How to retain local senior managers in international joint ventures: The effects of alliance relationship characteristics

Available online at www.sciencedirect.com

Journal of Business Research 61 (2008) 986 – 994

How to retain local senior managers in international joint ventures: The effects of alliance relationship characteristics Julie Juan Li ⁎,1 Department of Marketing, City University of Hong Kong, Kowloon, Hong Kong Received 1 October 2006; received in revised form 1 July 2007; accepted 1 December 2007

Abstract How can international joint ventures keep their local senior managers from quitting? Drawing on international business literature, social exchange theory, and social integration theory, this study examines how the characteristics of an alliance relationship affect local senior managers' turnover intentions. The findings from 139 ventures in China show that decision-making participation and social integration help retain local senior managers. However, the effects of participation and social integration decrease with high levels of foreign ownership control. Implications for researchers and practitioners are discussed. © 2007 Elsevier Inc. All rights reserved. Keywords: Turnover intention; Participation in strategic decision making; Social integration; International joint ventures

1. Introduction As the pace of globalization accelerates, firms increasingly establish operations in other countries in the form of international joint ventures (IJVs). Multinational corporations rely heavily on a host country's managerial talent to oversee their subsidiaries abroad, which makes the turnover of local senior managers an acute and chronic problem for IJVs (Krug and Hegarty, 2001), especially considering Wozniak's (2003) report that 43% of senior managers in China voluntarily leave their organizations each year. This frequent turnover of top management team members results in high recruitment costs and can tarnish a firm's corporate image and reputation (Foo et al., 2006). Senior managers' turnover appears even more catastrophic to IJVs in emerging economies such as China, because executive succession can cause fundamental changes in the strategic activities of an IJV, thereby affecting organizational continuity (Boone et al., 2004; Hambrick and Mason, 1984). ⁎ Tel.: +852 2788 7865; fax: +852 2788 9146. E-mail address: [email protected]. 1 This study received support from the CityU Start-up Grant (Project No. 7200065) from City University of Hong Kong. The author thanks two anonymous reviewers for their insightful suggestions. 0148-2963/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jbusres.2007.12.001

Moreover, turnover causes the leakage of important “inside” information to local competitors, or senior managers may take a portion of the firm's portfolio of customers with them when they leave (Gambel, 2000). Hiring replacement managers also becomes increasingly difficult due to the shortage of qualified local managers in emerging economies (Fryxell et al., 2004; Wong and Law, 1999). Despite the significance of maintaining local senior managers for both international business research and multinational corporations (MNCs), little work has focused specifically on the turnover of local senior managers (Reiche, 2007). Thus, an important question remains: How can IJVs retain their local senior managers? Outside the international business area, a significant body of literature from fields such as organizational behavior, human resource management, and applied psychology attempts to explain why people leave (e.g., Huselid, 1995; Michel and Hambrick, 1992; Mobley, 1977; Mossholder et al., 2005). In general, such research identifies several antecedents of employee turnover, namely, individual factors, perceived opportunity, and work-related factors. Although these models improve understanding of employee turnover, existing research appears mainly within the context of domestic firms. In contrast, IJVs are legally and economically distinct entities created by two or more organizations, at least one of which is headquartered in

J.J. Li / Journal of Business Research 61 (2008) 986–994

another country (Johnson et al., 2002; Wong and Ellis, 2002). This cross-border, hybrid form of organization often makes it difficult to manage and retain local managers because of its intercultural and interorganizational nature (Gong et al., 2005; Li et al., 2006). This study therefore takes the viewpoint of the IJV to uncover some unique alliance partner variables related to local senior managers' turnover intentions. In particular, this research examines the effects of two alliance partner relationship characteristics—participation in strategic decision making and social integration with overseas headquarters—on local senior managers' turnover intentions. Furthermore, we propose that their relationships may vary across different levels of ownership control. In this way, we provide an enhanced understanding of the turnover phenomenon by extending research from a domestic model to an international context and focusing on turnover within the local senior management team. We test our hypotheses in the context of IJVs in China, one of the largest recipients of foreign direct investment in the world. In China, IJVs face even greater management challenges due to the complexity, dynamics, and uncertainty associated with the country's transition to a market economy (Child and Mollering, 2003; Li, 2005). Therefore, IJVs in China provide a rich setting for investigating how to keep local senior managers from quitting. 2. Theoretical background and hypotheses 2.1. Turnover and turnover research Voluntary turnover incidents are those “wherein management agrees that the employee had the physical opportunity to continue employment with the company, at the time of termination” (Maertz and Campion, 1998, p. 50). The phenomenon of voluntary employee turnover attracts enormous attention in the fields of organizational behavior and applied psychology, and the vast associated research efforts offer several models to explain employee turnover behavior (e.g., Mobley, 1977; Mossholder et al., 2005; Wayne et al., 1997). Researchers identify three major classes of antecedents of employee turnover: individual factors such as age and tenure, economic opportunities such as job alternatives, and workrelated factors such as task repetitiveness and job challenge (cf. Griffeth et al., 2000; Maertz and Campion, 1998). Recent growing interest focuses on the turnover of a special category of employees: senior managers. As Michel and Hambrick (1992) and Wiersema and Bantel (1993) define it, the senior management team refers to the top two tiers of an organization's management, including (1) the CEO and general managers, as well as (2) those who have direct reporting relationships with these positions. Turnover of senior managers represents a significant event for any organization and remains a crucial managerial challenge (Foo et al., 2006; O'Reilly et al., 1989). Existing studies in this research stream provide some explanations of why top managers quit. For example, drawing on the demography theory, several studies of top management teams associate team demographic heterogeneity, in terms of age, education, and team tenure, with turnover intentions (e.g.,

987

Boone et al., 2004; Smith et al., 1994). Herrbach et al. (2004) find that managers' perceived external prestige directly influences their intentions to quit. Yet other explanations of managers' turnover include strategic change, low salary, limited advancement opportunities, and limited recognition of firm performance (Smith et al., 1994; Wiersema and Bantel, 1993). Despite the proliferation of research on senior managers' turnover in domestic firms , turnover in an international context remains underresearched (Reiche, 2007). IJVs are legally and economically distinct entities created by two or more organizations, at least one of which is headquartered in another country (Johnson et al., 2002). This hybrid, intercultural, interorganizational form of organization often suffers from high manager turnover rates, which are catastrophic to the venture's operation and performance (Wozniak, 2003; Zhang et al., 2006). Although insights from domestic turnover research provide a useful framework for international research, various venturerelated characteristics must be considered specifically in an international business context (Krug and Hegary, 2001; Naumann, 1992). Among the limited studies on turnover in MNCs, Krug and Hegary (2001) show that executives' perceptions of merger announcements and interactions with the acquiring firm's top managers influence their decision to leave. Zhang et al. (2006) find that local senior managers' perceptions of local staff incompetence positively affect their turnover intentions, but those intentions decrease with better communication quality with headquarters and more trust in headquarter management. Most recently, Reiche (2007) proposes that international staffing practice affects subsidiary staff retention through the local staff's perceived career prospects and organizational identification. Although these studies incorporate some international aspects into turnover research, less examined is how alliance-specific variables, such as shared strategic decision making or social integration between local senior managers and overseas headquarters, affect senior manager turnover. Nor does any research examine the role of the contingency effects of alliance partner characteristics, such as ownership control. Extending previous work, this study examines local senior manager turnover in the international context of IJVs. In addition to related literature in the area of international business, it draws on social integration theory and social exchange theory to posit that higher levels of participation in decision making and social integration lead to lower turnover intentions among local senior managers. Moreover, ownership control moderates such effects. 2.2. Participation in strategic decision making Social exchange theory describes how people form and maintain relationships in a social context and posits, as its key tenet, that human behavior reflects an exchange of rewards (Blau, 1964; Homans, 1961). According to this theory, employees feel obligated to remain with the company to reciprocate gestures of goodwill from their employers (Mossholder et al., 2005). For example, when MNCs localize their operations in the host country, it signifies their commitment to

988

J.J. Li / Journal of Business Research 61 (2008) 986–994

the local market and local employees (Child and Mollering, 2003). One of the specific objectives of localization is to replace expatriate managers with local managers and delegate decisionmaking authority to these local managers (Fryxell et al., 2004; Wong and Law, 1999). Involving local senior managers in strategic decision making signifies overseas headquarters' willingness to increase its vulnerability by relying on local talent (Li et al., 2006). Therefore, local senior managers likely perceive their participation in the IJVs strategic decision making as a way to meet their higher-order needs, such as for trust, respect, independence, and equality (Miller and Monge, 1986). As a result, local senior managers stay with the organization to reciprocate the goodwill and power implied by their decisionmaking rights, as endorsed by the IJV. In addition, participation and involvement promote work motivation and increase employees' satisfaction, which help keep them from leaving the organization. When managers participate more in alliance decision making, they feel they are important contributors to the organization and gain a sense of achievement (Jun et al., 2001). This increased level of participation sends a signal to local managers about their status as insiders, which then increases their job satisfaction and lowers their turnover intentions. Previous work on work-involvement practices also suggests that high-involvement work practices decrease employee turnover intentions (Huselid, 1995). H1: The higher their level of participation in strategic decision making, the lower are local top managers' turnover intentions. 2.3. Social integration Formed by two or more organizations from different countries, IJVs are described as “organizational marriages” (Wong and Ellis, 2002). Because the local partner and overseas headquarters are located in different nations, significant cultural diversity usually exists. These differences in organizational practices, values, and norms further complicate the relationship between the local partner and the foreign parent. In addition, local managers may represent a social segment that is vulnerable to any alliance partner conflict that stems from different cultural and organizational backgrounds (Fryxell et al., 2004; Zhang et al., 2006). The multifaceted concept of social integration reflects social interaction, group pride, and coordination (Krug and Hegarty, 2001; Michel and Hambrick, 1992). Social integration between local managers and overseas headquarters gives local senior managers the opportunity to build close relationships and a sense of shared identification (Dutton et al., 1994). If local executives get along well with the overseas headquarters and feel they are an integral part of the multinational corporation's network, they are less likely to think about withdrawing from the IJV (Zhang et al., 2006). Moreover, socially integrated parties become psychologically linked in their pursuit of a common objective (Foo et al., 2006). Therefore, a high level of social integration may engender cohesiveness and decrease turnover (O'Reilly et al., 1989). Furthermore, members of socially integrated groups experience higher morale and

satisfaction, which also lead to lower turnover intentions (O'Reilly et al., 1989; Smith et al., 1994). Thus, a high level of social integration acts as both a lubricating and a bonding mechanism that increases local senior managers' attachment and loyalty to the organization (Naumann, 1992). H2: The higher their level of social integration, the lower are local senior managers' turnover intentions. 2.4. Moderating role of ownership control Control in the context of IJVs is the ability to exercise authority and influence over the IJVs strategic and operational decisions, systems, and methods (Anderson and Gatignon, 1986). As a crucial coordinating mechanism in IJVs, control relates to alliance outcomes such as partner conflict, bargaining power, and firm performance (e.g., Barden et al., 2005; Gong et al., 2005). Existing IJV literature identifies two dimensions of control: ownership and management (Barden et al., 2005; Steensma and Lyles, 2000). Ownership control refers to “the residual rights to make decisions regarding an asset's use that is not contractually given to another party” (Steensma and Lyles, 2000, p. 833). The percentage of equity stakes each parent holds in the venture manifests ownership control (Geringer and Hebert, 1989). In contrast, management control is the observable pattern of decision-making power (Steensma and Lyles, 2000), that the partners' contributions of capital and other resources influence, such as technology, know-how, and market access (Yan and Gray, 1994). Although partners may have greater management control than their portion of equity would suggest (Madhok, 2006), empirical evidence shows that in practice, ownership control leads to management control in emerging economies, particularly through board membership (Child et al., 1997; Child and Faulkner, 1998). Thus, equity shareholdings represent a key factor that provides a basis for control in strategic alliances, and the extent of equity ownership can reflect the extent of management control (Child and Faulkner, 1998; Dhanaraj and Beamish, 2004). For example, a majority equity ownership offers dominant control, which gives firms a decisive say over decision making (Madhok, 2006). Because equity signifies the level of influence a partner can exercise in the IJV, equity share receives significant attention in most studies of alliance control (Child and Faulkner, 1998; Dhanaraj and Beamish, 2004; Li et al., 2006). Therefore, this study focuses on the ownership control of an alliance partner, as manifested by the partners' equity investments in an IJV. When foreign ownership is lower, the local partner exercises greater authority over the IJV, and the local senior managers enjoy the privilege of serving as decision makers for a variety of strategic issues. Such participation in strategic decision making may deter quitting. However, in conditions of high foreign ownership, overseas headquarters have more bargaining power and a stronger influence over the IJV's strategic decisions and probably are less willing to share authority (Lu and Hebert, 2005). Consequently, local partners yield to the foreign parent's wishes and relinquish some autonomy over their decisions,

J.J. Li / Journal of Business Research 61 (2008) 986–994

which restricts the scope and latitude of their decision making (Doh et al., 2004; Madhok, 2006). If the local senior managers experience constraints on their decision-making latitude, they enjoy less authority and freedom in alliance-specific strategic decision making. Thus, a higher level of foreign ownership control may weaken the effect of participation in strategic decision making on turnover intentions. H3a: The effect of participation in strategic decision making on lowering local senior managers' turnover intentions becomes weaker when foreign ownership control is high rather than low. Also when foreign ownership is high, the effect of social integration in terms of lowering turnover intentions also may decline. A high level of equity ownership by a foreign parent reflects administrative responsibilities (Doh et al., 2004; Lu and Hebert, 2005), which enables the overseas headquarters to bring in more Western managerial practices, organizational cultures, and values. Because Chinese local managers often are unfamiliar with modern management, marketing, financing, and international business practices (Fryxell et al., 2004; Li et al., 2006), they may feel incompetent and isolated. Moreover, a high level of control held by headquarters engenders a distinction between managers in superior positions (i.e., in the overseas headquarters) and those in subordinate positions (i.e., local senior managers) (Johnson et al., 2002). Local senior managers thus may feel that they have lower status and are treated less as MNC senior managers and more as local staff. Therefore, when the levels of foreign investment are higher, social integration may become less effective in deterring local manager's turnover. H3b: The effect of social integration on decreasing local senior managers' turnover intentions becomes weaker when foreign ownership control is high rather than low. 3. Method 3.1. Sample and data collection The hypotheses testing uses data collected from a sample of IJV local senior managers in China. As the world's largest emerging economy, China's rapid economic growth has created huge opportunities for foreign companies and made it one of the largest recipients of foreign direct investment in recent years. From January to July 2006, the number of newly approved foreign-invested enterprises reached 22,772, and actual foreign investments were US$32.707 billion. However, many ventures suffer from management problems, and some fail because of the high turnover rate among local senior managers, who act as bridges to connect foreign firms with local business networks and help the ventures adapt to the local environment (Wong and Law, 1999). A critical challenge for foreign investors therefore is to understand how to keep their local managers from quitting. From a database of foreign-invested enterprises in China, we randomly selected 350 IJVs operating in consumer durable and nondurable product categories, such as beverages, home appliances, and IT products. These ventures are located in several major cities in China (i.e., Beijing, Shanghai, Guangz-

989

hou, and Shenzhen). Telephone calls helped identify a senior manager who held a title such as general manager or regional manager as the respondent for each IJV. Further screening of this respondent ensured that he or she was committed to cooperating with the research. The data collection involved collaboration with local researchers, because in an emerging economy, such collaboration provides a key means to obtain reliable and valid information (Zhou et al., 2005). In addition, face-to-face onsite interviews provide the means to gather information, because mail and telephone surveys likely would incur a poor response rate. Most interviews occurred in the respondents' offices. The interviewer, a trained research assistant, informed respondents in advance of the confidentiality of their responses and provided an official university letter from the researchers to explain the academic purpose of the project. To encourage objective and truthful responses, the interviewer also told respondents that the completed questionnaires would be sent directly to researchers and no one in their organizations would see them. Each respondent received a valuable gift and the promise of a summary report of the survey. These efforts resulted in 139 complete surveys, representing a response rate of 39%. 3.2. Measures The questionnaire, prepared in English, underwent professional translation into Chinese and then, to ensure conceptual equivalence, subsequent back-translation into English. We pretest the questionnaire items with a sample of 10 senior managers and, on the basis of their responses, made several revisions to ensure the questionnaire items' clarity. All the perceptual scales use a seven-point Likert scale (1 = strongly disagree; 7 = strongly agree). The measurement items and the results of validity analyses appear in Appendix A. The measure of local senior managers' turnover intentions comes from Wayne et al.'s (1997) scale and uses three questions to gather information about local senior managers' tendency to leave the organization (Cronbach's α = .86). We develop three items to measure participation in decision making on the basis of our literature review and in-depth interviews. Example items include, “The Chinese partner of this brand participates little in the brand's strategy and positioning decisions” (reverse coded) and “In this joint operation, the Chinese partner is only responsible for production, and not brand strategy” (reverse coded) (Cronbach's α = .78). An adapted four-item scale from Smith et al. (1994) measures local senior managers' social integration with their overseas headquarters by asking respondents to assess their level of social interaction and cooperation with overseas headquarters (Cronbach's α = .93). Ownership control is measured by the percentage of the foreign parent company's equity share in the joint venture. 3.2.1. Control variables The analysis also includes controls for several factors, such as the individual-level variables of the senior managers' age, gender, and education. Respondents indicate their age ranges: below 26, between 26 and 30, between 31 and 35, between 36

990

J.J. Li / Journal of Business Research 61 (2008) 986–994

and 40, between 41 and 45, or 46 years of age and older. Gender uses a dummy variable (0 = female; 1 = male). Finally, respondents indicate their educational level: elementary school, junior high, senior high, community college, college graduate, master's degree, or doctoral degree. At the organizational level, the controls focus on the effects of industry type, firm history, firm size, and firm performance. Sales, obtained from secondary archival data, measure firm performance. Firm size is the logarithm of the number of employees, and firm history is the number of years since the IJV first was established in China. Industry type entails a dummy variable (0 = nondurable products; 1 = durable products). 3.2.2. Common method assessment Because the perceptual data come from the same respondents in the same survey, common method variance is a potential problem. A Harman one-factor test serves to check this potential (Podsakoff and Organ, 1986). A factor analysis of the 10 items from the three perceptual measures results in a three-factor solution, as expected, that accounts for 78.6% of the total variance; factor 1 accounts for 36.6% of the variance. Because a single factor does not emerge and factor 1 does not explain most of the variance, common method bias is unlikely to be a concern in the data. 3.2.3. Construct validity Refining the measures and assessing their construct validity proceeds as follows: First, we run exploratory factor analyses for the latent constructs and find theoretically expected factor solutions. Second, we run reliability analyses for each construct to determine if the measures demonstrate satisfactory coefficient reliability. Third, we conduct confirmatory factor analyses (CFA) to assess the convergent validity of the measures with latent variable structural equation modeling. A test of the threefactor model (turnover intentions, participation in strategic decision making, and social integration) yields a good fit (χ2(32) = 58.03, p = .003; goodness-of-fit index [GFI] = .92; root mean square error of approximation [RMSEA] = .07). The comparative fit index (CFI; Bentler, 1990) and incremental fit index (IFI: Bollen, 1989) also provide indicators of overall model fit, because they represent the best approximation of the

population value. Both the CFI and IFI are .97. All factor loadings are significant, indicating the unidimensionality of the constructs. Furthermore, the composite reliabilities of the latent constructs are .87, .79, and .93, greater than the .70 benchmark. In summary, the results indicate that the measures demonstrate adequate convergent validity and reliability (Fornell and Larcker, 1981). We further assess the discriminant validity of the three latent constructs with chi-square difference tests. We test all construct pairs (three tests together) to determine whether the restricted model (correlation fixed to 1) is significantly worse than the freely estimated model (correlation estimated freely). All the chisquare differences are highly significant, in support of discriminant validity (Anderson and Gerbing, 1988). For example, the comparison of turnover intentions and participation yields a χ2(1) = 14.14 (p b .01), the comparison of turnover intentions and social integration yields a χ2(1) = 13.74 (p b .01), and that of participation and social integration yields a χ2(1) = 7.41 (p b .01). Taken together, these results show that the measures possess adequate reliability and validity (Anderson and Gerbing, 1988). The results of the CFA, including the GFI, factor loadings, and composite reliability, appear in Appendix A. Table 1 presents the basic descriptive statistics and correlations of the measures. As expected, participation in strategic decision making and social integration are negatively and significantly related to turnover intentions. 4. Analyses and results We apply multiple moderated regression analyses to test our hypotheses and use the product term between the pertinent latent constructs to test the interaction effects. Because product terms can include collinearity, we mean-center the variables before constructing the terms (Aiken and West, 1991). We further check for potential collinearity among the variables by calculating the variance inflation factors (VIFs) associated with each of the predictors in our models. The value of the VIFs range 1.03 from 1.43, with a mean of 1.26, indicating no problem with collinearity in the analysis. Table 2 includes the regression results based on the mean-centered latent explanatory variables.

Table 1 Means, standard deviations, and correlations

1. Turnover intention 2. Participation 3. Social integration 4. Ownership control 5. Age 6. Gender 7. Education 8. Industry type 9. Firm history 10. Firm size 11. Firm performance

Mean

s. d.

1

2

3

4

5

6

7

8

9

10

11

3.08 4.67 5.23 63.08 3.36 0.77 4.48 0.42 8.26 6.03 63,792

1.15 1.43 1.15 21.38 1.60 0.42 0.83 0.50 8.48 1.40 147,168

1.00 −0.32 −0.40 −0.04 −0.10 −0.16 −0.03 −0.04 −0.04 0.01 0.09

1.00 0.35 − 0.15 0.24 0.12 − 0.08 − 0.20 0.09 0.03 − 0.16

1.00 − 0.17 0.00 − 0.03 0.01 0.10 0.06 − 0.03 0.06

1.00 − 0.11 − 0.18 0.06 0.13 0.09 − 0.03 − 0.01

1.00 0.31 − 0.27 0.01 − 0.13 − 0.21 − 0.12

1.00 − 0.10 − 0.02 − 0.10 0.08 − 0.07

1.00 − 0.10 0.08 0.19 0.10

1.00 −0.20 0.13 0.03

1.00 0.27 0.04

1.00 0.05

1.00

Notes: N = 139; p b .01 (2-tailed) for r greater than .24; p b .05 (2-tailed) for r greater than .16.

J.J. Li / Journal of Business Research 61 (2008) 986–994 Table 2 Standardized coefficient estimates: hierarchical moderated regressions Variables

Control variables Age Gender Education Industry type Firm history Firm size Firm performance Main effects Ownership control H1: Participation in decision making H2: Social integration Interactions H3a: Control × participation H4b: Control × social integration R2 Model F Significance df

4.2. Post-hoc analysis

Trust Model 1

Model 2

Model 3

−0.06 −0.16+ −0.08 −0.08 −0.09 0.05 0.08

−0.02 −0.15+ −0.08 −0.08 −0.05 0.04 0.07

− 0.02 − 0.12 − 0.04 − 0.03 − 0.04 0.05 0.07

−0.02 −0.19⁎ −0.33⁎⁎

− 0.03 − 0.19⁎ − 0.34⁎⁎

.236 3.958 b.001 10, 128

0.14⁎ 0.17⁎ .298 4.449 b.001 12, 126

.050 .984 .445 7, 131

991

p b .10 ,⁎ p b .05, ⁎⁎p b .01.

+

4.1. Hypotheses tests H1 pertains to the effect of participation in strategic decision making on the turnover intentions of local senior managers. As Table 2 (model 3) shows, and as predicted in H1, participation relates negatively to turnover intentions (β = .− 19, p b .05). Furthermore, social integration has a negative effect on top managers' turnover intentions (β = − .34, p b .001), in support of H2. Turning to the contingency effects of ownership control, H3a posits that ownership control moderates the relationship between participation and turnover intentions, such that the negative effect of participation on turnover is weaker when the foreign parent's ownership control is high. As Table 2 shows, the interaction between participation and ownership control is associated positively with turnover intentions (β = .14, p b .05). To further illustrate the nature of this interaction, Fig. 1, following convention (Aiken and West, 1991), depicts the pattern of the interaction and displays the negative relationship between participation in decision making and turnover intentions when foreign ownership control is low, in contrast to the lack of a trend among those reporting higher ownership control. Therefore, H3a receives full support. H3b predicts that ownership control moderates the relationship between social integration and turnover intentions; such that the relationship becomes weaker when foreign ownership control is high. As Table 2 shows, the two-way interaction effect (social integration × ownership control) is significant for turnover intentions (β = .17, p b .05). Fig. 1b displays this two-way interaction. It shows a negative relationship between social integration and turnover intentions when foreign ownership control is low. However, this relationship is less pronounced for high foreign ownership control. These results support H3b.

This study uses equity shares to measure ownership control, but equity does not always translate into control, especially in emerging economies in which the partners come from different sociopolitical cultural contexts (Muthusamy and White, 2006). In emerging economies, the local partner may enjoy a control advantage due to protection from the local government or institutional influence over IJV activities (Barden et al., 2005). Therefore, to test whether the results using equity shares are robust, we conduct additional analysis using management control as an indicator of control. Consistent with Steensma and Lyles (2000), we measure management control as the arithmetic average of the degree of influence (percentage) by the foreign parent in each of the following areas: financing, accounting, product development, pricing, sales/marketing, distribution, administrative support, and management decisions. These areas represent major specific aspects of control over IJVs (Child et al., 1997; Killing, 1983; Yan and Gray, 1994). The correlation between equity share and managerial control is .51, consistent with previous arguments that ownership control and management control are linked (Child and Faulker, 1998; Steensma and Lyles, 2000). A reestimation of model 3 using management control reveals that the interaction between participation and management control is significant and positive (β = .15, p b .05); the interaction between social integration and ownership control is also significantly positive (β = .19, p b .05). These results are highly consistent with the results reported in Table 2, suggesting the robustness of the findings.

Fig. 1. a. Interaction between participation in decision making and foreign ownership control on turnover intentions, b. interaction between social integration and foreign ownership control on turnover intentions.

992

J.J. Li / Journal of Business Research 61 (2008) 986–994

5. Discussion The issue of turnover has particular importance as local senior managers have become IJVs most valuable assets and often the antecedents of competitive advantage. Because these local managers possess more connections in the local business environment and more knowledge about local markets (Fryxell et al., 2004; Wong and Law, 1999), their turnover is an acute problem from which IJVs suffer. Thus, one challenge foreign investors continue to face is keeping local senior managers with the organization. In an attempt to provide some insights into this intriguing issue, this study examines the relationship between alliance variables and turnover in IJV, which receives only limited attention in extant literature. The investigation of the direct effects of two alliance-specific variables (i.e., participation in ventures' strategic decision making and social integration with overseas headquarters) and the moderating effect of ownership control provides strong support for the proposed relationships. Both participation in decision making and social integration have direct and negative impacts on local managers' turnover intentions. Therefore, the more local managers are involved in decision making, the less likely they will withdraw from the IJV. In addition, local managers have lower turnover intentions when they feel they are integral parts of the organization. Previous studies on domestic turnover identify individual factors, economic opportunities, and work-related factors as antecedents of employee turnover, and in adopting an international perspective, a few studies consider the value of alliance-related characteristics in staff turnover (e.g., Reiche, 2007; Zhang et al., 2006). Joining the latter group, these findings confirm the important role of alliance-specific variables in reducing local managers' intent to leave. The other interesting finding pertains to the moderating effect of ownership control. Apparently, high equity holdings by the foreign partner weaken the effect of participation in decision making and social integration on turnover intention. In summary, these findings represent an important step and help set an agenda for further research by demonstrating the potential integration of alliance-specific variables and turnover. The findings also provide several implications for researchers and practitioners. First, IJVs remain under the watch of both the overseas headquarters and the local partner. Their hybrid nature highlights the importance of decision-making authority in IJVs. Understanding the role of participation in strategic decision making as a means to deter managerial turnover remains limited, but this study suggests that a greater degree of participation by local senior managers will enhance their feelings of trust, respect, status, and satisfaction, which in turn will result in lower turnover intentions. According to social exchange theory, people's decisions to remain in or terminate relationships depend on their evaluations of the relationship's costs and benefits. If participation in IJV decision making helps meet local senior managers' higher-order needs, their withdrawal from the organization may entail a psychological loss that would make turnover personally costly. Second, consistent with previous studies (e.g., Foo et al., 2006; Krug and Hegarty, 2001; O'Reilly et al., 1989), social

integration plays an important role in influencing organizational outcomes and members' behavior. As a result of cultural and organizational differences, local senior managers may feel like less than integral parts of a corporation, but greater social integration decreases their turnover intentions. Social integration theory also suggests that members in socially integrated organizations experience higher satisfaction and maintain higher morale (Smith et al., 1994), which builds their loyalty and lowers their turnover intentions. Consistent with this logic, the findings suggest that without a sense of inclusion, local senior managers may be more likely to think about withdrawing from the company. Third, the study demonstrates a moderating effect of ownership control. Although its role in facilitating coordination and boosting performance receives widespread recognition and empirical support (e.g., Doh et al., 2004; Killing, 1983), existing literature also points to some negative effects in an alliance. For example, Hambrick and Cannella (1993) suggest increased control may deteriorate the relationship between executives from two companies. The current findings add to the evidence about the negative side of ownership control by highlighting that a high degree of ownership control decreases the effects of participation and social integration on managers' turnover intentions. Thus, control appears to be a double-edged sword that may be both beneficial and detrimental to a firm simultaneously. Further research should continue to scrutinize these effects of control structures in alliances. The findings also have important practical implications for MNCs operating in foreign countries. Overseas headquarters can initiate effective strategies to retain local senior managers, such as involving local senior managers more in the IJV's strategic decision making. It is also important for overseas headquarters to implement social integration programs to make local managers feel they represent an integral part of the corporation, rather than “just the locals.” With regard to the other major finding, that high ownership control weakens the effects of participation and social integration on turnover intentions, IJVs that want to increase ownership control to minimize the coordination costs of shared management should pay special attention the downside of this increased ownership control. Finally, this study contains several limitations that suggest directions for further research. First, the data collection procedure offers monetary incentives. Although the use of monetary incentives is a common practice, and evidence indicates that incentives strongly and consistently increase response rate (e.g., Armstrong and Yokum, 1994; Jobber et al., 2004), monetary incentives may lead to favorable answers. Previous research on the favorability of responses as a function of incentives is equivocal (James and Bolstein, 1990). To encourage objective and truthful responses, we informed all respondents in advance of the academic purpose of the project, the confidentiality of their responses, and that their responses would be used only in aggregated analyses. In addition, the interaction effects (H3a and H3b) are less likely affected by such biases because respondents probably do not have an interaction-based logic in mind that could bias their responses

J.J. Li / Journal of Business Research 61 (2008) 986–994

and produce interaction results (Aiken and West, 1991). Because both H3a and H3b receive support, such bias is less of a concern, but further research should consider the possible favorability when using incentives. Second, Chinese people care very much about face and tend to avoid circumstances in which they could lose face, so they may modify their responses to save face. To alleviate this potential bias, we employ some preventive measures in the data collection procedure. However, we also realize that the consideration of face may still create some social desirability bias and we encourage further research to refine our measures. Third, the proposed model examines two alliance-related variables in relation to turnover intentions. Thus, the constructs included in this study provide only a portion of the potentially relevant variables that might have been included. Additional research should expand this framework to identify more alliance-related factors that may affect local managers' turnover. Fourth, this study uses turnover intentions instead of actual turnover as the focal variable. Price and Mueller (1981) recommend this usage, because so many external factors affect actual behavior, making actual turnover very difficult to predict. Although turnover intention is an appropriate dependent variable, it does not absolutely lead to actual turnover. Therefore, further research should validate the findings by examining actual turnover among local senior managers in IJVs. 6. Conclusion In conclusion, this research aims to increase understanding of why local managers leave and how to reduce their turnover intentions. The study identifies two IJV-related factors (i.e., participation in ventures' strategic decision making and social integration with overseas headquarters) that influence managers' turnover. In addition, ownership control plays a moderating role in managers' turnover intentions. The results highlight the importance of specific alliance variables that may influence local managers' turnover intentions in IJVs and further suggest that IJVs effectively can deter quitting among local managers by involving managers more in strategic decision making and implementing social integration programs. In addition, high ownership control appears to weaken the effects of participation and social integration. Although this study has several limitations, the study provides new insights into how to deter local manager turnover and thus makes important contributions to turnover and international human resource literature. We hope this study helps stimulate further efforts to develop an integrative model of manager turnover in IJVs. Appendix A: Measurement items and validity assessment

Model : χ2(32) = 58.03, p = .003; GFI = .92, CFI = .97, IFI = .97; RMSEA = .07 Turnover intention: composite reliability = .87 I often think about quitting this job. I shall probably look for a new job soon.

Standardized Loading .94 .88

993

Appendix A (continued) I cannot imagine that I shall work for this organization for a .65 long time. Participation in strategic decision making: composite reliability = .79 The Chinese partner of this brand participates little in the .67 brand's strategy and positioning strategies. (reverse coded) In this joint operation, the Chinese partner is only responsible .82 for production, not brand strategy. (reverse coded) Good suggestions on the brand from the Chinese partner are .79 often not accepted. (reverse coded) Social integration: composite reliability = .93 We do not get along with the overseas headquarter. (reverse coded) We do not have a good working relationship with the overseas headquarter. (reverse coded) There is a great deal of competition between local senior managers and the overseas headquarter. (reverse coded) We are not always ready to cooperate and help the overseas headquarter. (reverse coded) Model : χ2(32) = 58.03, p = .003; GFI = .92, CFI = .97, IFI = .97; RMSEA = .07

.90 .94 .81 .88 Standardized Loading

References Aiken LS, West SG. Multiple regressions: testing and interpreting interactions. Newbury Park (CA): Sage; 1991. Anderson E, Gatignon H. Models of entry: a transaction cost analysis and propositions. J Int Bus Stud 1986;17(3):1–26. Anderson JC, Gerbing DW. Structural equation modeling in practice: a review of recommended two-step approach. Psychol Bull 1988;103(3):411–23. Armstrong JS, Yokum JT. Effectiveness of monetary incentives: mail survey to members of multinational professional groups. Ind Mark Manage 1994;23: 133–6. Barden JQ, Steensma HK, Lyles MA. The influence of parent control structure on parent conflict in Vietnamese international joint ventures: an organizational justice-based contingency approach. J Int Bus Stud 2005;36:156–74. Bentler PM. Comparative fit indexes in structural models. Psychol Bull 1990;107: 238–46. Blau PM. Exchange and power in social life. New York: Wiley; 1964. Bollen KA. A new incremental fit index for general structural models. Sociol Methods Res 1989;17:303–16. Boone C, Olffen WV, Witteloostuijn AV, Brabander ED. The genesis of top management team diversity: selective turnover among top management teams in Dutch newspaper publishing, 1970–94. Acad Manage J 2004;47(5): 633–56. Child J, Faulkner D. Strategies of cooperation: managing alliances, networks, and joint ventures. New York: Oxford University Press; 1998. Child J, Mollering G. Contextual confidence and active trust development in the Chinese business environment. Org Sci 2003;14(1):69–80. Child J, Yan A, Lu Y. Ownership and control in Sino-foreign joint ventures. In: Beamish PW, Killing JP, editors. Cooperative strategies: Asia pacific perspectives. San Francisco: New Lexington Press; 1997. p. 181–225. Dhanaraj C, Beamish P. Effect of equity ownership on the survival of international joint ventures. Strat Manage J 2004;25:295–305. Doh JP, Teegen H, Mudambi R. Balancing private and state ownership in emerging markets' telecommunications infrastructure: country, industry, and firm influences. J Int Bus Stud 2004;35:233–50. Dutton JE, Dukerich JM, Harquail CV. Organizational images and member identification. Admin Sci Q 1994;39:239–63. Foo M, Sin H, Yiong L. Effects of team inputs and intrateam processes on perceptions of team viability and member satisfaction in nascent ventures. Strat Manage J 2006;27:389–99. Fornell C, Larcker DF. Evaluating structural equation models with unobservable variables and measurement error. J Mark Res 1981;18(February):39–50.

994

J.J. Li / Journal of Business Research 61 (2008) 986–994

Fryxell GE, Butler J, Choi A. Successful localization programs in China: an important element in strategy implementation. J World Bus 2004;39:268–82. Gambel J. Localizing management in foreign-invested enterprises in China: practical, cultural, and strategic perspectives. Int J Hum Resour 2000;11(5): 883–903.f. Geringer JM, Hebert L. Control and performance of international joint ventures. J Int Bus Stud 1989;20(2):97–101. Gong Y, Shenkar O, Lou Y, Nyaw M. Human resources and international joint venture performance: a system perspective. J Int Bus Stud 2005;36:505–18. Griffeth RW, Hom PW, Gaertner S. A meta-analysis of antecedents and correlates of employee turnover: update, moderator tests, and research implications for the next millennium. J Manage 2000;26(3):463–88. Hambrick DE, Cannella AA. Relative standing: a framework for understanding departures of aquired executives. Acad Manage J 1993;36:733–62. Hambrick DE, Mason PA. Upper echelons: the organization as a reflection of its top managers. Acad Manage Rev 1984;9:193–206. Herrbach O, Mignonac K, Catigon AL. Exploring the roles of perceived external prestige in managers' turnover intentions. J Int Bus Stud 2004;15(8):1390–407. Homans G. Social behavior. New York: Harcourt, Brace & World; 1961. Huselid MA. The impact of human resource management practices on turnover, productivity, and corporate financial performance. Acad Mange J 1995;38: 635–72. James JM, Bolstein R. The effect of monetary incentives and follow-up mailings on the response rate and response quality in mail surveys. Public Opin Q 1990;54(3): 346–61. Jobber D, Saunders J, Mitchell VW. Prepaid monetary incentive effects on mail survey response. J Bus Res 2004;57:347–50. Johnson JP, Korsgaard MA, Sapienza HJ. Perceived fairness, decision control, and commitment in international joint venture management teams. Strat Manage J 2002;23:1141–60. Jun S, Gentry JW, Hyun YJ. Cultural adaptation of business expatriates in the host marketplace. J Int Bus Stud 2001;32(2):369–77. Killing P. Strategies for joint venture success. New York: Praeger; 1983. Krug JA, Hegarty WH. Predicting who stays and leaves after an acquisition: a study of top managers in multinational firms. Strat Manage J 2001;22:185–96. Li JJ. The formation of managerial networks in foreign firms in China: the effects of strategic orientations. Asia Pac J Manage 2005;22:423–43. Li JJ, Zhou KZ, Lam SK, Tse D. Active trust development of local senior managers in international subsidiaries. J Bus Res 2006;59:73–80. Lu, J.W., Hebert, L. Equity control and the survival of international joint ventures: A contingent. Madhok A. How much does ownership really matter?. Equity and trust relations in joint venture relationships. J Int Bus Stud 2006;37:4–11. Maertz CP, Campion MA. 25 years of voluntary turnover research: a review and critique. In: Copper CL, Robertson IT, editors. International review of industrial and organizational psychology, 13. New York: Wiley; 1998. p. 49–81.

Michel JG, Hambrick DC. Diversification posture and top management team characteristics. Acad Manage J 1992;35(1):9–37. Miller KI, Monge PR. Participation, satisfaction, and productivity: a metaanalytic review. Acad Manage J 1986;29(4):727–53. Mobley W. Intermediate linkages in the relationship between job satisfaction and employee turnover. J Appl Psychol 1977;62:237–40. Mossholder KW, Settoon RP, Henagan SC. A relational perspective on turnover: examining structural, attitudinal, and behavioral predictors. Acad Manage J 2005;48(4):607–18. Muthusamy SK, White M. Does power sharing mater? The role of power and influence in alliance performance. J Bus Res 2006;59:811–9. Naumann E. A conceptual model of expatriate turnover. J Int Bus Stud 1992;23: 499–531. O' Reilly CA, Caldwell DF, Barnett WP. Work group demography, social integration and turnover. Admin Sci Q 1989;34:21–37. Podsakoff PM, Organ D. Self-reports in organizational research: problems and prospects. J Manage 1986;12(Winter):531–43. Price JL, Mueller CW. Professional turnover. Bridgeport. CT: Luce; 1981. Reiche BS. The effect of international staffing practices on subsidiary staff retention in multinational corporations. Int J Hum Resour Manag 2007;18(4):523–36. Smith KG, Smith KA, Olian HP, Sims Jr HP, O, Bannon DP, Scully JA. Top management team demography and process: the role of social integration and communication. Admin Sci Q 1994;39:412–38. Steensma HK, Lyles MA. Explaining IJV survival in a transitional economy through social exchange and knowledge-based perspectives. Strat Manage J 2000;21:831–51. Wayne SJ, Shore LM, Liden RC. Perceived organizational support and leadermember exchange: a social exchange perspective. Aca Manage J 1997;40: 82–111. Wiersema MF, Bantel KA. Top management team turnover as an adaptation mechanism: the role of environment. Strat Manage J 1993;14:485–504. Wong PL, Ellis P. Social ties and partner identification in Sino-Hong Kong international joint ventures. J Int Bus Stud 2002;33(2):267–89. Wong CS, Law KS. Managing localization of human resources in the PRC: a practical model. J World Bus 1999;34(1):26–40. Wozniak L. Companies in China struggle to train, retain qualified managers. Wall Street J 2003:A8 December 30. Yan A, Gray B. Bargaining power, management control, and performance in United States–China joint ventures: a comparative case study. Acad Manage J 1994;37:1478–517. Zhang Y, George JM, Chan T. The paradox of dueling identities: the case of local senior executives' in MNC subsidiaries. J Manage 2006;32(3):400–25. Zhou KZ, Yim B, Tse D. The effects of strategic orientations on technology-and market-based breakthrough innovations. J Mark 2005;69:42–60.