performance, multiple regression is thus undertaken. Additional control variables used in the multiple regression include entry mode (JV = 0; WOS - 1), location (non-special zone = 0; special zone = 1), and investment motives (i.e., tour types derived from factor analysis of eight items on motives used in the survey: Strategic asset-seeking; regulation-avoidant/incentives-seeking; simple resource-seeking; and market-seeking, similar to Dunning's (1998) classification). Results of the multiple regression are presented in Table 5 below.
As shown in Table 5, two models (i.e., sales growth and market share) have been found to be statistically significant at the level of p < = 0.05, with F-ratios ranging from 1.094 to 2.899. While the relatively low R-squared values (0.05-0.13) are of some concern when compared to those (i.e., 0.16-0.29) of Brouthers (2002) and Pangarkar and Lira (2003), these levels are acceptable given the complex and multidimensional phenomenon of performance (Pangarkar/Lim 2003). These low values, however, warrant examination of how the affiliate is actually managed once a mode has been selected, as the entry mode decision is only one of many strategic decisions affecting affiliate performance.
The results for the entry mode fit (Group) variable shown in Table 5 are largely consistent with those of the exploratory two-sample t-test. While the fit group performs better, albeit not significant statistically, in all financial performance areas, this group shows significantly poorer performance in non-financial areas, i.e., market share and employee retention, than the no-fit group. These results are contrary to those of the prior literature (e.g., Brouthers 2002, Chen/Hu 2002). In his study of selected EU companies, Brouthers (2002) found that the fit group performed significantly better than the no-fit group in both financial and non-financial performance areas, and particularly in respect of non-financial measures including market share, marketing, reputation, and market access. Similarly, foreign subsidiaries operating in China that select entry mode according to the prescription of transaction costs theory have been found to be more likely to succeed than firms choosing otherwise (Chen/Hu 2002).
These contradictory results may be attributable to a number of factors. First, affiliate respondents may have found it too sensitive to reveal information about their relative financial performance within the industry. Second, an entry mode chosen in accordance with the proposed extended model largely based on transaction costs theory may not necessarily lead to superior financial performance. In particular, the significant, negative relationship between the mode fit and the non-financial performance measures appears to cast some doubt on the fundamental tenet of cost minimization posited by transaction costs theory. The focus on economizing costs by the MNE through the choice of an entry mode may itself have adversely affected the market share and the employee satisfaction levels at its affiliate. This lends further support to the weakness of transaction costs theory, i.e., a focus on cost minimization, not value maximization (Zhao et al. 2004). Third, the relatively limited power of the models (R-squared) in explaining the variance of performance calls for the inclusion of other potentially powerful variables affecting affiliate performance, away from a heavy dependence on transaction costs-based explanations. Fourth and last, the significance found in non-financial performance measures (i.e., market share and employee retention) in the opposite direction provides some interesting implications for MNEs entering host cultures like Korea where group harmony ('inhwa') and market expansion are traditionally more valued than the profit- or revenue-maximization of firms. A single-handed focus on cost minimization may not be as relevant as proposed by transaction costs theory in the Korean setting where indigenous local firms traditionally promote non-financial aspects of performance (Sohn 1994, Lee/Miller 1999). Foreign affiliates may, thus, have perceived a nurturing climate rather than blind emphasis on profit maximization to be conducive to generating loyalty and dedication of employees that is critical to their long-term survival and success.
The results of this study overall appear to fail to support the normative aspect of the extended entry mode theory. However, any hasty conclusion of this kind should be avoided because the results may arise from various reasons such as methodological limitations, including the use of inaccurate proxies to capture the constructs underpinned by theories, or our neglect of other potentially important variables. For example, structural tools such as optimal communication and coordination mechanisms (Hennart 1991) may affect performance more strongly because firms, through the use of these tools, can manage their international affiliates more efficiently and thus ultimately reduce internal communication costs. Superior firm performance is rather generated when a firm distinctively utilizes its resources and capabilities and thus enjoys an "organizational advantage" (Madhok 2002, p. 536). For this reason, Madhok (2002) suggests that researchers discard their heavy dependence on market failure or imperfection explanations that have been advocated by transaction costs theorists and adopt different approaches (e.g., resource- or capabilities-based view of the firm (Barney 1991, Peteraf 1993)) towards explaining how and why firms can enjoy enduring sources of competitive advantage.
A firm's capabilities and more generally, resources (e.g., knowledge) are an important driving force behind MNEs' behaviours. Referred to as a firm's processes of integrating, acquiring, and releasing its resources, dynamic capabilities become a source of competitive advantage not only by enhancing existing resource configurations but also by creating a new resource base as markets change and evolve (Eisenhardt/Martin 2000). Unlike transaction costs theory, this capabilities- and resource-based view shifts its focus from the transaction to the firm by linking entry mode decision to the firm's overall strategy (Peng 2001). More importantly, it highlights the importance of an MNE's capabilities and learning from prior entry experience (Chang 1995, Chang/Rosenzweig 2001) as well as the pulling effects of the capabilities of local firms in overseas markets (Peng/Wang 2000, Peng 2001). However, MNEs differ in capabilities. For instance, some MNEs successfully expand in overseas markets more easily than others, irrespective of mode. Contrary to objective and explicit knowledge, an MNE's capabilities (including implicit knowledge) are oftentimes tacit and invisible, even to the MNE itself as they are revealed through application and acquired through practice (Grant 1996). Due to this tacitness and invisibility, foreign expansion is often not limited to MNEs with superior capabilities. In other words, it is likely that MNEs, both successful and average, expand into overseas markets, which may partly explain the overall weak performance results found in the current study. Despite much academic interest in the potential connection between entry mode and performance, the results with respect to performance may in part be due to a lack of theory underlying the link between the two.
Among the control variables, JVs consistently perform better than WOS in all performance measures and in sales growth and market share in particular. A similar significant, positive relationship was also found for affiliate size. What is particularly notable is the consistently positive relationship between a particular investment motive (i.e., strategic asset-seeking motive) and performance, mostly at a statistically significant level. That is, the greater the level of the strategic-asset seeking motive (i.e., skilled labour, technology, or strategic production base for future entry into China and Japan), the better the performance of the Korean affiliate.
Conclusions and Limitations
We have examined the performance impacts of entry mode choice based on the perceptions of a large sample of managers of 222 foreign manufacturing firms in Korea. Using an extended transaction costs model, an evaluation of performance relative to predicted entry mode is carried out. This is an improvement on the majority of extant studies that simply compare performance between alternative modes of entry. Contrary to our initial expectation, firms following the entry mode predicted by the extended model perform only slightly better in respect of financial performance measures than non-following firms, but significantly worse for non-financial performance measures (i.e., market share and employee retention). This lack of support for the extended model appears to cast some doubt on its general applicability to the analysis of performance in a multidimensional and complex environment.
However, the limitation of the model may be attributable to the preoccupation of transaction costs theory with initial entry mode decisions, not the actual management of an established foreign operation and its network (Madhok 2002, Rugman/Verbeke 2003). Transaction costs analysis may be limited in explaining management behaviour (Birkinshaw 2000). As mentioned earlier, a firm's unique deployment and management of its internal resources and capabilities is potentially more instrumental in generating superior performance by creating organizational advantage (Barney 1991, Peteraf 1993). These capabilities, however, differ across firms and as a result, some firms successfully expand in overseas markets more easily than others, irrespective of mode.
The lack of significance for financial performance may also be related to the reluctance of the respondents to provide information as sensitive as financial performance. In addition, the significant negative relationship for non-financial performance measures calls for further examination of the usefulness of the extended model, especially in host cultures that value group harmony and employee satisfaction. The failure to find superior performance in the fit group may actually reflect the nature of the sample. It may be related to the intensity of competition in Korea's high-technology industries (e.g., chemical products, industrial machinery, and automobiles) that constituted approximately 60 percent of the sample firms' investments.
The present study has employed both financial and non-financial performance measures in the belief that the roles and performance expected of foreign affiliates are diverse and varied, and that the sole dependence on traditional financial measures, e.g., profitability, may not adequately capture affiliate performance. The measures used in this study are all based on the perceptions of affiliate senior managers/executives who are influential in determining the course of action of the affiliate. The reliance on these subjective measures can be justified by the high correlation between objective and subjective measures found in prior studies, as noted earlier. However, future research including objective measures will likely provide valuable insight into whether objective measures are consistent with respondents' perceptions about firm performance. The use of objective measures will also resolve the problem of common method variance (Luo 2003) as well as potentially different definitions (norms) of the same scale across firms (Brouthers 2002). The perceptions of home country managers will likely shed some useful light on affiliate performance as well. Future studies involving multiple entry modes which extend the current study's focus on FDI modes (i.e., shared vs. full control mode) will also further our understanding of the relationship between entry mode and performance. Finally, refining the theoretical model and employing more sophisticated measures of the variables would also be beneficial.
Despite these limitations, the findings of this study clearly enhance our understanding of foreign MNEs' entry mode decisions in a largely under-researched host country context, Korea. More importantly, the study explores the performance implications of entry mode choice and the normative aspect of an extended transaction costs model, thus improving on the simple comparison of performance across different modes of entry frequently found in the extant literature. Finally, these results make a useful, if somewhat controversial, contribution to the entry mode literature by reinforcing a recent call for re-examining transaction costs theory.
We appreciate the invaluable and insightful comments of three reviewers and the editors of mir. Our thanks also go to the KIET for their generous support in data collection.
Manuscript received July 2005, revised December 2005, final revision received July 2007.
Aczel, A. D., Complete Business Statistics, Chicago: Irwin 1996.
Agarwal, S., Socio-Cultural Distance and the Choice of Joint Ventures: A Contingency Perspective, Journal of International Marketing, 2, 2, 1994, pp. 63-80.
Anand, J./Delios, A., Location Specificity and the Transferability of Downstream Assets to Foreign Subsidiaries, Journal of International Business Studies, 28, 3, 1997, pp. 579-603.
Anderson, E., Two Firms, One Frontier: On Assessing Joint Venture Performance, Sloan Management Review, 31, 2, 1990, pp. 19-30.
Anderson, E./Gatignon, H., Modes of Foreign Entry: A Transaction Cost Analysis and Propositions, Journal of International Business Studies, 17, 3, 1986, pp. 1-26.
Anderson, S. E./Gansneder, B. M., Using Electronic Mail Surveys and Computer-Monitored Data for Studying Computer-Mediated Communication, Social Science Computer Review, 13, 1, 1995, pp. 34-47.
Barkema, H. G./Vermeulen, F., What Differences in the Cultural Backgrounds of Partners are Detrimental for International Joint Ventures?, Journal of International Business Studies, 28, 4, 1997, pp. 845-864.
Barney, J., Firm Resources and Sustained Competitive Advantage, Journal of Management, 17, l, 1991, pp. 99-120.
Bartlett, C. A./Ghoshal, S., Managing Across Borders: The Transnational Solution, Boston: Harvard Business School Press 1989.
Beamish, P. W./Delios, A., Incidence and Propensity of Alliance Formation by US, Japanese and European MNEs, in Beamish, P. W./Killing, J. P. (eds.), Cooperative Strategies: Asian Perspectives, San Francisco: The New Lexington Press 1997.
Birkinshaw, J., Entrepreneurship in the Global Firm, London: Sage 2000.
Brouthers, K. D., Institutional, Cultural and Transaction Cost Influences on Entry Mode Choice and Performance, Journal of International Business Studies, 33, 2, 2002, pp. 203-221.
Brouthers, K. D./Brouthers, L. E., Acquisition or Greenfield Start-Up? Institutional, Cultural and Transaction Cost Influences, Strategic Management Journal, 21, 1, 2000, pp. 89-97.
Brouthers, K. D./Brouthers, L. E., Explaining the National Cultural Distance Paradox, Journal of International Business Studies, 32, 1, 2001, pp. 177-189.
Brouthers, L. E./Brouthers, K. D./Werner, S., Is Dunning's Eclectic Framework Descriptive or Normative?, Journal of International Business Studies, 30, 4, 1999, pp. 831-844.
Buckley, P./Casson, M., The Future of Multinational Enterprise, London: McMillan 1976. Chang, E./Taylor, M. S., Control in Multinational Corporations (MNCs): The Case of Korean Manufacturing Subsidiaries, Journal of Management, 25, 4, 1999, pp. 541-565.
Chang, S. J., International Expansion Strategy of Japanese Firms: Capability Building Through Sequential Entry, Academy of Management Journal, 38, 2, 1995, pp. 383-407.
Chang, S. J./Rosenzweig, P. M., The Choice of Entry Mode in Sequential Foreign Direct Investment, Strategic Management Journal, 22, 8, 2001, pp. 747-776.
Chen, H./Hu, M. Y., An Analysis of Determinants of Entry Mode and its Impact on Performance, International Business Review, 11, 2, 2002, pp. 193-210.
Child, J./Markoczy, L., Host-Country Managerial Behaviour and Learning in Chinese and Hungarian Joint Ventures, Journal of Management Studies, 30, 4, 1993, pp. 611-631.
CCC (Chinese Culture Connection), Chinese Values and the Search for Culture-Free Dimensions of Culture, Journal of Cross-Cultural Psychology, 18, 2, 1987, pp. 143-164.
Cho, K. R./Padmanabhan, P., Revisiting the Role of Cultural Distance in MNC's Foreign Ownership Mode Choice: The Moderating Effect of Experience Attributes, International Business Review, 14, 3, 2005, pp. 307-324.
Coase, R. E., The Nature of the Firm, Economica, 4, 4, 1937, pp. 386-405.
Cook, C./Heath, F./Thompson, R. L., A Meta-Analysis of Response Rates in Web- or Internet-Based Surveys, Educational and Psychological Measurement, 60, 6, 2000, pp. 821-836.
Curhan, J. P./Davidson, W. H./Suri, R., Tracing the Multinationals: A Sourcebook on US-Based Enterprises, Cambridge: Ballinger 1977.
Delios, A./Beamish, P., Ownership Strategy of Japanese Firms: Transactional, Institutional and Experience Influences, Strategic Management Journal, 20, 10, 1999, pp. 711-727.
Dess, G. G./Robinson, R. B., Measuring Organizational Performance in the Absence of Objective Measures: The Case of the Privately-held Finn and Conglomerate Business Unit, Strategic Management Journal, 5, 3, 1984, pp. 265-273.
Dunning, J. H., Location and the Multinational Enterprise: A Neglected Factor?, Journal of International Bushless Studies, 29, 1, 1998, pp. 45-66.
Eisenhardt, K. M./Martin, J. A., Dynamic Capabilities: What Are They?, Strategic Management Journal, 21, 10/11, 2000, pp. 1105-1121.
Erramilli, M. K./Rao, C. P., Service Firms' International Entry-Mode Choice: A Modified Transaction-Cost Analysis Approach, Journal of Marketing, 57, 3, 1993, pp. 19-38.
Gatignon, H./Anderson, E., The Multinational Corporation's Degree of Control over Foreign Subsidiaries: An Empirical Test of a Transaction Cost Explanation, Journal of Law Economics and Organization, 4, 2, 1988, pp. 305-336.
Geringer, J. M./Hebert, L., Measuring Performance of International Joint Ventures, Journal of International Business Studies, 22, 2, 1991, pp. 249-263.
Ghoshal, S./Moran, P., Bad for Practice: A Critique of the Transaction Cost Theory, Academy of Management Review, 21, 1, 1996, pp. 13-47.
Gomes-Casseres, B., Firm Ownership Preferences and Host Government Restrictions: An Integrated Approach, Journal of International Business Studies, 21, 1, 1991, pp. 1-22.
Grant, R. M., Toward a Knowledge-Based Theory of the Firm, Strategic Management Journal, 17, Special Issue, 1996, pp. 109-122.
Guillen, M. F., Experience, Imitation, and the Sequence of Foreign Entry: Wholly Owned and Joint-Venture Manufacturing by South Korean Firms and Business Groups in China, 1987-1995, Journal of International Business Studies, 34, 2, 2003, pp. 185-198.
Hair, J. F. Jr./Anderson, R. E./Tatham, R. L./Black, W. C., Multivariate Data Analysis. Upper Saddle River: Prentice Hall 1998.
Hair, J. F. Jr./Black, W. C./Babin, B. J./Anderson, R. E./Tatham, R. L., Multivariate Data Analysis, Upper Saddle River: Prentice Hall 2006.
Harzing, A. W. K., Acquisitions versus Greenfield Investments: International Strategy and Management of Entry Modes, Strategic Management Journal, 23, 3, 2002, pp. 211-27.
Harzing, A. W. K., The Role of Culture in Entry Mode Studies: From Negligence to Myopia?, Advances in International Management, 15, 2003, pp. 75-127.
Hennart, J.-F., A Theory of Multinational Enterprise, Ann Arbor, Michigan: University of Michigan Press 1982.
Hennart, J.-F., The Transaction Cost Theory of Joint Ventures, Strategic Management Journal, 9, 4, 1988, pp. 361-374.
Hennart, J.-F., The Transaction Costs Theory of Joint Ventures: An Empirical Study of Japanese Subsidiaries in the United States, Management Science, 37, 4, 1991, pp. 483-497.
Hennart, J.-F./Larimo, J., The Impact of Culture on the Strategy of Multinational Enterprises: Does National Origin Affect Ownership Decisions?, Journal of International Business Studies, 29, 3, 1998, pp. 515-538.
Hill, C. W. L., Cooperation, Opportunism, and the Invisible Hand: Implications for Transaction Cost Theory, Academy of Management Review, 15, 3, 1990, pp. 500-513.
Hill, C. W. L./Hwang, P./Kim, W. C., An Eclectic Theory of the Choice of International Entry Mode, Strategic Management Journal, 11, 2, 1990, pp. 117-128.
Hofstede, G., Culture's Consequences: Comparing Values, Behaviours, Institutions and Organizations Across Nations, Second Edition, Thousand Oaks: Sage 2001.
Hofstede, G., Culture's Consequences: International Differences in Work-Related Values, Beverly Hills and London: Sage 1980.
Hofstede, G./Bond, M. H., The Confucius Connection: From Cultural Roots to Economic Growth, Organizational Dynamics, 16, 4, 1988, pp. 4-21.
Johanson, J./Vahlne, J. E., The Internationalization of the Firm: A Model of Knowledge Development and Increasing Foreign Market Commitments, Journal of International Business Studies, 8, 1, 1977, pp. 23-32.
Johnston, J., Econometric Methods, New York: McGraw-Hill 1972.
Kaye, B. K./Johnson, T. J., Research Methodology: Taming the Cyber Frontier-Techniques from Improving Online Surveys, Social Science Computer Review, 17, 3, 1999, pp. 323-337.
Kim, W. C./Hwang, E, Global Strategy and Multinationals' Entry Mode Choice, Journal of International Business Studies, 23, 1, 1992, pp. 29-54.
KISC (Korea Investment Service Centre), Introduction to the Foreign Direct Investment System in Korea. Seoul, Korea: KISC 2003.
Kitching, J., Winning and Losing with European Acquisitions, Harvard Business Review, 52, 2, 1974, pp. 124-133.
Kogut, B./Singh, H., The Effect of National Culture on the Choice of Entry Mode, Journal of International Business Studies, 19, 3, 1988, pp. 411-432.
KPMG, Foreign Direct Investment in Korea (Summary Report), Seoul: KPMG Consulting 2001.
Lee, J./Miller, D., People Matter: Commitment to Employees, Strategy and Performance in Korean Firms, Strategic Management Journal, 20, 6, 1999, pp. 579-593.
Li, J., Foreign Entry and Survival: Effects of Strategic Choices on Performance in International Markets, Strategic Management Journal, 16, 5, 1995, pp. 333-351.
Luo, Y., Guanxi and Performance of Foreign-Invested Enterprises in China: An Empirical Inquiry, Management International Review, 37, 1, 1997, pp. 51-70.
Luo, Y., Market-Seeking MNEs in an Emerging Market: How Parent-Subsidiary Links Shape Overseas Success, Journal of International Business Studies, 34, 3, 2003, pp. 290-309.
Maddala, G. S., Introduction to Econometrics, New York: MacMillan 1992.
Madhok, A., Reassessing the Fundamentals and Beyond: Ronald Coase, The Transaction Cost and Resource-Based Theories of the Firm and The Institutional Structure of Production, Strategic Management Journal, 23, 6, 2002, pp. 535-550.
MOCIE, Directory of Foreign Investment Companies, Seoul, Korea: MOCIE (in Korean) 2002.
Newbould, G. D./Buckley, P. J./Thurwell, J. C., Going International: The Experience of Smaller Companies Overseas, London: Associated Business Press 1978.
Nitsch, D./Beamish, P. W./Makino, S., Entry Mode and Performance of Japanese FDI in Europe, Management International Review, 36, 1, 1996, pp. 27-44.
North, D. C., Institutions, Institutional Change, and Economic Performance, Cambridge: Cambridge University Press 1990.
Padmanabhan, P./Cho, K. R., Ownership Strategy for a Foreign Affiliate: An Empirical Investigation of Japanese Firms, Management International Review, 36, 1, 1996, pp. 45-65.
Pan, Y., Influences on Foreign Equity Ownership Level in Joint Ventures in China, Journal of International Business Studies, 27, 1, 1996, pp. 1-26.
Pangarkar, N./Lim, H., Performance of Foreign Direct Investment from Singapore, International Business Review, 12, 5, 2003, pp. 601-624.
Peng, M. W., The Resource-Based View and International Business, Journal of Management, 27, 6, 2001, pp. 803-829.
Peng, M. W./Wang, D., Innovation Capability and Foreign Direct Investment: Toward a Learning Option Perspective, Management International Review, 40, 1, 2000, pp. 79-93.
Peteraf, A., The Cornerstones of Competitive Advantage: A Resource-Based View, Strategic Management Journal, 14, 3, 1993, pp. 179-191.
Read, R., Politics and Policies of National Economic Growth, Unpublished Doctoral Dissertation. Stanford University 1993.
Rho, K., A Study on Labour-Management Relations of Automotive Components Subsidiaries of Foreign MNEs in Korea: 1997-2000, Unpublished Master's Thesis, Seoul: Korean Catholic University 2002 (in Korean).
Root, F., Entry Strategies for International Markets, Lexington: D. C. Heath 1987.
Rugman, A. M., Inside the Multinationals: The Economics of Internal Markets, New York: Columbia University Press 1981.
Rugman, A. M./Verbeke, A., Extending the Theory of the Multinational Enterprise: Internalization and Strategic Management Perspectives, Journal of International Business Studies, 34, 2, 2003, pp. 125-137.
Schaefer, D. R./Dillman, D. A., Development of a Standard E-Mail Methodology: Results of an Experiment, The Public Opinion Quarterly, 62, 3, 1998, pp. 378-397.
Shaver, J. M., Accounting for Endogeneity When Assessing Strategy Performance: Does Entry Mode Choice Affect FDI Survival?, Management Science, 44, 4, 1998, pp. 571-585.
Shenkar, O., Cultural Distance Revisited: Towards a More Rigorous Conceptualisation and Measurement of Cultural Differences, Journal of International Business Studies, 32, 3, 2001, pp. 519-533.
Sills, S. J./Song, C., Innovations in Survey Research: An Application of Web-Based Surveys, Social Science Computer Review, 20, 1, 2002, pp. 22-30.
Sohn, J. H. D., Social Knowledge as a Control System: A Proposition and Evidence from the Japanese Behaviour, Journal of International Business Studies, 25, 2, 1994, pp. 295-324.
Taylor, C. R./Zou, S./Osland, G. E., A Transaction Cost Perspective on Foreign Market Entry Strategies of US and Japanese Firms, Thunderbird International Business Review, 40, 4, 1998, pp. 389-412.
Tsang, E. W. K., Influences on Foreign Ownership Level and Entry Mode Choice in Vietnam, International Business Review, 14, 4, 2005, pp. 441-463.
Tse, D. K./Pan, Y./Au, K. Y., How MNCs Choose Entry Modes and Form Alliances: The China Experience, Journal of International Business Studies, 28, 4, 1997, pp. 779-805.
Weisberg, S., Applied Linear Regression, New York: John Wiley 1985.
Williamson, O. E., Comparative Economic Organization: The Analysis of Discrete Structural Alternatives, Administrative Science Quarterly, 36, 2, 1991, pp. 269-296.
Williamson, O. E., The Economic Institutions of Capitalism, New York: Free Press 1985.
Woodcock, C. P./Beamish, P. W./Makino, S., Ownership-Based Entry Mode Strategies and International Performance, Journal of International Business Studies, 25, 2, 1994, pp. 253-273.
Xu, D./Pan, Y./Beamish, P. W., The Effect of Regulative and Normative Distances on MNE Ownership and Expatriate Strategies, Management International Review, 44, 3, 2004, pp. 285-307.
Xu, D./Shenkar, O., Institutional Distance and the Multinational Enterprise, Academy of Management Review, 27, 4, 2002, pp. 608-618.
Zajac, E. J./Olsen, C. E, From Transaction Cost to Transaction Value Analysis: Implications for the Study of International Strategies, Journal of Management Studies, 30, 1, 1993, pp. 131-145.
Zhao, H./Luo, Y./Suh, T., Transaction Cost Determinants and Ownership-Based Entry Mode Choice: A MetaAnalytical Review, Journal of International Business Studies, 35, 6, 2004, pp. 524-544.
(1) The regression model using a continuous dependent variable of ownership was also powerful in explaining he variance of ownership at p = 0.000. Statistical significance (two-tailed) was found in the CD, establishment mode, and size variables, which is consistent with the results of the logistic model using a dichotomous ownership variable.
Youngok Kim ([mail])
Lecturer of International Business, School of Organisation and Management, University of New South Wales, Sydney, Australia.
Sidney J. Gray ([mail])
Professor of International Business, Faculty of Economics and Business, University of Sydney, Sydney, Australia.
Table 1. Pairwise Pearson Correlation among Variables Variables 1 2 3 4 1. Sales Growth 1.000 2. Market Share 0.296 *** 1.000 3. Operating 0.379 *** 0.309 *** 1.000 Profit 4. Employee 0.057 0.118 * 0.216 *** 1.000 Retention 5. Productivity 0.418 *** 0.332 *** 0.502 *** 0.091 6. Group 0.037 -0.073 0.019 -0.106 7. Mode -0.091 -0.109 -0.071 0.037 8. Setup -0.034 -0.150 ** -0.039 -0.008 9. CD 0.084 0.041 0.053 -0.008 10. Transfer 0.135 ** 0.082 0.248 *** 0.118 * 11. HP 0.114 * -0.007 -0.097 -0.042 12. Location 0.078 0.025 0.096 -0.004 13. Industry 0.081 0.083 0.008 0.136 ** 14. Size 0.066 0.178 *** 0.122 * -0.025 15. Motive A 0.235 *** 0.053 0.160 ** 0.028 16. Motive B -0.031 -0.128 * -0.069 -0.003 17. Motive C -0.050 -0.064 -0.016 -0.092 18. Motive D -0.016 0.068 0.066 0.062 Variables 5 6 7 8 1. Sales Growth 2. Market Share 3. Operating Profit 4. Employee Retention 5. Productivity 1.000 6. Group 0.004 1.000 7. Mode -0.009 -0.336 *** 1.000 8. Setup -0.007 0.263 *** 0.413 *** 1.000 9. CD 0.114 -0.199 *** 0.251 *** 0.131 * 10. Transfer 0.127 * 0.005 -0.098 -0.063 11. HP -0.033 -0.035 0.075 -0.085 12. Location -0.033 -0.029 0.151 ** 0.009 13. Industry -0.044 0.021 0.006 -0.032 14. Size 0.077 0.037 -0.054 0.102 15. Motive A 0.170 ** -0.082 -0.002 -0.026 16. Motive B -0.016 0.026 -0.014 0.074 17. Motive C -0.090 -0.027 -0.010 -0.152 ** 18. Motive D 0.083 -0.177 ** 0.176 *** -0.003 Variables 9 10 11 12 1. Sales Growth 2. Market Share 3. Operating Profit 4. Employee Retention 5. Productivity 6. Group 7. Mode 8. Setup 9. CD 1.000 10. Transfer 0.045 1.000 11. HP 0.169 ** 0.065 1.000 12. Location 0.044 -0.120* -0.212 *** 1.000 13. Industry 0.029 0.092 0.092 0.077 14. Size -0.093 0.043 -0.296 *** 0.003 15. Motive A 0.143 ** 0.211 *** 0.226 *** -0.028 16. Motive B -0.085 -0.050 -0.016 -0.119 * 17. Motive C -0.036 0.027 -0.046 -0.069 18. Motive D 0.150 ** 0.177 ** -0.061 -0.035 Variables 13 14 15 16 17 1. Sales Growth 2. Market Share 3. Operating Profit 4. Employee Retention 5. Productivity 6. Group 7. Mode 8. Setup 9. CD 10. Transfer 11. HP 12. Location 13. Industry 1.000 14. Size -0.005 1.000 15. Motive A 0.065 -0.170 ** 1.000 16. Motive B -0.022 0.016 0.000 1.000 17. Motive C 0.022 0.019 0.000 0.000 1.000 18. Motive D 0.017 -0.091 0.000 0.000 0.000 Variables 18 1. Sales Growth 2. Market Share 3. Operating Profit 4. Employee Retention 5. Productivity 6. Group 7. Mode 8. Setup 9. CD 10. Transfer 11. HP 12. Location 13. Industry 14. Size 15. Motive A 16. Motive B 17. Motive C 18. Motive D 1.000 *** p < 0.01; ** p < 0.05; * p <= 0.10 (two-tailed); N = 222 Group = group classification (0 = no-fit group; 1 = fit group); Mode = entry mode (0 = JV; 1 = WOS); Setup =setup mode (0 = greenfield; 1 acquisition) CD = composite index of cultural distance between home and host country using Hofstede's five cultural dimensions Transfer = transfer of technology from foreign parent to Korean affiliate HP = host country policy on FDI (restrictive period = 0; vs. non-restrictive period = 1); Location = affiliate location within Korea (0 non-special zone; 1 special zone) Industry = line of business (0 = low-technology; 1 = high technology); Size = number of employees at the affiliate Motive A= strategic asset seeking (or advanced resource-seeking); Motive B = incentive-seeking/ regulation-avoiding Motive C = simple resource-seeking; Motive D = market-seeking Table 2. Results of Logistic Regression: Entry Mode Choice Independent Variables Technology transfer (proprietary assets) -0.096 (0.147) CD 0.365 (0.131) *** HP (non-restrictive) 0.151 (0.380) Technology transfer *Industry -0.216 (0.162) (#) Control Variables Setup (acquisitions) 21.756 *** (7510.269) Size -0.201 (0.121) * Industry (high-tech) 0.747 (0.643) Constant -0.935 (0.981) Chi-square 61.989 Correct ratio (%) 70.7% Significance 0.000 N 215 Baseline rate 50.19% Improvement 40.86% Cox & Snell R-square 0.250 Nagelkerke R-square 0.335 *** p < = 0.01; ** p < = 0.05; *p < = 0.10 (two-tailed); (#) p < = 0.10 (one-tailed); Joint venture = 0; Standard Error in Parentheses CD =aggregate index of cultural distance using Hofstede's five dimensions Setup = establishment mode (greenfield = 0 vs. acquisition = 1) Size = number of employees at the affiliate Industry = line of business (low-technology =0 vs. high-technology = 1) HP = host country policy on FDI (restrictive period = 0; vs. non-restrictive period = 1) The significance for setup is determined by score statistics (not Wald score) due to extremely large coefficients Table 3. Results of Logistic Regression: Entry Mode Choice (Industry sub-samples) Independent Variables High-tech industry Low-tech industry Technology transfer -0.117 (0.067) * 0.118 (0.153) CD 0.383 (0.146) *** 0.285 (0.322) HP (non-restrictive) 0.027 (0.415) 1.08 (0.994) Control Variables Setup (acquisitions) 21.620 (8520.258) *** 22.103 (15883.953) *** Size -0.279 (0.133) ** 0.289 (0.325) Constant 0.211 (0.860) -3.780 (2.290) * Chi-square 51.570 13.251 Correct ratio (%) 70.6% 78.9% Significance 0.000 0.021 N 177 38 Baseline rate 49.45% 50.14% Improvement 42.77% 57.36% Cox & Snell R-square 0.253 0.294 Nagelkerke R-square 0.338 0.393 *** p < = 0.01; **p < = 0.05; * p = 0.10 (two-tailed); Joint venture = 0; Standard Error in Parentheses CD =aggregate index of cultural distance using Hofstede's five dimensions Setup = establishment mode (greenfield = 0 vs. acquisition = 1) Size = number of employees at the affiliate Industry = line of business (low-technology = 0 vs. high-technology = 1) HP = host country policy on FDI (restrictive period = 0; vs. non-restrictive period = 1) The significance for setup is determined by score statistics (not Wald score) due to extremely large coefficients Table 4. Results of Two-Sample T-Tests For Difference for Subjective Performance Measures Fit group (n = 143) No-Fit group (n = 64) Performance Mean SD Mean SD Measures Sales growth 4.50 1.1 4.41 1.205 Market share 4.52 1.249 4.72 1.201 Operating profit 4.36 1.119 4.31 1.125 Employee retention 4.57 1.190 4.86 1.413 Productivity 4.65 0.988 4.64 1.132 Performance DF T -score P-value Measures (two-tailed) Sales growth 205 0.530 0.597 Market share 205 -1.046 0.297 Operating profit 204 0.276 0.783 Employee retention 206 -1.430 0.156 (#) Productivity 205 0.063 0.950 The score of '4' means the industry average in Korea; two-tailed 4 significant at p <=0.10 (one-tailed) Table 5. Results of Multiple Regression: Performance Analysis Variables Sales Growth Market Share Group 0.035 (0.317) -0.121 (0.048) ** (fit-group) Mode -0.130 (0.043) ** -0.231 (0.001) *** (wholly-owned) CD 0.106 (0.074) * 0.056 (0.217) Industry 0.046 (0.252) 0.054 (0.213) (high-tech) Affiliate size 0.108 (0.060) * 0.186 (0.003) *** Location 0.087 (0.108) -0.002 (0.0487) (special zone) Strategic-asset seeking 0.227 (0.000) *** 0.066 (0.171) Regulation/incentives- -0.009 (0.446) -0.112 (0.050) ** seeking Simple resource-seeking -0.026 (0.353) -0.149 (0.014) ** Market-seeking -0.015 (0.418) 0.106 (0.063) * Constant 3.607 (0.428) *** 3.970 (0.457) *** F-statistic 2.071 ** 2.899 *** R-squared 0.096 0.130 N = 204 204 Variables Operating Employee Profit Retention Group 0.026 (0.362) -0.130(0.043) ** (fit-group) Mode -0.110 (0.074) * -0.013 (0.432) (wholly-owned) CD 0.056 (0.226) -0.062 (0.204) Industry -0.007 (0.462) 0.158 (0.013) ** (high-tech) Affiliate size 0.168 (0.009) *** -0.007 (0.461) Location 0.105 (0.068) * -0.017 (0.404) (special zone) Strategic-asset seeking 0.200 (0.002) *** 0.021 (0.383) Regulation/incentives- -0.032 (0.323) -0.007 (0.462) seeking Simple resource-seeking -0.014 (0.418) -0.102 (0.075) * Market-seeking 0.103 (0.074) * 0.049 (0.248) Constant 3.472 (0.427) *** 4.736 (0.489) *** F-statistic 1.856 * 1.094 R-squared 0.088 0.053 N = 203 205 Variables Productivity VIFs Tol. Group 0.036 (0.317) 1.178 0.849 (fit-group) Mode -0.049 (0.261) 1.221 0.819 (wholly-owned) CD 0.105 (0.080) * 1.155 0.866 Industry -0.047 (0.253) 1.019 0.982 (high-tech) Affiliate size 0.128 (0.035) ** 1.034 0.967 Location -0.030 (0.336) 1.048 0.954 (special zone) Strategic-asset seeking 0.167 (0.010) *** 1.060 0.943 Regulation/incentives- -0.025 (0.362) 1.026 0.975 seeking Simple resource-seeking -0.071 (0.155) 1.022 0.978 Market-seeking 0.098 (0.087) * 1.072 0.933 Constant 4.074 (0.395) *** F-statistic 1.379 R-squared 0.066 N = 204 *** p < = 0.01: ** p < = 0.05; * p < = 0.10 (one-tailed); standardized coefficients with p-values in parentheses; Constant (unstandardized coefficients with standard errors in parentheses) VIFs = variance inflation factors; Tol. = tolerance level Group = entry mode fit (no-fit group = 0; fit group = 1); Mode = entry mode (JV = 0; WOS = 1) CD = a composite index of cultural distance using five dimensions; Affiliate size = number of employees at the affiliate Industry = line of business (low-technology = 0; high-technology = 1); Location = affiliate location in Korea (non-special zone = 0; special zone = 1)
|Printer friendly Cite/link Email Feedback|
|Author:||Kim, Youngok; Gray, Sidney J.|
|Publication:||Management International Review|
|Date:||Apr 1, 2008|
|Previous Article:||Regional economic integration and foreign direct investment: the case of NAFTA.|
|Next Article:||Innovation in multinational firms: does cultural fit enhance performance?|