Ownership structure, family leadership, and performance of affiliate firms in large family business groups

Hsi-Mei  Chung

Asia Pac J Manag (2012) 29:303–329 DOI 10.1007/s10490-011-9281-5 Ownership structure, family leadership, and performance of affiliate firms in large family business groups Hsi-Mei Chung & Shu-Ting Chan Published online: 4 February 2012 # Springer Science+Business Media, LLC 2012 Abstract Utilizing the agency viewpoint, this research attempts to shed light on the issue of family leadership by examining ethnic Chinese family business groups in Taiwan. The study examines the performance implications of this kind of leadership under the ownership structure concern. The research results indicate that the ownership structure of the affiliate firm influences the likelihood that family leadership will be used. Specifically, if the founding family owns more direct ownership of the affiliate firm, the family will be likely to appoint a family leader at the affiliate firm. However, when the founding family has a greater degree of pyramidal ownership of an affiliate firm, family leadership will be less likely at that affiliate firm. Additionally, family leadership mediates the relationship between ownership structure and affiliate firm performance in a family business group. Family leadership positively affects the effect of direct family ownership on affiliate firm performance but does not significantly affect the negative relationship between pyramidal ownership and affiliate firm performance. The implications of these findings for future research on leadership in family business groups are discussed. Keywords Agency theory . Family business groups . Family leadership . Family ownership structure The current authors greatly appreciate the comments from the Special Issue Editors and the discussants at the Special Issue Conference, and also sincerely appreciate the sponsorship provided to this study by “Aim for the Top University Plan” of the National Sun Yat-sen University and Ministry of Education, Taiwan, R.O.C., and National Science Council, Executive Yuan, Taiwan, R.O.C. (grant no. NSC 99-2410-H-214007-MY2, 2010/08–2012/07). H.-M. Chung (*) Department of Business Administration, I-Shou University, Kaohsiung 840, Taiwan e-mail: smchung@isu.edu.tw S.-T. Chan Department of Business Management, National Sun Yat-sen University, Kaohsiung 804, Taiwan e-mail: shuting0607@gmail.com 304 H.-M. Chung, S.-T. Chan Outside of Japan, almost all large publicly traded firms in Asia are family businesses (Claessens, Djankov, & Lang, 2000; La Porta, Lopez-de-Silanes, & Shleifer, 1999; La Porta, Lopez-de-Silanes, Shleifer, & Vishny, 2002; Young, Peng, Ahlstrom, Bruton, & Jiang, 2008). In North America and Europe, families enjoy a large degree of ownership of some publicly traded firms, but the control mechanisms on which such firms rely are substantively different than in Asia. For example, in publicly traded family firms in North America and Europe, the presence of independent directors on a board of directors partially mitigates the influence of the founding family, ensuring that the firm is managed in a way that is in the best interests of all shareholders (Anderson & Reeb, 2003; Miller & Le Breton-Miller, 2005). In contrast, in Asia, a large firm that is publicly traded will have a board of directors made up of family members and close friends (Claessens et al., 2000; La Porta et al., 1999, 2002; Peng & Jiang, 2010). Similarly, whereas publicly traded family businesses in North America and Europe rely on senior management teams not dominated by family members, one of the central control and coordination strategies that family businesses in Asia employ is the positioning of family members in key leadership positions (Claessens et al., 2000; Luo & Chung, 2005; Morck, Wolfenzon, & Yeung, 2005). Despite these clear differences between family businesses in Asia and North American or Europe, there has been little effort made to understand how families ensure coordination and control in affiliate firms in the Asian context. Again, one coordination and control mechanism that has been identified is the designation of family members as the heads of business units. However, the number of family members who might be available to fill such leadership positions is finite. As a result, families must choose where to position family members with the aim of maximizing their impact on the affiliate firm. This issue is particularly difficult for large family business groups with control of multiple, diverse affiliate firms (Almeida & Wolfenzon, 2006; Luo & Chung, 2005; Morck, 1996; Young et al., 2008). Therefore, to better understand the performance implications of choices regarding coordination and control in family business groups, we examine the placement of family members as leaders of affiliate firms in large groups of this type in the Chinese context. This research will employ agency theory and utilize multi-level, longitudinal data to examine this important topic. The key question addressed in this paper is the performance implications of family leadership in affiliate firms under the ownership structure consideration in large family business groups. The founding family in a large family business group usually utilizes the direct and pyramidal ownership structures at the same time. Previous studies have examined the performance implications of ownership structure for family business groups with public affiliate firms (Chang, 2003; Claessens et al., 2000). However, the relationships between different ownership structures, family leadership, and performance in affiliate firms have not been sufficiently analyzed (Carney, Gedajlovic, Heugens, van Essen, & van Oosterhout, 2011). Agency theory is widely employed to examine issues related to coordination and control because it can be theoretically expected that as investment increases, issues of coordination and control will be more closely addressed because of information asymmetry and risk (Eisenhardt, 1989; Jensen & Meckling, 1976). However, researchers have tended to use agency theory in the same manner for different institutional contexts (Eisenhardt, 1989; Fama & Jensen, 1983; Jensen & Meckling, 1976; Sanders & Hambrick, 2007). At the same time, scholars have increasingly Ownership structure, family leadership, and performance… 305 recognized that this sort of universal approach to agency theory is not appropriate (Granovetter, 1985). Rather, as some have suggested, direct or pyramidal family ownership structures are associated with different risk considerations, and thus, the form of ownership used may encourage different coordination and control decisions by affiliate firms in family business groups (Almeida & Wolfenzon, 2006; Chang, 2003; Kim & Sung, 2009). Previous studies have usually assumed that ownership and management manifest in the same way in affiliate firms in family business groups (Hamilton, 1997; Luo & Chung, 2005). However, the strategic importance and the risk that each affiliate firm poses for family members are not equally distributed, and as a result, the level of control that a family business group has over its affiliate firms may vary (e.g., Carney et al., 2011; Kim & Sung, 2009). Therefore, this research aims to investigate the possible relationships between family ownership structures, family leadership decisions, and the performance of affiliate firms in large family business groups. This research will make several contributions to the academic literature. First, although researchers have recognized that family businesses are likely to utilize family ownership and family leadership as key control mechanisms (i.e., Cruz, Gómez-Mejía, & Becerra, 2010; Luo & Chung, 2005; Peng & Jiang, 2010; Villalonga & Amit, 2006), there has been little prior research conducted on the diversity of ownership and management configurations used in family business groups. This study will fill that gap. Second, we will use agency theory to expand knowledge of family business groups and the coordination and control mechanisms that they use with affiliate firms. Business groups are the dominant form of business organization outside North America, but they remain poorly understood (Chang, 2006; Yiu, Lu, Bruton, & Hoskisson, 2007; Young et al., 2008). The examination of the issue of affiliate control will have strategic implications for business group management around the world (Carney et al., 2011; Chang, 2003). Moreover, examining the performance implications of family leadership and differential ownership structure will help to expand knowledge of ownership and management issues for family business groups. Finally, the multi-level and longitudinal data used in this research will provide insight into whether and how the founding family in a large family business group makes leadership decisions spanning multiple industries. To date, this longitudinal approach to leadership issues in family business groups has been rare. Hypothesis development Leadership in Asian family business groups The many complex institutional developments that have occurred in Asia offer researchers an opportunity to expand their theories, originating fresh insights based on context-specific considerations (Leung, 2007; Meyer, 2006). However, in addressing the issue of leadership in Asia, researchers have created fruitful theories and models but have focused too much on simplistic comparisons between Asia and Western societies (Tjosvold, Wong, & Hui, 2004; White, 2001). Typically, this crosscultural leadership research considers the role of cultural context in determining 306 H.-M. Chung, S.-T. Chan leadership effectiveness but provides insufficient evidence of whether and in what ways leadership matters to specific businesses in the Asian context (Bruton & Lau, 2008; White, 2001). Moreover, previous studies indicate that at the micro-level of analysis, leadership and trust issues for groups and teams are prominent topics in Asian management research (Bruton & Lau, 2008; Meyer, 2006; Tjosvold et al., 2004). Working with data from within different Asian countries, researchers usually employ mainstream leadership theories and examine issues such as the Asian leadership style (Ismail & Ford, 2010), guanxi and trust (Chen, Chen, & Xin, 2004; Lau, Liu, & Fu, 2007), leader-team member exchanges and the relationship between leadership and performance (Wang, Law, Hackett, Wang, & Chen, 2005), and the development of leadership in the Asian context (e.g., Leung, 2007; Meyer, 2006). In addition to the prominent micro-level leadership research, an increasing amount of research has addressed the leadership issues in Asia from a macro-level perspective— for instance, by exploring the role of leadership in the business group or network context (Bruton & Lau, 2008). A business group is a collection of legally independent firms that are linked by multiple ties, including ownership, economic connections (such as inter-firm transactions), and/or social relationships (family, kinship, friendship). Through these connections, the firms in question coordinate to achieve mutual objectives (Granovetter, 1995; Khanna & Rivkin, 2001; Yiu et al., 2007). In Asia, family ownership is the most prominent governance structure in business groups (La Porta et al., 1999; Morck et al., 2005; Young et al., 2008). As with other kinds of family businesses, one distinctive aspect of family business groups that results from their ownership structure is that their firms emphasize family leadership and trust among family members in decision-making (Almeida & Wolfenzon, 2006; Luo & Chung, 2005). Thus, assigning family members to appropriate leadership roles is a key strategic decision, one that ensures that family business groups achieve their goals (Morck et al., 2005; Peng & Jiang, 2010; Young et al., 2008). However, as family business groups become larger and/or grow internationally, they may find that they no longer have sufficient family members to serve as the leaders of their affiliate firms. Therefore, to achieve the mutual objectives of the affiliate firms in a family business group, the controlling family must consider whether and where to assign each family member as a leader within the formal ownership structure so as to ensure that each affiliate’s strategic direction is consistent with the goals of the family business group. Compared to the research in Asian leadership at the micro-level of analysis, the research on the antecedents and consequences of family leadership in family business groups is scarcer. Family leadership is strategically necessary to achieve family control and wealth (Arregle, Hitt, Sirmon, & Very, 2007; Cruz et al., 2010; GómezMejía, Haynes, Núñez-Nickel, Jacobson, & Moyano-Fuentes, 2007; Miller & Le Breton-Miller, 2005). In addressing family leadership-outcomes issues, previous research indicates that different types of family leadership may influence distinctive outcomes (e.g., Dyer, 1986; Pellegrini & Scandura, 2008; Sorenson, 2000). For example, based on family businesses in Western society, the participative leadership compared with the autocratic and expert leadership is positively related to both family and business financial outcomes as well as to employee satisfaction and commitment (Sorenson, 2000). Additionally, in the Asia Pacific, paternalistic leadership style is found to be one of the most remarkable dimensions in Chinese family business Ownership structure, family leadership, and performance… 307 leadership studies (Pellegrini & Scandura, 2008). Research indicates that the unidimensional construct of paternalistic leadership can be delineated into three dimensions, that is, benevolence, authoritarianism, and morality (Cheng, Chou, & Farh, 2000). Prior studies have examined how these different dimensions of paternalistic leadership in Chinese family business affect subordinates’ organizational commitment, satisfaction with a team’s leader, and the commitment to the team (e.g., Cheng, Huang, & Chou, 2002a; Cheng, Shieh, & Chou, 2002b; Farh, Cheng, Chou, & Chu, 2006). However, as we argued before, previous research in addressing the family leadership-macro outcomes issues has under-explored the role of family leadership in the business group or network context (Bruton & Lau, 2008). Thus, although Vallejo (2009) argued that leadership is more transformational in family businesses than in non-family businesses, he also indicated that it still needs much effort to further the current research work of leadership and its possible consequences in the family business context. Therefore, this research will fill the gap in examining the possible influence of family leadership on performance under various ownership structures in family business groups. Agency theory provides a good perspective from which to examine the issue above, especially in a family business group context characterized by multiple links among affiliate firms. Thus, the following will examine the relationships between ownership structure, family leadership decisions, and performance implications in family business groups from the agency theory viewpoint. Ownership structure and family leadership Family ownership is a key characteristic that distinguishes the structure of family governance from that of more widely held and non-family firms (Anderson & Reeb, 2003; Villalonga & Amit, 2006). In addressing the strategic issues associated with ownership structures in family businesses, previous studies have usually examined the performance implications of such structures (e.g., Anderson & Reeb, 2003; Miller, Le Breton-Miller, & Scholnick, 2008; Peng & Jiang, 2010). Evidence derived by examining family businesses in developed economies indicates that family ownership will generate a greater level of motivation to manage firms, helps to monitor managers, and thus has positive implications for performance (Anderson & Reeb, 2003; Demsetz & Lehn, 1985; Miller et al., 2008). However, we know that family members usually exert control over businesses through both ownership structure and family leadership. Thus, whether and how the ownership structure will influence family leadership decisions is a key issue for family businesses (Peng & Jiang, 2010; Villalonga & Amit, 2006). Furthermore, in those family business groups in emerging economies, the founding family usually relies on a pyramidal ownership structure to control its multiple affiliate firms (Almeida & Wolfenzon, 2006; Claessens et al., 2000; La Porta et al., 1999, 2002; Morck et al., 2005; Young et al., 2008). Compared with the evidence from the family businesses in developed economies indicating that family ownership generates positive motivation, whether and how the utilization of family direct ownership and pyramidal ownership in family business groups in emerging economics will influence family leadership decisions has not been sufficiently explored. This research will fill that gap. 308 H.-M. Chung, S.-T. Chan The different ownership structures of family business groups, which include both direct family ownership and the pyramidal ownership structure, may generate distinctive effects on the strategy and power of family business groups (Claessens, Djankov, Fan, & Lang, 2002; Kim & Sung, 2009; La Porta et al., 2002). From the agency theory viewpoint, the founding family’s ownership structure entails distinctive risk and thus may influence the coordination and control mechanisms used (Demsetz & Lehn, 1985; Fama & Jensen, 1983). The key concern in agency theory is contract design, as contracts should minimize possible agency costs (Eisenhardt, 1989; Fama & Jensen, 1983; Jensen & Meckling, 1976). Given that agency relationship exists under the attitudinal and information consideration, the principal should choose an incentive contract or a monitoring mechanism to reduce agency cost (Eisenhardt, 1989; Fama & Jensen, 1983; Jensen & Meckling, 1976; Sanders & Hambrick, 2007). Family businesses—or more specifically, family business groups as a subset of family businesses—rely on trust relationships to manage their business and overcome possible transaction uncertainty and risk (Anderson & Reeb, 2003; Cruz et al., 2010; Fama & Jensen, 1983). Family members focus primarily on running the business to generate wealth for the family; thus, in contrast to non-family managers, family members have a greater incentive to obey the instructions of the business group leader and do not necessarily require close monitoring (Arregle et al., 2007; Cruz et al., 2010; Gómez-Mejía et al., 2007). Previous research indicates that when a founding family has majority shareholdings, the family has a greater incentive to serve as good stewards in its businesses and is motivated to monitor non-family managers. Thus, the agency problem should be less acute in family businesses (Anderson & Reeb, 2003; Demsetz & Lehn, 1985; Fama & Jensen, 1983). From an agency perspective, it appears that trust relationships among family members will reduce the need to monitor a unit’s decisions closely because family members will have a unique incentive to maximize the founding family’s wealth (Cruz et al., 2010; Fama & Jensen, 1983). In other words, the combination of founding family shareholdings and family leadership will lessen the agency problem for founding family managers as opposed to non-family managers (Anderson & Reeb, 2003; Demsetz & Lehn, 1985; Fama & Jensen, 1983). Following this logic, with reference to family business groups, we argue from the agency theory viewpoint that direct founding family ownership may increase the likelihood of family leadership in family business groups and generate better performance (Claessens et al., 2002). Compared with these arguments regarding motivation and direct family ownership, the strategic implications of the pyramidal ownership structure are controversial (Almeida & Wolfenzon, 2006; Claessens et al., 2002; Kim & Sung, 2009). In this type of pyramidal ownership structure, the founding family achieves control of the constituent firms via a chain of ownership relationships: the family directly controls a firm, which in turn controls another firm, which might itself control other firms (Almeida & Wolfenzon, 2006: 1). The pyramidal ownership structure allows a founding family to control a chain of affiliate firms based on a relatively small investment (La Porta et al., 1999). By utilizing this ownership structure, family members can leverage their financial risk by expanding their businesses; they can exert control over multiple affiliate firms without providing larger financial inputs (Chang, 2003; Kim & Sung, 2009; Morck et al., 2005). Thus, a founding family can utilize the pyramidal ownership structure to meet its financial requirements if Ownership structure, family leadership, and performance… 309 external funds are more costly than internal funds. Furthermore, such an ownership structure constitutes an optimal financial choice when family members expect a higher return on costs associated with an affiliate firm (Almeida & Wolfenzon, 2006; Young et al., 2008). The family business group must also confront increases in uncertainty and risk as it and its firms grow in size and further diversify, creating coordination and control problems (Claessens et al., 2002; Levy, 2009). This issue is a particularly difficult problem for family business groups because there are limits to the number of family members who can be called on to manage multiple affiliate firms (Morck, 1996; Morck et al., 2005). If the founding family utilizes the pyramidal ownership structure, the investment risk for the family business group will be less substantial than under direct ownership (Morck, 1996). The use of this ownership structure will generate a weaker incentive for founding families to assign family members to management positions because the potential loss and risk under this structure are smaller for them (Almeida & Wolfenzon, 2006; Levy, 2009). Thus, one could expect that the founding family in a family business group would choose a non-family manager to head such an affiliate firm because the potential loss of wealth and potential risk would be smaller. In other words, from the agency theory viewpoint, the direct and pyramidal forms of family ownership have differential strategic implications for coordination and control mechanism such as leadership decisions made by family business groups (Claessens et al., 2002; Kim & Sung, 2009; Levy, 2009). The following hypotheses are thus presented: Hypothesis 1 The greater the degree of direct ownership of an affiliate firm by members of a family business group, the higher the likelihood that this affiliate firm will be headed by a family member. Hypothesis 2 The greater the degree of pyramidal ownership of an affiliate firm by members of a family business group, the lower likelihood that this affiliate firm will be headed by a family member. Ownership structure, family leadership, and affiliate firm performance Family members usually play a strategic role in coordination and control in family businesses (Anderson & Reeb, 2003; Arregle et al., 2007; Cruz et al., 2010; GómezMejía et al., 2007). Assigning family members to key decision-making positions can ensure coherence and solidarity in family businesses and facilitate family goal achievement (Arregle et al., 2007; Cruz et al., 2010; Gómez-Mejía et al., 2007). From the agency theory viewpoint, family leadership can be helpful because it can mitigate the possible conflict of interest between controlling shareholders and nonfamily professional managers in family businesses, thus improving performance (Anderson & Reeb, 2003; Fama & Jensen, 1983). However, previous research has also indicated that family leadership may cause entrenchment and issues with selfcontrol in family businesses and may thus be detrimental to performance (GómezMejía, Núñez-Nickel, & Gutierrez, 2001; Morck, 1996; Schulze, Lubatkin, Dino, & Buchholtz, 2001). Indeed, prior research shows that family leadership can simultaneously generate both positive management motivation and issues with self-control (Miller et al., 2008; Morck, 1996; Villalonga & Amit, 2006). Ownership structure is 310 H.-M. Chung, S.-T. Chan thus a key concern when researchers address the performance of family leadership in family businesses (Claessens et al., 2002; Miller et al., 2008; Villalonga & Amit, 2006). The relationship among ownership structure, family leadership, and performance is a key issue in family business field (Miller et al., 2008; Peng & Jiang, 2010; Villalonga & Amit, 2006). In the family business group context, it seems difficult to tell a priori whether the benefits of family control—through family leadership or pyramidal ownership structure—in large firms outweigh the costs, or vice versa (Carney et al., 2011; Peng & Jiang, 2010). In this study, we propose that when a founding family owns the majority share of a family business, it will hope to preserve its social-emotional wealth while also helping the business to grow (Arregle et al., 2007; Gómez-Mejía et al., 2001, 2007). Agency theory suggests the more shares owned by the founding family, the higher resource commitment for the founding family in this family business (Anderson & Reeb, 2003; Fama & Jensen, 1983). Therefore, if the founding family has higher ownership, it will increase the founding family’s risk of loss and thus the higher positive motivation in management (GómezMejía et al., 2001, 2007; Mishra & McConaughy, 1999). In explaining the relationship between family ownership and performance, previous studies indicate that the more shares owned by family members the better performance in this family business (Anderson & Reeb, 2003; Chu, 2011; Cruz et al., 2010). Others also argue that founding family ownership will generate greater goal achievement and management motivation, and accordingly better strategic decisions and performance implications (Gómez-Mejía et al., 2001; Mishra & McConaughy, 1999; Villalonga & Amit, 2006). Whereas it would seem that direct family ownership creates greater motivation for founding families with regard to management, a pyramidal ownership structure should generate a situation in which family members and family business groups can reduce their financial and loss risk (Claessens et al., 2000, 2002). The choice of pyramidal ownership is based on financial concerns related to expansion (Almeida & Wolfenzon, 2006). Previous literatures indicate that the use of the pyramidal ownership structure may be detrimental to performance under some conditions, such as insufficient shareholder protection or inefficient financial markets in institutional environments (Morck et al., 2005; Young et al., 2008). In these under-developed institutional environments, a lower investment by the founding family under the pyramidal ownership structure will generate a separation between cash flow rights and control rights in public affiliate firms and thus is bad for minority shareholders’ interests (La Porta et al., 1999, 2002; Morck et al., 2005; Young et al., 2008). Therefore, the use of pyramidal ownership will be detrimental to the public affiliate firm’s performance in a family business group (Claessens et al., 2000; La Porta et al., 1999). Besides the distinctive relationship between family ownership and performance, furthermore, it is argued that if the founding family does not assign qualified family members to manage its business, the poor management of the business will be more detrimental to the family’s goals and social-emotional wealth in family businesses (Demsetz & Lehn, 1985; Gómez-Mejía et al., 2007). Thus, the founding family’s ownership will decide whether to choose a family manager or a non-family manager to manage the family business, as we argue previously. Agency theory suggests that in a family business context, family leadership compared with the non-family management is helpful to mitigate the agency problem between managers and Ownership structure, family leadership, and performance… 311 shareholders since the founding family can maintain control over non-family managers (Chu, 2011; Cruz et al., 2010; Demsetz & Lehn, 1985). In general situations, family leadership will have positive performance implications (Anderson & Reeb, 2003; Chu, 2011; Fama & Jensen, 1983). Furthermore, for those family business groups that are embedded in an under-developed institutional environments, family leadership may help to access to unique resources through kinship networks (Liu, Yang, & Zhang, 2012). Family leadership will contribute to the family business group’s performance during market-oriented institutional transition (Luo & Chung, 2005). In a large family business group, the family members can utilize both direct and pyramidal ownership to control the public and private affiliate firms, and the different ownership structure will decide whether the founding family will rely on a family member to lead this affiliate firm or not, as we argued before. Thus, no matter the affiliate firm is owned by family members directly or indirectly in a pyramidal ownership structure, family leadership will impose authorized power in strategic decisions (Chang, 2003; Morck et al., 2005), and accordingly influence the performance in family businesses or family business groups in a positive manner (Anderson & Reeb, 2003; Cruz et al., 2010; Luo & Chung, 2005). Following the logic above, in a family business group, the more founding family’s direct ownership on an affiliate firm, this affiliate will perform better. The use of pyramidal ownership structure on the other hand will generate negative affiliate firm performance implications in a family business group. Since the founding family’s ownership in an affiliate firm will decide whether this affiliate firm is headed by a family member or not, the family leadership decision will influence affiliate firm performance in family business groups. Thus, we expect family leadership will mediate the positive relationship between family ownership and performance in an affiliate firm in a family business group. Furthermore, since family business under pyramidal ownership is less likely to appoint family members, we expect family leadership to mediate the negative relationship between pyramidal ownership and affiliate firm performance in a family business group. The following hypotheses are presented accordingly: Hypothesis 3 Family leadership will mediate the relationship between ownership structure and affiliate firm performance in family business groups. Hypothesis 3a Family leadership will mediate the positive relationship between direct family ownership and affiliate firm performance in family business groups. Hypothesis 3b Family leadership will mediate the negative relationship between the pyramidal ownership structure and affiliate firm performance in family business groups. Research method Sample and data The current research uses longitudinal, multi-level data from 1988 to 2004 to examine the relationship between ownership structures, family leadership, and the performance of affiliate firms in Taiwan’s family business groups. We chose to focus 312 H.-M. Chung, S.-T. Chan on business groups in Taiwan because Taiwan is one of the world’s 25 largest economies despite having a relatively small population of less than 30 million (IMD, 2006). The island also has well-established legal traditions that help to ensure that the public data reported by the business groups are reliable (IMD, 2006). Taiwan is representative of a number of other newly industrialized economies with a history of dominant business groups, including South Korea, Hong Kong, and Singapore (Chang, 2006; Lasserre & Schütte, 2006). Taiwan is also suitably representative of other nations with many Chinese residents, including Hong Kong, the People’s Republic of China, Singapore, Malaysia, and other Asian countries (Lasserre & Schütte, 2006). The data for the largest 100 business groups in Taiwan from 1988 to 2004 were derived from the China Credit Information Service (1990, 1992, 1994, 1996, 1998, 2000, 2001, 2002, 2003, 2004). Moreover, since there was a marketoriented institutional transition in Taiwan in 1988 (Luo & Chung, 2005), the years examined in this study are helpful to understand the role of family control in performance in large family business groups after the institutional transition in Taiwan. The data culled from this resource have also been employed in previous studies of Taiwan’s business groups and have proven reliable (e.g., Chung, 2003; Luo & Chung, 2005). The directory of Taiwanese business groups was published every two years prior to 1999; thus, 12 yearly datasets were examined for the 16-year period (1988, 1990, 1992, 1994, 1996, 1998, 1999, 2000, 2001, 2002, 2003, and 2004). Our focus is on family business groups in the textile, electronic and household appliance, financial service, and constructing investment industries, which play a critical economic role in both the domestic and export markets in Taiwan, making up more than 50% of the island’s export market during the time period examined (Ministry of Finance, 2008). We classified a business group as family-owned if it was managed and controlled by a specific family or set of families at that time and if a previous generation of the same family (or families) had owned that business. As in previous studies, a business was considered a family business only when there was more than one family member involved (Cruz et al., 2010; Miller & Le Breton-Miller, 2005). However, the use of longitudinal data can result in family business groups moving in and/or dropping out of the top-100 business groups. To ensure that any effects on performance were long-term and not a product of misleading accounting practices, we required that the family business groups be part of the top-100 business groups for at least 10 years during the 16 years examined. In this way, we derived a sample of 21 family business groups. The analysis was conducted at the affiliate firm level. Because multiple affiliate firms are nested in one family business group, the data analysis addressed the multilevel nature of the data; we considered the influence of both group-level and affiliate firm-level variables on the dependent variable (Rabe-Hesketh & Skrondal, 2008). Moreover, the longitudinal and multi-level nature of the data made it appropriate to use a hierarchical linear model (Bryk & Raudenbush, 1992) to consider the growth pattern of the outcomes. The benefits and importance of longitudinal data analysis in strategic management analysis (Certo & Semadeni, 2006) and business group research (i.e., Khanna & Rivkin, 2001; Luo & Chung, 2005) is widely recognized. During the 12 yearly datasets examined, 252 family business groups were employed in this research (21 [number of family business groups] × 12 [number of year Ownership structure, family leadership, and performance… 313 datasets] 0 252). Again, the level of analysis that we used was the affiliate firms in each family business group. Thus, the 252 family business groups provided us with the opportunity to investigate 6,916 affiliate firms. Variables Dependent variable (Y): Affiliate firm’s performance We considered data for all affiliate firms in these family business groups, including both public and private affiliate firms. Because the data explored in the database were limited, we considered the financial performance of each affiliate firm to indicate whether this affiliate firm performs efficiently or is adequately productive (Pfeffer, 1997). Thus, we utilized sales revenue as a proxy for affiliate firm performance when a standard public performance index was not available (Miller et al., 2008). The sales revenue of each affiliate firm can indicate the affiliate firm’s efficiency and the outcomes for operation (Pfeffer, 1997; Scott, 1977). Mediating variable (M): Family leadership The mediating variable is family leadership in an affiliate firm. In this research, we used the information available on whether the affiliate head is a family member as a proxy for this variable (e.g., Anderson & Reeb, 2003; Miller & Le Breton-Miller, 2005). We identified a family member as the affiliate head based on data from the China Credit Information Service and other public information. For example, if the affiliate head is described in such data sources as “having kinship relationships with the founding family members in the group,” we classified this affiliate head as a family member. The dependent variable was dummycoded as “1” if the affiliate head was a family member and as “0” otherwise. Independent variables (X1–X2): Affiliate ownership structure To calculate the affiliate ownership structure used by each family business group, we used data from the China Credit Information Service for the 12 datasets (1988, 1990, 1992, 1994, 1996, 1998, 1999, 2000, 2001, 2002, 2003, 2004). In distinguishing between direct and pyramidal family ownership structure, we followed Claessens et al. (2000, 2002) and La Porta et al. (1999, 2002), who defined control by computing the voting and cash-flow ownership of the controlling shareholder (or family) based on their direct and indirect rights. Here, we extended this concept by dividing family ownership into two categories: direct and pyramidal. We identified the affiliates’ major shareholders and their shares using multiple secondary databases, including the China Credit Information Service and other public databases. The data from the China Credit Information Service indicate most of each affiliate firm’s major shareholders and their shares. Affiliate firms’ major shareholders may be individuals, firms, or both. If the data on how many shares the main shareholders had were missing, we attempted to use the appropriate annual reports to determine this information. In this study, family direct ownership (X1) is the sum of the following two types of ownership (Claessens et al., 2000, 2002; La Porta et al., 1999, 2002): (1) the shares directly owned by family members, and (2) the shareholdings of the nominal agents controlled by the family—for example, through nonprofit organizations such as universities and hospitals. 314 H.-M. Chung, S.-T. Chan In family business groups, family members can exert overall control over their group by using both the direct and the pyramidal ownership structures. In definition, an affiliate firm is controlled by a founding family (or the ultimate owner) if this founding family’s direct and indirect voting rights in this affiliate firm exceeds the direct votes controlled by this family (La Porta et al., 2002). For the founding family in a large family business group, this family’s control over the cash-flow rights and the voting rights are separate since the founding family can use relatively smaller cash-flow rights to control higher percentages of voting rights in an affiliate firm in this family business group (Claessens et al., 2000; La Porta et al., 1999). Information about the actual ultimate owner’s voting rights, cash-flow rights, and the pyramidal ownership structure is limited since previous research has only examined the public affiliate firms in family business groups (Claessens et al., 2000, 2002; La Porta et al., 1999, 2002). Thus, this approach may create a bias in terms of ownership structure and firm valuation and thus underestimate the effect of ownership structure on firm valuation (Claessens et al., 2002: 2747). Claessens, La Porta, and their colleagues’ measurements only used information about public affiliate firms to estimate the controlling shareholder’s voting rights and cashflow rights and did not estimate the pyramidal structure in a business group with both public and private affiliate firms (Claessens et al., 2000, 2002; La Porta et al., 1999, 2002). Therefore, in calculating the pyramidal ownership structure, we revised the limitation in the previous calculation (Claessens et al., 2000, 2002; La Porta et al., 1999, 2002) and redefined the founding family’s (or the ultimate owner’s) weakest link in the chain of voting rights among the public and private affiliate firms in a family business group. As in prior research (Claessens et al., 2000, 2002; La Porta et al., 1999, 2002), a founding family has x percent indirect control over firm A under the following situations: (1) if this founding family directly controls firm B, which in turn directly controls x percent of the votes for firm A; or (2) if this founding family directly controls firm C, which in turn controls x percent of the votes for firm B, and if firm B directly controls x percent of the votes for firm A. Thus, when the founding family has personal shareholders or institutional shareholders in the control chain, and accordingly these multiple shareholders have different percentages of the votes in the voting rights of control chain, we select the weakest link in the chain of voting rights (Claessens et al., 2000, 2002). The rationale to pick up the weakest link in the chain of voting rights is that this founding family can utilize the weakest votes to control the target affiliate firm (i.e., firm A). Therefore, pyramidal ownership structure (X2) is the sum of the cross-held shares held by more than one public affiliate firm in the same family business group. For example, suppose that family members have directly 25% of the voting rights for firm A, and firm A owns 100% of firm B. On the other hand, firm B in turn owns 17% of firm A. Thus, in this simple case, the founding family has 42% of the control rights in firm A, 25% directly and 17% through a pyramidal chain. For more details, please refer to the Appendix for a more complex case in pyramidal chain. Control variables The database used contains cross-sectional dominant data (e.g., N > T, N 0 6916, T 0 12). To control for possible contemporaneous correlation (i.e., the residuals of the units observed in each time period being correlated), Ownership structure, family leadership, and performance… 315 following Khanna and Rivkin (2001), we used 12 dummy-coded year variables (y1988, y1990, y1992, y1994, y1996, y1998, y1999, y2000, y2001, y2002, y2003, and y2004). Because of collinearity, y1988 and y2004 were dropped from the statistical model. Again, this research was conducted at the affiliate firm level, and each affiliate firm is nested in a specific business group. Thus, we were obliged to control for the influence of group-level variables. In this paper, we also controlled for business group size and age (Khanna & Rivkin, 2001). In addition, we controlled for the possible influence of the number of group affiliates in the 12 year-long sets (Luo & Chung, 2005). Furthermore, in this study, we controlled two variables related to the influence of the family at the group level. One variable, named “Founding family leadership,” indicates whether the founder still holds a key leadership position in the family business group (e.g., Anderson & Reeb, 2003; Cruz et al., 2010). This variable was dummy-coded, with “1” indicating that the founder still served as the CEO or chair of the family business group’s core company during the time period examined. Additionally, we also controlled for “Family members in the inner circle” to address whether each family controlled the strategic direction of the business group (Hamilton, 1997; Luo & Chung, 2005). Family members in the inner circle were identified based on whether they had ties to the particular family. In addition to controlling for the influence of group-level variables, we also controlled for the possible influence of affiliate-level variables, including each foreign affiliate’s industry, age, assets, and location. Whether a firm engages in manufacturing or provides services can influence the characteristics of that firm’s resources, which in turn can influence the firm’s staffing decisions or even performance (e.g., Cruz et al., 2010). Furthermore, the affiliate’s assets, age, and location influence the firm’s accumulation of experience and the process of knowledge transfer from the core company to the affiliate, thus also impacting what coordination and control mechanisms are chosen (e.g., Anderson & Reeb, 2003; Chang, 2003). Therefore, we controlled for these characteristics. Data analysis The family member assigned as the affiliate head and the structure governing the family ownership of the affiliate firms are determined based on strategic decisions that occur at the affiliate firm level. However, the character of the nested family business group will influence each affiliate decision (Dansereau, Yammarino, & Kohles, 1999). Thus, in this type of research, the statistical model must be sufficient to address multi-level issues and to account for the characteristics of the longitudinal data. We used a growth model for this purpose (Bryk & Raudenbush, 1992; RabeHesketh & Skrondal, 2008), with the multi-level mixed-effects maximum likelihood model (ML model) together with the STATA 9.0 software. Researchers use the ML model in regressions when the data are characterized by both random effects generated by the nested group-level variables and fixed effects associated with the affiliatelevel independent variables (Bryk & Raudenbush, 1992; Rabe-Hesketh & Skrondal, 2008). The repeated affiliate observations from the same family business group may lead to the affiliate observations nested in the same group (Dansereau et al., 1999). The ML model is a variance-components model that can be used to compensate for 316 H.-M. Chung, S.-T. Chan the possible nested effects that occur with variables with two or more levels. Moreover, the ML model also assumes the variables at the affiliate level to have an effect on the dependent variable only via the residual term, which has a different value for each firm-group combination (Rabe-Hesketh & Skrondal, 2008). Thus, the ML model is a version of a hierarchical linear model (Bryk & Raudenbush, 1992) that has allowed us to consider the growth patterns in the data and address the possible intra-level correlation problems that occur in multi-level data. Additionally, to ensure that homoskedasticity did not negatively impact our analysis, we conducted the test suggested by Brown and Forsythe (1974) to test the null hypothesis of equality of variances across groups. The test results led us to reject the null assumption of homoskedasticity across the panel. Therefore, it is clear that the ML model allowed us to control for heteroskedasticity, intra-correlation problems that might have occurred with the longitudinal, multi-level data. Results Correlation analysis The results presented in Table 1 indicate that the dependent variable, the affiliate firm’s performance, is positively correlated with the mediating variable: family leadership (p ≤ .01). Moreover, affiliate firm performance is positively correlated with direct family ownership in affiliate firms (p ≤ .05) and negatively correlated with pyramidal family ownership in affiliate firms (p ≤ .01). Table 1 indicates that family leadership is positively correlated with direct family ownership of affiliate firms and negatively correlated with pyramidal family ownership of affiliate firms (p ≤ .01 for both correlation relationships). The mean and standard deviations of each variable are indicated in Table 1. Thus, it is clearly important to control for these variables. Causal analysis To determine the antecedents and consequences of family leadership in family business groups, we used two serial ML models that helped us to understand the relationship between ownership structure, family leadership, and the performance of affiliate firms in family business groups (see Figure 1). The first series of ML models (M1 to M4) was used to test Hypotheses 1 and 2, regarding the relationship between ownership structure and family leadership with regard to affiliate firms in family business groups. We used the second series (M5 to M10) to test the mediating effects of family leadership, examining the role of independent, mediating, and dependent variables. An influential statistical approach to determining mediation was presented in Baron and Kenny (1986), with an updated account made by Kenny, Kashy, and Bolger (1998). The Sobel test was also utilized to illustrate the rigor requirement in mediating effects (Sobel, 1982). In this study, we follow the instructions discussed extensively in articles on testing mediation (Baron & Kenny, 1986; Kenny et al., 1998; Lam, Loi, & Leong, 2012; Lau et al., 2007; Sobel, 1982). First, it is necessary to show that the independent variables are correlated with the mediator (i.e., family leadership in the affiliate firm). Second, the direct effects from Mean 1. Affiliate’s performance S.D. 1 2 3 4 85494.87 523946.51 2. Family leadership in affiliate firm .47 .50 .099** 3. Family direct ownership in affiliate .04 .16 .012* 4. Family pyramidal ownership in affiliate .53 .44 −.103** −.196** −.265** 5. Affiliate asset 263793.05 6. Affiliate age 31.55 1892932.28 .770** 199.39 −.004 .076** −.015 –.100** .052** .001 −.003 .154** −.315** .67 .47 .057** .050** 8. Manufacturing affiliate .29 .45 .006 .069** −.019 15482220.51 18748625.77 .070** −.153** .008 .014 10. Group age 40.13 14.37 11. Number of affiliate 72.56 46.63 −.036** −.187** −.137** 12. Percentage of family member in inner circle .78 .27 13. Founder leadership .34 .47 −.026* 6 7 8 9 10 11 12 .156** 7. Affiliate location 9. Group asset 5 −.004 .077** −.083** −.118** −.036** .168** .031** −.017 −.023* .073** −.018 .331** −.021* .239** −.020 .087** −.030** .136** −.027* .129** −.076** −.067** −.014 .022* .198** −.185** −.016 .043** −.007 −.015 Ownership structure, family leadership, and performance… Table 1 Correlation matrix. −.163** .177** .058** −.254** −.083** .513** .000 .072** −.048** .254** −.036** .316** −.026* .084** .002 .215** .093** −.077** .089** *p ≤ .05, ** p ≤ .01 (two-tailed). Unit of affiliate revenue and asset is USD1,000. (Average exchange rate of USD / NTD is 29.56 from 1988 to 2004; Source: Central Bank of R.O.C. http://www.cbc.gov.tw/ content.asp?CuItem027029). 317 318 H.-M. Chung, S.-T. Chan X1: Family direct ownership H1 (+) M: Family leadership H2 (–) X2: Family pyramidal ownership H3a & H3b H3b (–) D: Affiliate firm’s performance H3a (+) Figure 1 Model of mediated role of family leadership on the relationship between ownership structure and affiliate firm’s performance Note: Solid paths indicate hypothesized relationships. Dotted paths indicate potential indirect influences in terms of ownership structure effects on performance. independent variables on the dependent variable must be statistically significant to establish that there is an effect that may be mediated. Moreover, we need to test that the effects from independent variables on the dependent variable are changed after considering the mediator. In that, if the significant relationship between the independent variables and the dependent variable is no longer significant after considering the mediator, it can be concluded that the full mediation occurs. However, if the significant relationship becomes less significant after considering the mediator, we will conclude that the mediator partially mediates the relationship between the independent variable and the dependent variable. Third, the use of the Sobel test provides a more rigorous requirement in mediating effects (Baron & Kenny, 1986; Sobel, 1982). The Sobel test is utilized to determine whether the mediator carries the effects of the independent variables on to the dependent variables. In this study, we use the Sobel test to judge whether the ownership structures, including direct and pyramidal family ownership, influence family leadership, which in turn influences affiliate firm performance. In this study, the Wald Chi-square values were significant for all models. Therefore, the model fitness and setting were deemed satisfactory. Regarding the influence of the group- and affiliate-level control variables on family leadership within an affiliate firm, there are several key control variables. The nested effects of group-level variables including business group size, age, and the degree of family control are significant. Furthermore, in addressing the influence of affiliate-level control variables, Table 2 demonstrates that some of the affiliates’ characteristics significantly influence the family leadership of those affiliate firms in family business groups. Specifically, an affiliate firm’s assets and age have a positive effect on family leadership decisions, and thus, the larger and older an affiliate firm is, the more likely it is that a family member will head it (p ≤ .01 for both the influence from the affiliate firm’s assets and age on performance). Furthermore, if the affiliate firm is based domestically, it is more likely that a family member will be assigned as the affiliate firm’s leader. Finally, in the particular years examined, Ownership structure, family leadership, and performance… 319 Table 2 Relationship between ownership structure and family leadership of affiliate firms in family business groups. Family Leadership Constant M1 .102 (.054)† M2 .064 (.051) M3 .226 (.057)** M4 .176 (.055)** Independent Variables X1: Family direct ownership .466 (.050)** X2: Family pyramidal ownership .418 (.050)** −.125 (.016)** −.107 (.016)** Control Variables Year 1990 .026 (.046) .027 (.045) .013 (.046) .015 (.045) Year 1992 −.011 (.042) .002 (.042) −.019 (.042) −.010 (.041) Year 1994 −.018 (.039) −.010 (.039) −.029 (.040) −.022 (.039) Year 1996 −.031 (.037) −.014 (.037) −.038 (.037) −.025 (.037) Year 1998 −.030 (.034) −.004 (.034) −.022 (.035) −.011 (.034) Year 1999 .005 (.024) .014 (.024) .012 (.024) .015 (.024) Year 2000 .019 (.020) .024 (.020) .024 (.020) .027 (.020) Year 2001 .023 (.020) .028 (.020) .025 (.020) .029 (.020) Year 2002 −.017 (.019) −.013 (.019) −.013 (.019) −.011 (.019) Year 2003 −.018 (.019) −.014 (.019) −.016 (.019) −.012 (.019) Affiliate’s assets .039 (.002)** .038 (.002)** .034 (.002)** .035 (.002)** Affiliate’s age .038 (.006)** .036 (.006)** .034 (.006)** .033 (.006)** Affiliate’s location −.021 (.013)† −.028 (.013)* −.044 (.013)** −.045 (.013)** Manufacturing affiliate −.043 (.012)** −.047 (.012)** −.051 (.012)** −.053 (.012)** .008 (.001)** .007 (.001)** .008 (.001)** .007 (.001)** Control variablesRandom parts Wald Chi-Square 427.37** 520.21** 495.10** 568.53** Number of observations 6,916 6,861 6,861 6,861 Number of Groups 21 21 21 21 The random effect contains group’s size, group’s age, number of affiliates, founding family leadership, and percentage of family member in inner circle. Standard deviation data given in parentheses. † p ≤ .10, * p ≤ .05, ** p ≤ .01. if the affiliate firm is in the manufacturing sector, it is less likely for a family member to be appointed to a leadership position (p ≤ .01). The results presented in Model 2 (M2) of Table 2 show that the higher the degree of direct ownership of a founding family in an affiliate firm, the greater the possibility that the founding family will appoint a family member to a leadership position in the affiliate firm (β 0 .47, p ≤ .01). Additionally, according to Model 3 (M3) of Table 2, a greater degree of pyramidal ownership of an affiliate firm has a significantly negative influence on family leadership decisions regarding the affiliate firm (β 0 −.13, p ≤ .01). 320 H.-M. Chung, S.-T. Chan The current data analysis therefore supports the agency theory argument made in Hypotheses 1 and 2. Furthermore, following the previous instructions on testing mediation (Baron & Kenny, 1986; Kenny et al., 1998; Lam et al., 2012; Lau et al., 2007; Sobel, 1982), Table 3 indicates that family leadership demonstrates differential mediating influence on the relationship between ownership structure and affiliate firm performance in family business groups. Specifically, according to Models 6 (M6) and 7 (M7) of Table 3, during the years examined, family leadership positively mediates the relationship between direct family ownership and affiliate firm performance. In that period, the higher degree of direct family ownership in an affiliate firm, the higher possibility that this affiliate firm is headed by a family member, and the family leadership generates positive performance in this affiliate firm. More importantly, the significance of family direct ownership on firm performance disappears after considering family leadership in the regression model (see M7). There are evidences of mediation according to Baron and Kenny’s test (1986). The Sobel test statistic also reveals that family leadership mediates the positive effects of family direct ownership on affiliate firm performance (Z 0 2.94, p < .05). Thus, these results indicate that Hypothesis 3a is supported. However, according to Models 8 (M8) and 9 (M9) of Table 3, the founding family’s pyramidal ownership has a more negative impact on affiliate firm performance when family leadership is included, thus the result definitely does not support Baron and Kenny’s (1986) conditions. Thus, family leadership does not mediate the relationship between family pyramidal ownership and affiliate firm performance. Moreover, the Sobel test indicates that family leadership does not significantly mediate the relationship between family pyramidal ownership and affiliate firm performance (Z 0 −2.85, n.s.). Therefore, Hypothesis 3b is not supported. These unexpected results are discussed below. Discussion and conclusion Family businesses constitute a distinctive worldwide governance system. One of the key mechanisms through which family businesses (or family business groups as a set of family businesses) ensure coordination and control is family leadership (Cruz et al., 2010; Gómez-Mejía et al., 2007; Luo & Chung, 2005; Peng & Jiang, 2010). However, prior to this research, the academic work done on this control mechanism in family business groups was limited. Based on a longitudinal, multi-level analysis of datasets for 12 different years, this research uses the agency theory viewpoint to address the importance and unique character of the coordination and control issue for family business groups in Asia. The results reveal that the different ownership structures of affiliate firms influence the likelihood that family members will hold leadership positions in firms associated with a family business group. Specifically, if the founding family enjoys more direct ownership of the affiliate firm, the family will be likely to appoint a trusted family member to the leadership position within the affiliate firm. However, when the founding family has a greater degree of pyramidal ownership of an affiliate firm, the family will be less likely to appoint a family leader from within. Additionally, family leadership mediates partially the relationship between the ownership structure Affiliate’s Performance Constant M5 .039 (.192) M6 .020 (.193) M7 −.016 (.193) M8 .420 (.208)* .105 (.056)† M: Family leadership M9 M10 .416 (.208)* .406 (.209)† .128 (.056)* .132 (.057)* −.371 (.072)* −.365 (.072)* Independent Variables X1: Family direct ownership .235 (.216)* .287 (.217) X2: Family pyramidal ownership .140 (.219) −.342 (.071)** Control Variables Year 1990 .518 (.183)** .513 (.183)** .522 (.183)** .464 (.183)* .472 (.182)** Year 1992 .475 (.170)** .463 (.171)** .468 (.171)** .409 (.170)* .411 (.170)* .469 (.182)** .411 (.170)* Year 1994 .578 (.158)** .580 (.159)** .584 (.159)** .523 (.158)** .524 (.158)** .523 (.158)** Year 1996 .565 (.149)** .576 (.150)** .573 (.150)** .527 (.149)** .520 (.149)** .522 (.149)** Year 1998 .504 (.139)** .494 (.141)** .496 (.141)** .459 (.140)** .459 (.140)** .459 (.140)** Year 1999 .091 (.105) .890 (.106) .088 (.106) .085 (.105) .083 (.106) .085 (.106) Year 2000 −.048 (.092) −.048 (.092) −.046 (.093) −.041 (.092) −.037 (.092) −.035 (.092) Year 2001 −.195 (.090)* −.194 (.091)* −.196 (.091)* −.195 (.090)* −.196 (.090)* −.194 (.091)* Year 2002 −.061 (.091) −.058 (.091) −.060 (.091) −.049 (.091) −.051 (.091) −.049 (.091) Year 2003 −.204 (.089)* −.204 (.089)* −.203 (.089)* −.196 (.089)* −.194 (.089)* −.193 (.089)* Affiliate’s assets .834 (.012)** .835 (.012)** .840 (.012)** −.821 (.012)** .827 (.012)** .827 (.012)** Affiliate’s age .358 (.027)** .360 (.028)** .366 (.028)** .350 (.028)** .355 (.028)** .355 (.028)** Affiliate’s location −.086 (.059) −.092 (.060) −.091 (.060) −.150 (.061)* −.154 (.061)* Ownership structure, family leadership, and performance… Table 3 The mediating role of family leadership in ownership structure and affiliate firm’s performance in family business groups. −.155 (.061)* Manufacturing affiliate .747 (.056)** .745 (.057)** .741 (.057)** .726 (.057)** .720 (.057)** .720 (.057)** Control variables-Random parts .012 (.002)** .012 (.002)** .012 (.002)** .011 (.002)** .011 (.002)** .011 (.002)** 321 322 Table 3 (continued). Affiliate’s Performance M5 M6 M7 M8 M9 M10 Wald Chi-Square 7759.73** 7708.06** 7666.69** 7774.79** 7744.44** 7743.77** Number of Observations 5,892 5,841 5,813 5,840 5,813 5,813 Number of Groups 21 21 21 21 21 21 The random effect contains group’s size, group’s age, number of affiliates, founding family leadership, and percentage of family member in inner circle. Standard deviation data given in parentheses † p ≤ .10, * p ≤ .05, ** p ≤ .01. H.-M. Chung, S.-T. Chan Ownership structure, family leadership, and performance… 323 and an affiliate firm’s performance in a family business group. Family leadership mediates the positive relationship of direct family ownership on affiliate firm performance; however, family leadership does not mediate the negative relationship between pyramidal ownership and affiliate firm performance in a family business group. Compared with the direct family ownership, pyramidal ownership will generate a relatively lower financial risk situation for founding family in a family business group. The more pyramidal ownership in an affiliate firm will lower the likelihood of family leadership choice, and accordingly the possibility that family leadership will generate positive motivation in affiliate firm’s management and performance will be reduced. Evidences indicate that family leadership cannot account for the negative affiliate firm performance from pyramidal ownership in a family business group. The possible solutions in addressing the pyramidal ownership-performance issues will be raised after. Contributions The results, which are consistent with those of previous research (e.g., Chu, 2011; Cruz et al., 2010), provide further insight into agency theory, expanding it by explaining the mechanisms of coordination and control in family businesses, especially the complex network among affiliate firms in a family business group. Examining unbalanced control by considering direct and pyramidal ownership at the affiliate firm level will provide insights into the use of agency theory to analyze family business groups. The agency or performance issues associated with pyramidal ownership at the group level are the typical concerns (Almeida & Wolfenzon, 2006; Morck et al., 2005; Young et al., 2008). Performance issues associated with both direct and pyramidal ownership at the affiliate firm level are particularly a key concern in a family business group (Carney et al., 2011; Chang, 2003). Thus, examining the possible performance implications of family leadership under the direct and pyramidal ownership structures in a family business group can provide more information about control and leadership issues in family business groups, which are the typical governance structure in the Asian context (Carney et al., 2011; Yiu et al., 2007). In short, the findings presented in this study provide a better theoretical understanding of leadership issues and results associated with family business groups. The findings presented here also indicate the strategic implications of ownership structure and family leadership in the context of family businesses (e.g., Miller et al., 2008; Peng & Jiang, 2010; Villalonga & Amit, 2006). Ownership structure and family leadership need to be considered carefully in addressing performance. The personal trust established via family ties is good for intra-business and intra-business group coalitions, and these ties are a strategic choice under different ownership structures. We know from this study that family leadership will be conducive to improve the affiliate firm’s shareholder value under specific conditions, such as family direct ownership concern, in a family business group. However, we also believe that this kind of leadership is not the only solution to improve affiliate firm performance in a family business group. Indeed, besides the family leadership concern on performance issues in family business groups, there are still other kinds of external and internal governance arrangements, such as independent directors in 324 H.-M. Chung, S.-T. Chan board of directors (Anderson & Reeb, 2004), or the presence of legal institutions and multiple blockholders structure (Jiang & Peng, 2011), that can provide positive performance implications for majority and minority shareholders in family businesses or family business groups. The longitudinal evidence from this study indicates that family leadership can provide positive performance implications in Asian family business groups, but this study also leaves answers to the advanced performance issues that concern the relationship between family leadership and institutional arrangements. Future work can address performance issues to illustrate the interwoven nature among family business’s ownership structure, leadership choices, performance implications, and the institutional contexts, especially in the Asian region that is characterized by divergent cultural backgrounds and institutional contexts (Ahlstrom, Chen, & Yeh, 2010; Bruton & Lau, 2008; Jiang & Peng, 2011). The results of this research may also be useful in other domains. One is in business groups, more broadly, as noted in the introduction. This application is especially helpful because even though such groups are the dominant organizational form outside North America, they still remain poorly understood. The evidence presented here facilitates a better understanding of coordination and control issues in such organizations (Carney et al., 2011). Future work to examine the relationships among ownership structure, family leadership, and affiliate firm performance in distinctive institutional contexts can provide more fruitful understandings in coordination and control issues in business groups in Asia or around the world (Jiang & Peng, 2011; Peng & Jiang, 2010; Yiu et al., 2007). In addition, the use of longitudinal and multi-level data provides a rare look into the performance of such business groups over time. To date, there has been little effort to conduct this type of analysis. Limitations and future research directions Since this research relies on secondary data, there are several limitations. For example, we cannot identify the influence of family dynamics or even the family leadership style on coordination and control in family business groups. Future research might consider the family dynamics influence or family leadership style from the questionnaire to create a comprehensive model of control and coordination within family businesses. Moreover, this research provides possible answers on addressing the antecedents and consequences of family leadership in family business groups; future work can address family leadership types-outcomes issues from questionnaires to advances the understanding of family leadership styles and the distinctive macro outcomes in family businesses or family business groups. For example, previous studies have provided fruitful evidences on addressing the paternalistic leadership style-micro outcomes issues in Asian family businesses, especially focusing on Chinese family businesses (Cheng et al., 2000, 2002a, 2002b; Farh et al., 2006). Future work can examine the paternalistic leadership style-macro outcomes relationships in Asian family business enterprises to provide more understanding on family leadership issues in the Asian Pacific region. Besides the leadership-macro outcomes issues, what leadership patterns (e.g., transformational, transactional, and ethical leadership) Ownership structure, family leadership, and performance… 325 are more conducive in enhancing shareholder’s value when family members are appointed as leaders is also an advanced issue in the family business field. Additionally, this study’s exclusive use of data from one country may create concerns regarding generalization. Comparative research on family businesses in Asia may help researchers to further advance our understanding of coordination and control in Asian family businesses. Furthermore, comparative studies to address the performance issues by family leadership choice under differential institutional contexts will be helpful to provide solutions in mitigating the principal-principal concerns in Asian family business groups. This research has laid an important foundation for the better understanding of family businesses, especially in Asia. Future research should continue to expand our understanding of this issue given that the importance and impact of such firms will likely only grow in the future as the economic power of this region increases. Moreover, this research also advances our knowledge of leadership. Future research can continue to probe the importance of leadership within family enterprises. Appendix Example of direct and pyramidal ownership structure in family business group In this Appendix, we illustrate the example raised under the Independent variables section with a picture. The founding family controls the target affiliate firm A by three kinds of chain in the chain of voting rights (i.e., chain a, chain b, and chain c). The calculation of family direct ownership and pyramidal ownership structure are indicated below. (1) Family direct control (chain a) is the shares directly owned by family members. Thus, family direct control of firm A is 25%. (2) Family pyramidal control (chain b and chain c) is the sum of shares that crossshareholding through more than one listed affiliates in the same business group. In this example, family pyramidal control of firm A includes chain b (in this chain, family controls 25% rights of firm A which control 100% rights of firm B, and firm B in turn has 17% rights of firm A) and chain c (in this chain, family has 25% of firm C, and firm C controls 7% of firm A). According to La Porta et al. (1999, 2002), the highest minimum voting stake along the control chain was picked when the firm’s shareholders have over 10% voting rights. However, because we illustrate all possible controlling situations of family members and not just focus on large or public firms, we select the minimum voting rights along the control chain and do not take 10% voting rights as picking limitation. This argument is similar with Claessens et al.’s (2000, 2002) study. Based on Claessens et al.’s (2000, 2002) study, the weakest link in the chain of voting rights means the minimal voting rights across the ownership chain. Therefore, the weakest link of chain b is 17% (the minimum of 25%, 100%, and 17%) and of chain c is 7% (the minimum of 25% and 7%). We would say that the family controls 17% of firm A along the control chain b and 7% of firm A along the control chain c. Thus, the total 326 H.-M. Chung, S.-T. Chan family pyramidal control of firm A is 24% which is the sum of 17% from chain b and 7% from chain c. 25 100% a Firm A b c Family 7% 25 17 Firm B Firm C 49 Family direct control (from chain a): 25% Family pyramidal control (17% from chain b plus 7% from chain c): 24% References Ahlstrom, D., Chen, S.-j., & Yeh, K. S. 2010. Managing in ethnic Chinese communities: Culture, institutions, and context. Asia Pacific Journal of Management, 27(3): 341–354. Almeida, H., & Wolfenzon, D. 2006. A theory of pyramidal ownership and family business group. Journal of Finance, 61: 2637–2680. Anderson, R. C., & Reeb, D. M. 2003. Founding-family ownership and firm performance: Evidence from the S&P 500. Journal of Finance, 58: 1301–1329. Anderson, R. C., & Reeb, D. M. 2004. Board composition: Balancing family influence in S&P 500 firms. Administrative Science Quarterly, 49: 209–237. Arregle, J.-L., Hitt, M. A., Sirmon, D. G., & Very, P. 2007. The development of organizational social capital: Attributes of family firms. Journal of Management Studies, 44: 73–94. Baron, R. M., & Kenny, D. A. 1986. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51: 1173–1182. Brown, M., & Forsythe, A. 1974. Robust test for the equality of variances. Journal of the American Statistical Association, 69(346): 364–367. Bruton, G. D., & Lau, C. M. 2008. Asian management research: Status today and future outlook. Journal of Management Studies, 45: 636–659. Bryk, A. S., & Raudenbush, S. W. 1992. Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage. Carney, M., Gedajlovic, E. R., Heugens, P. P. M. A. R., van Essen, M., & van Oosterhout, J. 2011. Business group affiliation, performance, context, and strategy: A meta-analysis. Academy of Management Journal, 54: 437–460. Certo, S. T., & Semadeni, M. 2006. Strategy research and panel data: Evidence and implications. Journal of Management, 32: 449–471. Chang, S. J. 2003. Ownership structure, expropriation, and performance of group-affiliated companies in Korea. Academy of Management Journal, 46: 238–253. Chang, S. J. (Ed.). 2006. Business groups in East Asia: Financial crisis, restructuring, and new growth. New York: Oxford University Press. Chen, C. C., Chen, Y. R., & Xin, K. 2004. Guanxi practices and trust in management: A procedural justice perspective. Organization Science, 15: 200–209. Ownership structure, family leadership, and performance… 327 Cheng, B. S., Chou, L. F., & Farh, J. L. 2000. A triad model of paternalistic leadership: The constructs and measurement. Indigenous Psychological Research in Chinese Societies, 14: 3–64. Cheng, B. S., Huang, M. P., & Chou, L. F. 2002a. Paternalistic leadership and its effectiveness: Evidence from Chinese organizational teams. Journal of Psychology in Chinese Societies, 3: 85–112. Cheng, B. S., Shieh, P. Y., & Chou, L. F. 2002b. The principal’s leadership, leader-member exchange quality, and the teacher’s extra-role behavior: The effects of transformational and paternalistic leadership. Indigenous Psychological Research in Chinese Societies, 17: 105–161. China Credit Information Service. 1988. 1990/1991 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 1990. 1992/1993 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 1992. 1994/1995 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 1994. 1996/1997 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 1996. 1998/1999 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 1998. 2000 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 1999. 2001 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 2000. 2002 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 2001. 2003 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 2002. 2004 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 2003. 2005 business groups in Taiwan. Taipei: CCIS. China Credit Information Service. 2004. 2006 business groups in Taiwan. Taipei: CCIS. Chu, W. 2011. Family ownership and firm performance: Influence of family management, family control, and firm size. Asia Pacific Journal of Management, 28(4): 833–851. Chung, C. N. 2003. Managerial structure of business groups in Taiwan: The inner circle system and its social organization. Developing Economies, 41: 37–64. Claessens, S., Djankov, S., Fan, J., & Lang, L. H. P. 2002. Disentangling the incentive and entrenchment effects of large shareholdings. Journal of Finance, 57: 2741–2771. Claessens, S., Djankov, S., & Lang, L. H. P. 2000. The separation of ownership and control in East Asian corporations. Journal of Financial Economics, 58: 81–112. Cruz, C., Gómez-Mejía, L. R., & Becerra, M. 2010. Perceptions of benevolence and the design of agency contracts: CEO-TMT relationships in family firms. Academy of Management Journal, 53: 69–89. Dansereau, F., Yammarino, F. J., & Kohles, J. C. 1999. Multiple levels of analysis from a longitudinal perspective: Some implications for theory building. Academy of Management Review, 24: 346–357. Demsetz, H., & Lehn, K. 1985. The structure of corporate ownership: Causes and consequences. Journal of Political Economy, 93: 1155–1177. Dyer, W. G., Jr. 1986. Cultural change in family firms: Anticipating and managing business and family transitions. San Francisco, CA: Jossey-Bass. Eisenhardt, K. M. 1989. Agency theory: An assessment and review. Academy of Management Review, 14: 57–74. Fama, E. F., & Jensen, M. C. 1983. Agency problems and residual claims. Journal of Law and Economics, 26: 327–349. Farh, J. L., Cheng, B. S., Chou, L. F., & Chu, X. P. 2006. Authority and benevolence: Employees’ responses to paternalistic leadership in China. In A. S. Tsui, Y. Bian & L. Cheng (Eds.). China’s domestic private firms: Multidisciplinary perspectives on management and performance: 230–260. New York: Sharpe. Gómez-Mejía, L. R., Haynes, K. T., Núñez-Nickel, M., Jacobson, K. J. L., & Moyano-Fuentes, J. 2007. Socioemotional wealth and business risks in family-controlled firms: Evidence from Spanish olive oil mills. Administrative Science Quarterly, 52: 106–137. Gómez-Mejía, L. R., Núñez-Nickel, M., & Gutierrez, I. 2001. The role of family ties in agency contacts. Academy of Management Journal, 44: 81–95. Granovetter, M. 1985. Economic action and social structure: The problem of embeddedness. American Journal of Sociology, 91: 481–510. Granovetter, M. 1995. Coase revisited: Business groups in a modern economy. Industrial and Corporate Change, 4: 93–140. Hamilton, G. G. 1997. Organization and market processes in Taiwan’s capitalist economy. In M. Orrù, N. W. Biggart & G. G. Hamilton (Eds.). The economic organization of East Asian capitalism: 237– 293. Thousand Oaks, CA: Sage. IMD (International Institute for Management Development). 2006. World competitiveness yearbook 2006. Lausanne, CH: IMD. Ismail, K., & Ford, D., Jr. 2010. Organizational leadership in central Asia and the Caucasus: Research considerations and directions. Asia Pacific Journal of Management, 27(2): 321–340. 328 H.-M. Chung, S.-T. Chan Jensen, M. C., & Meckling, W. 1976. Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics, 3: 305–360. Jiang, Y., & Peng, M. W. 2011. Principal-principal conflicts during crisis. Asia Pacific Journal of Management, 28(4): 683–695. Kenny, D. A., Kashy, D., & Bolger, N. 1998. Data analysis in social psychology. In D. Gilbert, S. Fiske & G. Lindzey (Eds.). Handbook of social psychology, 4th ed.: 233–265. New York: McGraw-Hill. Khanna, T., & Rivkin, J. W. 2001. Estimating the performance effects of business groups in emerging markets. Strategic Management Journal, 22: 45–74. Kim, D., & Sung, T. 2009. Cross-ownership, takeover threat and control benefit. Applied Financial Economics, 19: 659–667. Lam, L. W., Loi, R., & Leong, C. 2012. Reliance and disclosure: How supervisory justice affects trust in supervisor and extra-role performance. Asia Pacific Journal of Management. doi:10.1007/s10490-0119249-5. La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. 1999. Corporate ownership around the world. Journal of Finance, 54: 471–517. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. W. 2002. Investor protection and corporate valuation. Journal of Finance, 57: 1147–1170. Lasserre, P., & Schütte, H. 2006. Strategies for Asia Pacific: Meeting new challenges, 3rd ed. London: MacMillan Press Ltd. Lau, D. C., Liu, J., & Fu, P. P. 2007. Feeling trusted by business leaders in China: Antecedents and the mediating role of value congruence. Asia Pacific Journal of Management, 24(3): 321–340. Leung, K. 2007. The glory and tyranny of citation impact: An East Asia perspective. Academy of Management Journal, 50: 510–513. Levy, M. 2009. Control in pyramidal structures. Corporate Governance: An International Review, 17: 77– 89. Liu, W., Yang, H., & Zhang, G. 2012. Does family business excel in firm performance?. An institutionbased view. Asia Pacific Journal of Management. doi:10.1007/s10490-010-9216-6. Luo, X., & Chung, C. N. 2005. Keeping it all in the family: The role of particularistic relationships in business group performance during institutional transition. Administrative Science Quarterly, 50: 404–439. Meyer, K. E. 2006. Asian management research needs more self-confidence. Asia Pacific Journal of Management, 23(2): 119–137. Miller, D., & Le Breton-Miller, I. 2005. Managing for the long run: Lessons in competitive advantage from great family businesses. Boston: Harvard Business School Press. Miller, D., Le Breton-Miller, I., & Scholnick, B. 2008. Stewardship vs. stagnation: An empirical comparison of small family and non-family businesses. Journal of Management Studies, 45: 51–78. Ministry of Finance. 2008. Values of major export commodities. http://www.mof.gov.tw/public/Data/ statistic/trade/2370.htm, Accessed Nov. 11, 2008. Mishra, C., & McConaughy, D. 1999. Founding family control and capital structure: The risk of loss of control and the aversion to debt. Entrepreneurship Theory Practice, 23: 53–64. Morck, R. K. 1996. On the economics of concentrated ownership. Canadian Business Law Journal, 26: 63–85. Morck, R., Wolfenzon, D., & Yeung, B. 2005. Corporate governance, economic entrenchment, and growth. Journal of Economic Literature, 43: 655–720. Pellegrini, E. K., & Scandura, T. A. 2008. Paternalistic leadership: A review and agenda for future research. Journal of Management, 34: 566–593. Peng, M. W., & Jiang, Y. 2010. Institutions behind family ownership and control in large firms. Journal of Management Studies, 47: 253–273. Pfeffer, J. 1997. New directions for organization theory—Problems and prospects. Oxford: Oxford University Press. Rabe-Hesketh, S., & Skrondal, A. 2008. Multilevel and longitudinal modeling using Stata, 2nd ed. College Station, TX: Stata Press Publication. Sanders, W. M. G., & Hambrick, D. C. 2007. Swinging for the fences: The effects of CEO stock option options on company risk taking and performance. Academy of Management Journal, 50: 1055–1078. Schulze, W. S., Lubatkin, M. H., Dino, R. N., & Buchholtz, A. K. 2001. Agency relationships in family firms: Theory and evidence. Organization Science, 12: 99–116. Scott, W. R. 1977. Effectiveness of organizational effectiveness studies. In P. S. Goodman & J. M. Pennings (Eds.). New perspectives on organizational effectiveness: 63–95. San Francisco, CA: Jossey-Bass. Sobel, M. E. 1982. Asymptotic intervals for indirect effects in structural equations models. In S. Leinhart (Ed.). Sociological methodology 1982: 290–312. San Francisco, CA: Jossey-Bass. Ownership structure, family leadership, and performance… 329 Sorenson, R. L. 2000. The contribution of leadership style and practices to family and business success. Family Business Review, 13: 183–200. Tjosvold, D., Wong, A., & Hui, C. 2004. Leadership research in Asia-developing relationships. In K. Leung & S. White (Eds.). Handbook of Asian management: 373–395. Boston: Kluwer. Vallejo, M. C. 2009. Analytical model of leadership in family firms under transformational theoretical approach: An exploratory study. Family Business Review, 22: 136–150. Villalonga, B., & Amit, R. 2006. How do family ownership, control and management affect firm value?. Journal of Financial Economics, 80: 385–417. Wang, H., Law, K. S., Hackett, R. D., Wang, D., & Chen, Z. X. 2005. Leader-member exchange as a mediator of the relationship between transformational leadership and followers’ performance and organizational citizenship behavior. Academy of Management Journal, 48: 420–432. White, S. 2001. Asian management matters: Regional relevance and global impact. Asia Pacific Journal of Management, 18(1): 121–124. Yiu, D., Lu, Y., Bruton, G. D., & Hoskisson, R. E. 2007. Business groups: An integrated model to focus future research. Journal of Management Studies, 44: 1551–1579. Young, M. N., Peng, M. W., Ahlstrom, D., Bruton, G. D., & Jiang, Y. 2008. Corporate governance in emerging economies: A review for the principal-principal perspective. Journal of Management Studies, 45: 196–220. Hsi-Mei Chung (PhD, National Sun Yat-sen University) is an associate professor of business management at Department of Business Administration, I-Shou University, Kaohsiung, Taiwan. She also participates in the family business research at Electronic Commerce Research Center, National Sun Yat-sen University, Kaohsiung, Taiwan. Her research interests are focused on strategy and performance issues in the family business and business group field. She is invited to serve as the guest editor of the Special Issue on Corporate Governance and Strategic Implementation, Organization and Management (in Chinese). Her publications can be found in the Asia Pacific Journal of Management, Multinational Business Review, Organization and Management (in Chinese), Industry and Management Forum (in Chinese). She also has three book chapters on Family Enterprising in Asia: Where East meets West, Advances in Mergers and Acquisitions, and The Handbook of Research on Asian Entrepreneurship. Shu-Ting Chan is a PhD student at Department of Business Management, National Sun Yat-sen University, Kaohsiung, Taiwan. She is also an assistant researcher of family business at Electronic Commerce Research Center, National Sun Yat-sen University, Kaohsiung, Taiwan. Her research interests are focused on strategic management and control issues in the family business and business group field. She has two conference papers in the Asia Academy of Management 2010 Conference and Special Issue Conference on Global Financial Crisis, Corporate Governance: An International Review.

RELATED PAPERS

RELATED TOPICS

Log In

Ownership structure, family leadership, and performance of affiliate firms in large family business groups

Ownership structure, family leadership, and performance of affiliate firms in large family business groups

Related Papers

RELATED PAPERS

RELATED TOPICS