MNC characteristics and global learning.

ABSTRACT

This study attempted to identify firm characteristics related to global learning. The results of this study suggest that cultural proximity, ownership percentage, similarity of processes used and products produced, and competitive advantage are positively related to global learning. The coefficient of the variable representing the size of the MNC, however, is not statistically significant in any of seven regression equations. That might suggest that rapid advancement in information and transportation technology and subsequent decrease in costs of communication and transportation has made global learning affordable regardless of firm size. Meantime, there are no generally accepted measures to examine global learning. Consequently, this study developed seven measures to examine global learning, i.e., scope of global learning, diversity of global learning, diversity of production-related global learning, diversity of non-production-related global learning, intensity of global learning, intensity of production-related global learning, and intensity of non-production-related global learning. As such, this research contributes to collective knowledge on the knowledge-based view of the firm.

1. INTRODUCTION

With rapid advancement of communication and information technology and increasing interdependence of markets, the concept of global learning has received considerable research attention over the last decade. The value of global learning in the multinational corporation (MNC) can be particularly high because foreign markets often provide access to new ideas and stimuli that can be subsequently applied in other countries. By leveraging knowledge in different markets, MNCs are in a position to capitalize on market imperfections and achieve higher returns on their investments. Numerous researchers argued that higher levels of global learning lead to higher MNC performance.

Meantime inward foreign direct investments (FDIs) tend to have high levels of knowledge and to be the dominant channel of knowledge diffusion (Hejazi and Safarian, 1999). Research inquiry into knowledge leakage has long existed. Previous research (Caves, 1974; Globerman, 1979) showed that greater foreign presence is correlated with greater productivity in host countries. Thus many governments offer explicit and implicit incentives to MNCs to establish subsidiaries in their countries. Thus, the practical reasons for investigating factors affecting global learning seem compelling.

A learning corporation is a holographic organization in the sense that information about the whole is stored in each part of the company (Hedlund and Rolander, 1990). In a learning organization, the basic strategy, guiding principles of behavior, and access to detailed information are widely shared. Recent shifts in the business environment allow MNCs to acquire efficiently knowledge critical to their success. The most important of these shifts is the revolutionary improvements in information and communications technologies that make information and knowledge acquisition substantially more efficient for all firms (Porter and Millar, 1985; Ohmae, 1991). Thus information technology is of crucial importance to global learning.

In light of its managerial significance, one would expect a large international business literature on global learning. Despite its clear importance, however, global learning appears to remain one of the most persistently understudied areas in international business, and so little is known about MNC's global learning. Moreover, past research related to knowledge transfer in international business has traditionally focused on technology and other technical knowledge transfers (e.g., Davidson and McFetridge, 1985; Simonin, 1999).

The goal of this paper is to undertake a careful test of the proposition that MNC characteristics are related to the extent of learning and knowledge transfer between MNC headquarters and subsidiaries and among MNC subsidiaries. Correspondingly, the central question to be investigated in this paper is to what extent does MNC characteristics affect global learning among the MNC's components spread throughout the world? Meantime, there have been no generally accepted measures to examine global learning. This study attempts to develop such measures. As such, this research contributes to collective knowledge on the knowledge-based view of the firm. In the following section we review the literature related to MNC characteristics and global learning. Then the issue of their relationship is further developed with empirical research. Meantime this study's use of a sample of MNC subsidiaries in Korea extends the empirical scope of both the multinationality and global learning literatures. Concluding remarks discuss implications of the findings.

2. THEORETICAL BACKGROUND AND HYPOTHESES

A key asset of the MNC is the diversity of environments in which it operates. This diversity exposes it to multiple stimuli, allows it to develop diverse capabilities, and provides it with a broad learning opportunity. The mere existence of diversity, however, does not enhance learning. It only creates opportunities for global learning. The open exchange of information and knowledge is necessary for managers in one organizational unit to learn about and benefit from the experiences of other units in other national markets (White and Poynter, 1990). Such exchanges allow for the diffusion of innovative ideas from the originating unit to the rest of the MNC. In practice, however, such exchanges do not occur frequently. For example, US automobile manufacturers have benefited relatively little historically from the know-how in their European subsidiaries, although the potential for realistic experimentation with small car and performance car production certainly was there (Hedlund & Rolander, 1990). Meantime, global learning has two aspects. One is its scope, and the other is its diversity. In this paper, the scope of global learning addresses to whether learning occurred among their global components, and whether it is dyadic between headquarters and subsidiaries or globally if it occurred. Meantime the diversity of global learning addresses to the question of in how many value activities (Porter, 1985) and to what extent in those activities global learning takes place.

MNCs are broken down by countries, and the MNC network may operate in four different ways. First, many MNCs that have facilities in several countries still tend to run their operations independently. Geographic distance, divergence in customer requirements, and differences in cultures in countries where MNC components are located will all make global learning ever more difficult to take place. In other cases local subsidiaries may fail to exploit global learning benefits because of their inability to receive and synthesize knowledge and expertise developed in different components of the MNC network. In still other cases local loyalties, and turf protection may restrict internal flow of information across national borders, which is essential for global learning. As a result, these subsidiaries do not exchange information including management practices and knowledge with either their parent or their foreign sister subsidiaries. A company cannot innovate globally if its managers identify primarily with local parochial interests.

For MNCs, however, their early strategies were based on the transfer of technology and superior organizational abilities to overseas markets. And the organizational structure and operating systems of early MNCs were designed for the management of dyadic relationships between headquarters and overseas subsidiaries (Kogut, 1990). In those firms, where all organizational resources are centralized and where local subsidiaries are seen as mere delivery pipelines to supply the organization's value-added to different countries, global learning may not take place, in those cases, local subsidiaries may not possess appropriate sensing, analyzing, and responding capabilities to learn from other subsidiaries since the centralized decision processes in the MNC headquarters may be insensitive to knowledge accumulated outside the corporate headquarters. In other cases local subsidiaries may fail to exploit global learning benefits because of their inability to receive and synthesize knowledge and expertise developed in different components of the MNC network except in headquarters.

The implementation of a global strategy also tends to enhance the forces of centralization and to shift organizational power from the subsidiaries to the headquarters (Ghoshal, 1987). This may result in demotivation of subsidiary managers and may erode the potential for learning from its many environments, which is one key asset of the MNC. Although both centralization and decentralization may hinder global learning, MNC's initial stock of knowledge, which was the strength that allows it to go abroad in the beginning, may produce better results compared with the decentralized MNC that is on its own.

Although in many organizations expertise is still held locally, the overall corporate outlook within a MNC is shifting in directions that will demand more innovative effort by all parts of the organization (McGuinness & Conway, 1986). In many cases, the most powerful insights come from looking outside one's immediate environment to gain a new expertise. Enlightened managers know that even companies in completely different businesses can be fertile sources of ideas and catalysts for creative thinking (Garvin, 1993). For learning to be more than a local affair, knowledge must spread quickly and efficiently throughout the MNC. Ideas will carry maximum impact when they are shared globally throughout the MNC's network rather than held locally. The components of MNC will be able to advance their skills in many fields by learning from each other, and apply what they learned to renew the capabilities of the MNC. In global economy, a well-developed ability to create and sustain fruitful global learning will give the MNC network a significant competitive advantage. Any MNC, which neglects this key source of competitive advantage, will pay clearly in its economic performance. The strategic interdependence of global markets implies that competitive position in one national market is affected by its competitive position in other national markets (Porter, 1986). The MNC network will be able to create competitive advantage by integrating good ideas from different parts of the world. Transferring knowledge to different parts of the MNC network will help share the wealth of knowledge. Transfers may be from the headquarters to subsidiaries, or from subsidiaries to the headquarters, or even to other subsidiaries. They may involve senior, middle, or first-level managers. In MNCs good at learning, for example, a supervisor experienced in just-in-time production might move to a plant in another subsidiary to transfer the methods there, or a successful plant manager might transfer to a lagging subsidiary to invigorate it with already proven ideas. The first step of global learning would take place between MNC parent and a specific subsidiary. But as the organization accumulates learning experience, the scope of learning would widen to include the entire MNC global network.

As mentioned above, meantime, the diversity of global learning addresses to the question of in how many value activities and to what extent in those activities global learning takes place. Competitive advantage cannot be understood by looking at any one activity a firm performs or a firm as a whole. Porter (1985) identified value activities that are physically and technologically distinct activities a firm performs. Those activities include procurement of raw materials, parts and components, manufacturing, product design, process design, marketing, sales, finance, accounting and legal, and human resource management. Each of these activities is a building block of competitive advantage (Porter, 1985). And learning may take place in any managerial activities, in a global environment where change is fast and competition intense, fast learning in multiple functions would be crucial for the improvement of MNC performance. The resource-based view of the firm posits that performance differences across firms can be attributed to the variance in the firms' resources and capabilities. Resources that are valuable, unique, and difficult to imitate can provide the basis for firms' competitive advantages (Amit and Schoemaker, 1993; Barney, 1991). In turn these advantages produce positive returns (Peteraf, 1993). Likewise, extensive global learning taking place in many of the above-mentioned value activities are more likely to produce a competitive advantage because it is often rare and socially complex, thereby making it difficult to imitate. Thus diversity of knowledge transferred among various parts of MNC can permit the MNC subsidiary to enhance competitive advantage.

While the transfer of knowledge between departments or between sister units in the same country is difficult, it is clear that the problems associated with knowledge transfer will increase with cultural distance. In this connection, the agency theory suggests the difficulty of specifying tacit information embedded in a firm's trained personnel and technicians to local employees in the host country. This will partially depend on the absorptive capacity of local employees learning the new routines. The agency theory also suggests the increasing difficulty of assessing the abilities of foreign employees, and monitoring them for performance as the cultural difference between source and host countries increases. Subsidiaries established in culturally distant countries encounter larger knowledge barriers regarding local political, cultural and societal norms (Hennart and Larimo, 1998). As the cultural distance between source and host countries increases, the investment in the non-redeployable assets in the host country becomes riskier (Gatignon and Anderson, 1988). This is partly because of the higher cost of acquiring information to monitor and evaluate business (Jones and Hill, 1988; Kogut and Singh, 1988). Thus the exchange of ideas and knowledge between culturally distant MNC components will be costly. Knowledge creation from the information acquired from more culturally distant subsidiaries is also likely to be more difficult, and costly. Thus the aggregate cost of global learning and its associated knowledge creation would be greater for more culturally distant markets than are near. This suggests that MNCs may be expected to exchange knowledge better with those units that are located in countries that are culturally more proximate.

Hypothesis 1 : The cultural similarity between MNCs and host countries will be positively related to MNC subsidiary's global learning

A standard approach to the MNC is to note that market exchange cannot perform certain functions as efficiently as do firms. In this connection numerous authors have successfully used transaction cost ideas to analyze the role of ownership in international business. Through FDI, the MNC appropriates larger returns on its investment in firm-specific assets like knowledge than it would through the market mechanism. Some more complex knowledge which involves a great deal of learning are transmitted more effectively or at a lower cost through more closely related parties (Hejazi and Safarian, 1999). Prior studies also showed that firms frequently relied on majority ownership to achieve effective management control of their foreign subsidiaries' activities (Tomlinson, 1970; Friedman and Beguin, 1971; Stopford and Wells, 1972). They assume that control is achieved primarily through ownership rather than through other management or network mechanisms. Many authors found the control exercised by parents over a venture's activities as a critical determinant of its performance (Geringer and Hebert, 1989). Obviously managers would prefer that valuable knowledge not to get transferred to other firms. High foreign ownership coupled with knowledge gaps reduces knowledge spillovers to local companies (Kokko, 1994).

Hypothesis 2: MNC ownership will be positively related to MNC subsidiary's global learning.

While numerous researchers argued that higher levels of global learning lead to higher MNC performance, the transaction cost theory suggests that global learning involves several specific and substantial transaction costs. Transaction costs associated with global learning can include the costs of additional payroll and overhead, added equipment, costs of administration as well as potential irrelevancies and the opportunity costs associated with having resources committed to global learning. And for effective global learning, a constant search for new opportunities also needs to take place constantly throughout the organization. In order to facilitate the transfer of information and knowledge, standard procedures must be established regarding methods of documentation, standards for blueprints, and a host of other details. In spite of the rapid development and pervasive impact of new communications technologies, efficient exchange of knowledge may have to rely heavily on personal face-to-face communication. The transfer of strategically significant knowledge involves demonstrations, negotiations and common problem solving activities that require direct personal contact. Thus global learning takes substantial administrative, organizational and monitoring support that tend to be only available to large firms.

Hypothesis 3: The size of subsidiary will be positively related to MNC subsidiary's global learning.

Transferring management practices and technological improvements effectively between headquarters and subsidiaries and among subsidiaries may require similarity in processes used and products produced by the global units. An organization's ability to acquire new knowledge depends on the similarity of the information to the existing knowledge base. This concept is known as absorptive capacity (Cohen and Levinthal, 1990). Extending the concept to intra-organizational learning suggests that global learning can take place most successfully among MNC's global units when they operate from the same knowledge base. Thus the degree of similarity in products produced and processes used by internationally dispersed units of MNC will condition the extent and types of global learning that is possible (Simonin, 1999). Similar products and processes also mean that it is easier to communicate among units dispersed around the world. The more similar MNC network's production processes are, for example, the more easily production process improvements discovered at one subunit can be communicated and transferred to other components.

Hypothesis 4: Similarities in products produced and processes used by MNC headquarters and subsidiaries will be positively related to MNC subsidiary's global learning.

Competitive strategy examines the way in which a firm can compete more effectively to strengthen its market position. The central question in competitive strategy involves the development of sustainable competitive advantage (Porter, 1985). Competitive advantage grows fundamentally out of the value a firm is able to create for its buyers (Porter, 1985). As far as global leaning is concerned, the learning subsidiary will try to learn those knowledge and expertise proven to be competitive in the world market. Likewise, the MNC will try to transfer only those knowledge and expertise proven to be competitive in the world market.

Hypothesis 5: MNC competitive advantage will be positively related to MNC subsidiary's global learning.

3. METHODOLOGY

3.1 The Sample and Questionnaire

Questionnaire development proceeded mostly as suggested by Cullen, Johnson and Sakano (1995). Firstly, the academic and trade literature were reviewed. Secondly, relevant existing measures were assembled and bases for new measures were developed. Some existing scales were modified for the research, and new measures were developed in case there were no generally accepted measures. Thirdly, two international business professors reviewed a preliminary version of the questionnaire, it was revised on the basis of their comments. The questionnaire was accompanied by a cover letter explaining the aims of the study, guaranteeing confidentiality, and urging response. In order to improve the response rate, the questionnaire had to be short, concise and of current interest to the respondent. The data for this study were obtained from Present Status of Firms with Investments by Foreigners 2000 published by Foreign Trade and Investment Bureau, Ministry of Industry and Resources, Republic of Korea. The data included the name of the firm, the name of the investing firm, the foreign investor's nationality, the name of the industry in which the firm operates, the founding date of the firm, and the share of foreign investment. Only manufacturing firms were studied to partially control the industry effect. There were 2,817 manufacturing firms with FDIs. Out of them, 1,802 firms had their telephone numbers listed in the above-mentioned directory. From 1,802 firms we selected those firms that received investments from only one foreign country, and made telephone calls to each of them to explain the purpose of the study and ask cooperation for the study. 180 firms expressed the willingness to participate in the study.

The questionnaires were sent to top managers of the sample firms. Respondents were asked to answer all questions in the questionnaire and were assured of the secrecy of the responses. Ten days after the questionnaire was mailed out, a reminder postcard was sent to all companies that had not yet responded. And another ten days after the first reminder was mailed out, another reminder postcard was sent to all companies that had not yet responded. After two reminders, eighty-six completed questionnaires were received. Of these, five were found to be unusable for various reasons, leaving eighty-one valid responses (45.0%) for evaluation. Non-response bias was investigated with the widely used method suggested by Armstrong and Overton (1977). This involved comparing early and late respondents. Late respondents were defined to be those who responded after receiving the reminder postcard2. Three sample measures (sales, number of employees, and equity ratio) were compared using an test of independence. The responses from early and late respondents were virtually identical.

3.2 Variables and Measurement

Our models analyze the degree of global learning undertaken by MNC subsidiaries. There are, however, no generally accepted measures to examine the degree of global learning. Consequently, this study develops seven measures to examine global learning, i.e., scope of global learning, diversity of global learning, diversity of production-related global learning, diversity of non-production-related global learning, intensity of global learning, intensity of production-related global learning, and intensity of non-production-related global learning.

Dependent variables

Scope of global learning (LSCP). LSCP was designed to operationalize the specific scope of global learning among global components of MNCs. Specifically respondents were asked to indicate whether learning occurred among their global components, and whether it occurred between them and headquarters or globally if it occurred. The responses were coded as follows: 1= Do not exchange information including management practices and knowledge with either our parent or its foreign sister subsidiaries. 2= Receive information including management practices and knowledge from our parent and utilize it in our business. 3= Not only receive information including management practices and knowledge from our parent but also transfer those practices and knowledge developed and proven to be effective by us to our parent and the parent utilizes them in its business. 4= Do things mentioned in 3 not only with our parent but also with our foreign sister subsidiaries.

Diversity of global learning (LDIV). LDIV was designed to operationalize the specific contents of learning as being related to competitive advantage. Porter (1985) identified procurement of raw materials, parts and components (LDV1), manufacturing (LDV2), process design and improvements (LDV3), product design and improvements (LDV4), marketing (LDV5), sales (LDV6), finance (LDV7), accounting and legal (LDV8), and human resource management (LDV9) as physically and technologically distinct activities a firm performs. Respondents were asked to rate transfer of information about each of these nine value activities among global subunits on a seven-point scale ranging from 7 "very much" to 1 "not at all". The reliability of the measure of this construct (LDIV) was tested using Cronbach's alpha coefficient. The result (.9064) suggested adequate internal consistency. Diversity of global learning was measured by the averages of the rating of each statement.

Diversity of production-related global learning (LDVM). To develop this measure, I divide the aforementioned nine value activities into two kinds of activities, i.e., production-related activities and non-production-related activities. LDVM represents the mean of LDV1, LDV2, LDV3, and LDV4. These four indicators represent production-related global learning. The reliability of the measure of this construct (LDVM) was tested using Cronbach's alpha coefficient. The result (.8986) suggested adequate internal consistency.

Diversity of non-production-related global learning (LDVO). LDVO represents the mean of LDV5, LDV6, LDV7, LDV8, and LDV9. These five indicators represent non-production-related global learning. The reliability of the measure of this construct (LDVO) was tested using Cronbach's alpha coefficient. The result (.8876) suggested adequate internal consistency. Intensity of global learning (LSLD). LSLD is LSCP times LDIV. LSLD measures the scope and diversity of global learning of MNC units together.

Intensity of production-related global learning (LSLM). LSLM is LSCP times LDVM. LSLM measures the scope and diversity of production-related global learning of MNC units together. Intensity of non-production-related global learning (LSLO). LSLO is LSCP times LDVO. LSLO measures the scope and diversity of non-production-related global learning of MNC units together.

Independent variables

Cultural distance (CLDT). CLDT is a measure of the cultural similarity between South Korea and the investing country. We used two homophily items to operationalize cultural similarity. Our measure allowed the respondents to reflect on the extent to which the attitudinal and behavioral patterns of people of source and host countries were similar (Lin and Germain, 1998). CLDT also is a multiple-item construct. The overall score of CLDT was measured by the averages of the rating of each statement. Each statement concerning CLDT was rated by the respondent on a seven-point scale ranging from 7 "totally agree" to 1 "totally disagree". The reliability of the measure of this construct was tested using Cronbach's alpha coefficient. The result (.8866) suggested adequate internal consistency.

MNC ownership (SHAR). Prior studies showed that firms frequently relied on majority ownership to achieve effective management control of their foreign subsidiaries' activities (Tomlinson, 1970; Friedman and Beguin, 1971; Stopford and Wells, 1972). SHAR is the percentage of equity owned by the MNC Subsidiary size (EMPL). EMPL is measured by the total number of employees employed by the subsidiary. In order to avoid the problems commonly associated with cross-national research such as different accounting methods we followed Gatignon and Anderson's (1988) approach and measured the firm size by the total number of employees employed by the subsidiary.

Similarity of processes and products (PPSM). To measure the degree of similarity of processes used and products produced by MNC units, respondents were asked to rate the degree of similarity on a seven-point scale ranging from 7 "very much" to 1 "not at all".

Competitive advantage (VANT). Four variables (product quality, service quality, R&D expenditures, and advertising expenditures, all relative to its major market competitors) were used to measure competitive advantages. These indicators tap the dimensions underlying the typologies of generic competitive strategies developed by Miles and Snow (1978) and Porter (1980). VANT is a multiple-item construct. The reliability of the measure of this construct was tested using Cronbach's alpha coefficient. The result (.7571) suggested adequate internal consistency. The overall score of VANT was measured by the averages of the rating of each statement. Each statement concerning VANT was rated by the respondent on a seven-point scale ranging from 7 "totally agree" to 1 "totally disagree".

Control variables.

Given the multivariate determinants of global learning, we also sought to control a number of other factors that may have influence on global learning. First, subsidiaries vary in age (SAGE). As time goes by, a subsidiary may become familiar with the host country market. With age come enhanced understanding, competence, and confidence, as well as more accurate perception of foreign risks and returns. Just like a human being, a firm matures as it ages in the local market. And then it may be able to overcome the negative impact of cultural difference through deliberate strategies (Gomez-Mejia and Palich, 1997). That will help a subsidiary share knowledge and expertise more extensively with other MNC units. Second, we include two national dummies, i.e., NAT1 (U.S.) and NAT2 (Japan) to distinguish two countries that made the largest number of investments in Korea and to control whatever effect remained after controlling cultural distance effects. Finally, we included three industry dummies, i.e., IND1 (electronics), IND2 (machinery and metals), and IND3 (chemical products) to distinguish three major Korean industries in which largest number of foreign investments was made.

4. RESULTS

4.1 Sample Characteristics

The average number of employees of the sample firm was 488, and the average sales amount in 1999 was 187.3 billion won, approximately U$146 million. The average asset in 1999 was 188.1 billion won, approximately U$147 million. The sample firm has been in business in Korea for 14 years on average, and the average equity share of the foreign parent was 57.88%. The nationalities of the sample firms include U.S. (23.0%), Japan (44.9%), Germany (4.4%), France (2.1%), Great Britain (2.0%), Switzerland (1.3%), and others (22.3%). Others are two Italian, two Danish, one Swedish, one Belgian, One Taiwanese, and one Saudi Arabian. Industries in which sample firms operate include electricity and electronics (22.8%), machinery and metals (28.6%), chemicals (14.6%), textiles and clothes (6.2%), transportation equipment (5.2%), pharmaceuticals (2.7%), wood products (1.8%), and others (18.4%).

4.2 Estimation of the Model

A correlation analysis was performed to assess the degree of multicollinearity present in the sample data. The correlation matrix (not reported here due to space constraints) suggests that any impact of multicollinearity on the parameter estimates should not be substantial. In Table 1, four models were estimated using ordinary least squares (OLS) regression analysis to identify the relationship between MNC characteristics and global learning. Model 1 summarizes the regression results using LSCP and LDIV as dependent variables. Beginning with LSCP as the dependent variable in model 1, we found support for H1, where we predicted that the cultural similarity between MNCs and host countries would be positively related to MNC subsidiary's global learning. This suggests that the aggregate cost of global learning and its associated knowledge creation would be greater for more culturally distant markets than are near. We also found support for H2, where we predicted that MNC ownership would be positively related to MNC subsidiary's global learning. Meantime, we did not find support for H3, H4, and H5. The coefficient for NAT1 is statistically significant and positive. The regression analysis explained 29.2% of the variance of the dependent variable, and achieved overall significance (p<.01).

The OLS regression analysis results with LDIV as the dependent variable in model 1 support H1, where we predicted that the cultural similarity between MNCs and host countries would be positively related to MNC subsidiary's global learning. We also found support for H4, where we predicted that similarities in products produced and processes used by MNC headquarters and subsidiaries would be positively related to MNC subsidiary's global learning. The regression results also support H5. The coefficient of VANT is positive, and statistically significant. This suggests that competitive advantage is positively related to subsidiary's global learning. Meantime, we did not find support for H2 and H3. The coefficient for NAT1 is statistically significant and positive. The regression analysis explained 32.3% of the variance of the dependent variable, and achieved overall significance (p<.01).

In model 2, I divide LDIV into two kinds of learning. LDVM measures subsidiary's global learning involving value activities related to production (LDV1 to LDV4), while LDVO measures subsidiary's global learning involving value activities related to non-production (LDV5 to LDV9). Model 2 summarizes the regression results using LDVM and LDVO as dependent variables. Beginning with LDVM as the dependent variable in model 2, we found support for H1, where we predicted that the cultural similarity between MNCs and host countries would be positively related to MNC subsidiary's global learning. We also found support for H4, where we predicted that similarities in products produced and processes used by MNC headquarters and subsidiaries would be positively related to MNC subsidiary's global learning. Meantime, we did not find support for H2, H3, and H5. The coefficients for NAT1 and NAT2 are statistically significant and positive. The regression analysis explained 21.0% of the variance of the dependent variable, and achieved overall significance (p<.01).

In model 3, LSLD is LSCP times LDIV. LSLD measures the scope and diversity of global learning of MNC units together. In model 3, we found support for H1, where we predicted that the cultural similarity between MNCs and host countries would be positively related to MNC subsidiary's global learning. We also found support for H2, where we predicted that MNC ownership would be positively related to MNC subsidiary's global learning. However, we did not find support for H3, H4, and H5. The coefficient for NAT1 is statistically significant and positive. The regression analysis explained 37.2% of the variance of the dependent variable, and achieved overall significance (p<.01).

In model 4, LSLM is LSCP times LDVM, and LSLO is LSCP times LDVO. LDVM measures how extensively the subsidiary exchanges knowledge about production-related value activities with other MNC units. LDVO measures how extensively the subsidiary exchanges knowledge about non-production-related value activities with other MNC units. Model 4 summarizes the regression results using LSLM and LSLO as dependent variables. Beginning with LSLM as the dependent variable in model 4, we found support for H1, where we predicted that the cultural similarity between MNCs and host countries would be positively related to MNC subsidiary's global learning. We also found support for H2, where we predicted that MNC ownership would be positively related to MNC subsidiary's global learning. Meantime, we did not find support for H3, H4, and H5. The coefficient for NAT1 is statistically significant and positive. The regression analysis explained 30.0% of the variance of the dependent variable, and achieved overall significance (p<.01).

The OLS regression analysis results with LSLO as the dependent variable in model 4 support H1, where we predicted that the cultural similarity between MNCs and host countries would be positively related to MNC subsidiary's global learning. We also found support for H2, where we predicted that MNC ownership would be positively related to MNC subsidiary's global learning. Meantime, we did not find support for H3, H4, and H5. The coefficient for NAT1 is statistically significant and positive. The regression analysis explained 38.8% of the variance of the dependent variable, and achieved overall significance (p<.01).

5. CONCLUDING REMARKS

Although ideas are formed in the minds of individuals, interaction between individuals typically plays a critical role in developing these ideas (Nonaka, 1994). Likewise communities of interaction contribute to the amplification and development of new knowledge in organizations. In MNCs, these communities for organizational knowledge creation and learning would span national boundaries. Reciprocal knowledge transfer, which is cross transfer of knowledge between units of MNCs, comes with considerable potential for the creation of new knowledge embodied in new competitive products and systems (Bresman, Birkinshaw and Nobel, 1999). As knowledge and information are transferred and learned around a wider community of MNC units, they would be developed and clarified. Gradually, concepts that are thought to be of value would obtain a wider currency and become crystallized. Thus innovation produced by one part of organization in turn creates a stream of related information and knowledge, which might then trigger changes in the organization's wider knowledge systems (Nonaka, 1994).

MNCs need to foster the increasingly important tasks of learning and knowledge transfer. Being able to transfer knowledge from one country to another provides a key source of advantage for MNCs. Good ideas and solutions to problems sought throughout the company and flowed rapidly through the entire organization would surely contribute to the improvement of MNC network's performance. Ford Motor Company which is one of the most global automobile companies, actually saved itself financially during the difficult 1979-1984 period by combining its German subsidiary's special strengths in manufacturing with the British subsidiary's excellent product development and general enterprise management (Yip 1992, p. 183). Global learning is likely to become more important as MNCs adopt increasingly more transnational strategies.

A learning organization actively seeks to move knowledge from one part of the organization to another, to ensure that relevant knowledge finds its way to the organizational unit that needs it most. Learning organizations also spend a lot of energy looking outside their own boundaries for knowledge (Hamel and Prahalad, 1994). Learning is the process of searching for, acquiring, and absorbing both tacit and explicit information into the organization and translating it into knowledge which is then applied to some purpose (Nonaka and Takeuchi, 1995). Learning may facilitate the strategy-environment configuration, improve resource deployment, and enhance the competitive edge of the MNC. In more concrete terms, the long-term effects of global learning of the MNC network may be found in gains in turnover, rationalization of production, improved innovative performance, decreasing costs and finally improved performance for the combined components of the MNC. To the extent that the less easily quantifiable value of the experience gained in one component of the MNC can also be exploited in other components, the MNC network's future performance will be significantly enhanced by being able to avoid repeating costly mistakes. In that case, multinationality itself can be a source of advantage.

Despite its clear importance, however, global learning appears to remain one of the most persistently understudied areas in international business, and so little is known about MNC's global learning. This study attempted to identify firm characteristics related to global learning. The results of this study suggest that cultural proximity, ownership percentage, similarity of processes used and products produced, and competitive advantage are positively related to global learning. The coefficient of CLDT is positive and statistically significant in all seven regression equations. The coefficient of SHAR is positive, and statistically significant in four of seven equations. The coefficients of PPSM, and VANT are statistically significantly positive in two of seven equations. The coefficient of EMPL, the variable representing the size of the MNC, is not statistically significant in any of seven regression equations. That might suggest that the costs of building necessary infrastructure for global learning might not be too great, contrary to our expectations. Rapid advancement in information and transportation technology and subsequent decrease in costs of communication and transportation may have rendered global learning affordable regardless of firm size.

Meantime, there have been no generally accepted measures to examine global learning. Consequently, this study developed seven measures to examine global learning, i.e., scope of global learning, diversity of global learning, diversity of production-related global learning, diversity of non-production-related global learning, intensity of global learning, intensity of production-related global learning, and intensity of non-production-related global learning. As such, this research contributes to collective knowledge on the knowledge-based view of the firm.

Like any multivariate study, the interpretation of our results hinges on how well we control for other factors that might affect MNC's global learning. In this study we have controlled firm-level, industry-level, and national level variables that might affect MNC's global learning. Nevertheless, such controls are necessarily imperfect because our knowledge of the determinants of MNC's global learning is still limited. There are other limitations to this study. First, it had a rather small sample, and thus it may not be representative of MNC subsidiaries in Korea. And that may be responsible for the finding that independent variables are not statistically significant in some equations. And our study was limited to manufacturing industries. The generalizability of the results of our study therefore remains unknown.

In interpreting these results, therefore, some of the limitations of this study should be kept in mind. Many of the proxies used here are also crude, even though they are similar to those used in previous studies. Several measures used in the analysis may need further refinement. Future studies should examine whether the same results can be obtained by using different measures. Furthermore, although multicollinearity does not seem to have been a problem, it might still have inflated the standard errors, leading to low t-statistics. As a result, some of the insignificant coefficients may well be due to the methods and proxies used. The use of some perceptual measures may also introduce possible concern about the subjectivity of response. Thus we may be able to get better results if we as researchers can find better variables in the future.

Learning organizations will not be built overnight. MNC's global learning will be products of carefully cultivated attitudes, commitment, and management processes that will accrue slowly and steadily over time. MNCs that wish to become a learning organization will have to foster an environment that is conducive to global learning by eliminating barriers that impede learning. They need open up boundaries between the components of their global network and stimulate the exchange of ideas among themselves. In short, effective global learning will be facilitated when MNC management makes it an explicit organizational goal. To make the most of global learning, the MNC need be unfettered by bureaucracy, hierarchical thinking, and expensive existing information systems (Liesch and Knight, 1999). They need to adapt their systems, routines, and the collective employee mindset to the imperatives of global learning. Future research should investigate these organizational matters facilitating global learning.

TABLE 1 a, b
MNC CHARACTERISTICS AND GLOBAL LEARNING

              Dependent Variables

              Model 1                 Model 2

Independent
Variables     LSCP        LDIV        LDVM        LDVO

CLDT           .389 ***    .234 **     .195 *      .219 **
SHAR           .338 ***    .118        .023        .117
EMPL           .082       -.040        .031       -.094
PPSM          -.139        .226 *      .263 **     .146
VANT           .055        .229 **     .168        .234 **
SAGE          -.148        .003       -.002        .006
NAT1           .230 *      .247 *      .249 *      .194
NAT2           .053        .186        .249 *      .092
IND1          -.039       -.080        .047       -.174
IND2          -.016       -.050        .016       -.098
IND3           .034       -.115       -.026       -.169
F             3.919 ***   4.387 ***   2.889 ***   3.727 ***
Adjusted      .292         .323        .210        .278
  [R.sup.2]

              Dependent
              Variables

              Model 3     Model 4

Independent
Variables     LSLD        LDLM        LSLO

CLDT           .367 ***    .340 ***    .368 ***
SHAR           .298 ***    .230 **     .337 ***
EMPL           .058        .110        .013
PPSM           .024        .078       -.022
VANT           .091        .063        .108
SAGE          -.108       -.128       -.086
NAT1           .285 **     .289 **     .265 **
NAT2           .081        .101        .060
IND1          -.075        .011       -.140
IND2          -.020        .039       -.068
IND3          -.037        .011       -.075
F             5.192 ***   4.032 ***   5.495 ***
Adjusted       .372        .300        .388
  [R.sup.2]

a. Standardized regression coefficients are shown.

b. *; p<0.1, **; p<0.05, ***; p<0.01

6. REFERENCES

Amit, R. and Shoemaker, P. J. H., "Strategic Assets and Organizational Rent", Strategic Management Journal, 14, 1993, 33-46.

Armstrong, J. S. and Overton, T., "Estimating Non-response Bias in Mail Surveys", Journal of Marketing Research, 51 (7), 1977, 71-86.

Barney, J. B., "Firm Resources and Sustained Competitive Advantage", Journal of Management, 17, 1991, 99-120.

Bresman, H., Birkinshaw, J. and Nobel, R., "Knowledge Transfer in International Acquisitions", Journal of International Business Studies, 30 (3), 1999, 439-462.

Caves, R. E., "Multinational Firms, Competition, and Productivity in Host-country Markets", Economica, 41, 1974, 176-193.

Cohen, W. and Levinthal, D., "Absorptive Capacity: A New Perspective on Learning and Innovation", Administrative Science Quarterly, 35, 1990, 128-152.

Cullen, J. B., Johnson, J. L. and Sakano, T., "Japanese and Local Partner Commitment to IJVs: Psychological Consequences of Outcomes and Investments in the IJV Relationship", Journal of International Business Studies, 26 (1), 1995, 91-115.

Davidson, W. H. and McFetridge, D. G., "Key Characteristics in the Choice of International Technology Transfer Mode", Journal of International Business Studies, 16 (2), 1985, 5-21.

Friedman, W. G. and Beguin, J. P., Joint International Business Ventures in Developing Countries, Columbia University Press, New York, 1971.

Garvin, D. A., "Building A Learning Organization", Harvard Business Review, 71 (4), 1993, 78-91.

Gatignon, H. and 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 (Fall), 1988, 305-336.

Geringer, M. J. and Hebert, L., "Control and Performance of International Joint Ventures", Journal of International Business Studies, 20 (2), 1989, 235-254.

Ghoshal, S., "Global Strategy: An Organizing Framework", Strategic Management Journal, 8, 1987, 425-440.

Globerman, S., "Foreign Direct Investment and 'Spillover' Efficiency Benefits in Canadian Manufacturing Industries", Canadian Journal of Economics, 12 (1), 1979, 42-56.

Gomez-Mejia, L. R. and Palich, L. E., "Cultural Diversity and the Performance of Multinational Firms", Journal of International Business Studies, 28 (2), 1997, 309-335.

Hamel, G,. and Prahalad, C. K., Competing for the Future, Harvard Business Press, Boston, Mass., 1994.

Hedlund, G. and Rolander, D., "Action in Heterarchies- New Approaches to Managing the MNC", In Christopher A. Bartlett, Y. Doz and G. Hedlund, Editors, Managing the Global Firm, Rouledge, London, 1990.

Hejazi, W. and Safarian, A. E., "Trade, Foreign Direct Investment, and R&D Spillovers", Journal of International Business Studies, 33 (3), 1999, 491-511.

Hennart, J. and 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, 515-538.

Jones, G. R. and Hill, C. W. L., "Transaction Cost Analysis of Strategy-structure Choice", Strategic Management Journal, 9 (2), 1988, 159-72.

Kogut, B., "International Sequential Advantages and Network Flexibility", In Christopher A, Bartlett, Y. Doz and G. Hedlund, editors, Managing the Global Firm, Rouledge, London, 1990.

-- and Singh, H., "The Effect of National Culture on the Choice of Entry Modes", Journal of International Business Studies, 19 (3), 1988, 411-432.

Kokko, A., "Technology, Market Characteristics, and Spillovers", Journal of Development Economics, 43 (2), 1994, 279-293.

Liesch, P. W. and Knight, G. A., "Information Internalization and Hurdle Rates in Small and Medium Enterprise Internationalization", Journal of International Business Studies, 30 (2), 1999, 383-94.

Lin, X. and Germain, R., "Sustaining Satisfactory Joint Venture Relationships: The Role of Conflict Resolution Strategy", Journal of International Business Studies, 29 (1), 1998, 179-96.

McGuiness, N. W. and Conway, H. A., "World Product Mandates: The Need for Directed Search Strategies", In H. Etemad and L. S. Dulude, Editors, Managing the Multinational Subsidiary, Croom Helm, London, 1986.

Miles, R. E. and Snow, C. C., Organizational Strategy, Structure, and Process, McGraw Hill, New York, 1978.

Ministry of Industry and Resources, Republic of Korea, Present Status of Firms with Investments by Foreigners 2000, 2000.

Nonaka, I., "A Dynamic Theory of Organizational Knowledge Creation", Organization Science, 5 (1), 1994, 14-37.

-- and Takeuchi, H., The Knowledge-Creating Company, Oxford University Press, New York, 1995.

Ohmae, K., The Borderless World, Fontana, London, 1991.

Peteraf, M. A., "The Cornerstones of Competitive Advantage: A Resource-based View", Strategic Management Journal, 14, 1993, 179-191.

Porter, M. E., Competitive Advantage: Creating and Sustaining Superior Performance, Free Press, N.Y., 1985.

--, Competition in Global Industries, Harvard Business School Press, Boston, Mass., 1986.

-- and Millar, V. E., "How Information Gives You Competitive Advantage", Harvard Business Review, 63 (July-August), 1985, 149-160.

Simonin, B. L., "Transfer of Marketing Know-how in International Strategic Alliances: An Empirical Investigation of the Role and Antecedents of Knowledge Ambiguity", Journal of International Business Studies, 30 (3), 1999, 463-490.

Stopford, J. M. and Wells, L. T., Managing the Multinational Enterprise, Basic Books, New York, 1972.

Tomlinson, J. W. C., The Joint Venture Process in International Business: India and Pakistan, MIT Press, Cambridge, Mass., 1970.

White, R. E. and Poynter, T. A., "Organizing for Worldwide Advantage", In C. A. Bartlett, Y. Doz and G. Hedlund, Editors, Managing the Global Firm, Rouledge, London, 1990.

Yip, G. S., Total Global Strategy, Prentice Hall, Englewood Cliffs, N. J., 1992.

Dr. Jangho Lee earned his Ph.D. at University of Washington in 1996. Currently he is a professor of international business at Sogang University, Seoul, Korea.. This work was supported by the Brain Korea 21 Project in 2003.