THE DETERMINANTS OF THE SUCCESSION IN TAIWAN’S FAMILY BUSINESS GROUPS

Based on the panel data analysis of Taiwan’s family business groups from 1988 to 2002, this research attempts to investigate the relationships among overlapping investment, use of particularistic ties, group performance, and succession in family business group. The results show that the family business group’s overlapping investment between the owner-managers and family members occupying the decisive positions of group affiliates significantly influence its leader change. This study highlights the importance of alternative control choices within the family business. Furthermore, it also provides a good comparing start-point for researches interested in understanding the succession issue of Chinese family business in Great China.


1.Introduction
Business group is a conglomerate of enterprises interlocked through overlapping investments by a small number of core owner-managers (Numazaki, 1986;Granovetter, 1995;Yiu, Bruton, & Lu, 2005;Chang, 2006).This enterprise system is a cluster of enterprises owned and controlled by a group of persons tied by a network of various relationships, such as kinship, marriage, friends and other social bonds (Hamilton, 1997;Kao, 1996;Luo & Chung, 2005;Numazaki, 1986;Redding, 1990;Wong, 1985).Scholars have researched this organizational from in developed, developing, and newly industrialized economies (Granovetter, 1995;Shiba & Shimotani, 1997).In a family-owned business, family members are involved in the ownership, control, and management of the business group.In contrast to other business groups, there is no separation of the roles.Previous studies presented that Taiwanese listed companies and Taiwanese business groups are typically governed by family control (e.g., Luo & Chung, 2005;Yeh, 2003).Even after a company goes public, family ownership or control tends to play a dominant role in the decision-making process (Liu & Yeh, 2000).
In Taiwan, among various kinds of relationship, family-owned business groups, constituted by kinship and marriage, are the typical type.Researches highlighted the family business is a basic constitution for Chinese business (Amyot, 1973;Hamilton, 1997;Kao, 1996;Kiong, 1996;Luo & Chung, 2005;Redding, 1990;Whitley, 1992;Wong, 1985).Researcher even described the Chinese type of capitalism in Taiwan, -guanxi capitalism‖ (Hamilton, 1997), emphasizing the importance of personal connections among business and family.
In discussing the issues confronting family business, succession process, i.e., the business is transferred from one generation to the next, is the most critical one (e.g., Gersick et al., 1997;Lee, Lim, & Lim, 2003).Regarding how family business transits successfully over time, most researches focused on the family dynamic or the psychological problems in the family business.For example, the interdependencies among family members (Beckhard & Dyer, 1983), the psychological and emotional problems experienced by the owner-parent and the person's children, such as the stress, fear of abandonment, mutual role adjustment et al. (e.g., Handler, 1990;Kepner, 1983; Kets de Vires, 1993), the development and structure of a formal succession plan (e.g., Kets de Vires, 1993; Sharma, Chrisman, & Chua, 2003; Ward, 1987) and the demographic characteristic of the family-business owners (e.g., Bates, 1990;Cooper & Artz, 1995) seem as the key determinants in family business succession.Almost all the related literatures about this issue concern the internal relationship and conditions of the focal family business; little is regarding the influence from the family ownership and control.Claessens et al. (2000) found that control is enhanced through pyramid structures and cross-holdings among firms in all the East Asian countries studied, so that even where control by founding families is not direct, the families have retained control through indirect shareholding networks.The overlapping investment established by the core owner-managers is the foundation of interlocking ties among the affiliates within the group.Although the family values associated with the family ties distinguishes the family business from others when discussing Taiwanese business groups (Hamilton, 1997;Luo & Chung, 2005;Redding, 1990;Whitley, 1992), differences exist between business groups established on the basis of different types of guanxi in the control and decision-making preferences, including those relating to diversification (e.g., Chung, 2006;Fox & Hamilton, 1994) or the succession in this case.
In addition to ownership and control considerations, the group performance also merits investigation because previous studies have highlighted that the poor performance is the possible trigger for succession (e.g., Coughlan & Schmidt, 1985;Weisbach, 1988;Murphy & Zimmerman, 1993).Specifically, does the poor performance mitigate the family controlling power in family business succession?If the answer is yes, it means that although the entrenchment of the family is the key concern in family business, this family business still need to face the poor performance, and to deal with the possible challenges from the non-family shareholders.However, if the answer is no, it means that the good performance is not the only goal for family business.The family members will accept the poor performance to keep the family control over the business group (Gómez-Mejía et al., 2001;Gómez-Mejía et al., 2007).
The goal of this paper is to examine whether the succession is affected by overlapping investment and by the family control power, and whether poor performances interact with overlapping investment and family control power to influence the replacement of the leader in family business group.Investigating the possible relationships among those factors is not only meaningful in its own right, it also provides a good comparing start-point for researches interested in understanding the succession issue of Chinese family business in Great China.

2.Literature Review 2.1.Overlapping investment and CEO succession in family business group
The major control and management mechanism inside the business group is hinged upon a group of closely related core players, the so-called inner circle (Useem, 1984).The members of the inner circle in Taiwan's business groups coordinate group business by overlapping investment in various group firms.The overlapping investment between owner-managers indicates the degree of interlocking shareholders among the various affiliates in a business group.By Numazaki's (1986) definition, a business group is a conglomerate of enterprises interlocked through overlapping investments by a small number of core owner-managers.Therefore, to understand the operation of business group, examining the overlapping investment among core players is a crucial task.
In family business group, although the core owner-managers who are bonded by particularistic relationship (such as family and prior social ties) are the majority shareholders (Granovetter, 1995), there still exist other minority shareholders in the board of directors within each affiliate.With the opening-up of capital markets and economic liberalization in Taiwan (Tien, 1989) and with the expansion and diversification of group, the family business group might bring in external investors, which would induce a more dispersed and open structure of group shareholdings.As a result, family ownership would be diluted and the size of the inner circle would be enlarged.Chung's (2004) research showed that the institutional changes in Taiwan between 1987 and 1993 induced more institutional investors and inter-group cross-shareholding, and this started blurring the group boundaries.
In order to achieve dominant control, the original inner circle members would occupy overlapped shareholder positions in various affiliates.Furthermore, they would entice different external investors to join the group so that these external shareholders would not cross-sharehold the affiliates.In other words, the higher degree of overlapping investment over the core owners-managers in the family business group is a kind of alternative method for the family to control over the business group.Interlocking shareholdings among members of the inner circle contribute to a cohesive business group; from this perspective, overlapping investment provides a mechanism for the members that help to preserve their control power.When the degree of overlapping investment within a business group is high, there exist a greater number of common shareholders among the various affiliates, of which other non-owner-managers can hardly challenge the status of the original controllers.Thus we hypothesize under the situation that the overlapping investment within a family business group represents a higher degree, the possibility of the succession is lower.H1: The higher degree of overlapping investment between owner-managers, the lower possibility that the family business group have the succession.

2.2.Family control and CEO succession in family business group
In the management structure of Taiwan's business groups, particularistic ties between the core individuals in decision-making are prominent and are related with a degree of centrality in decision-making (  1996).
One distinctive feature of family businesses is that they emphasize personal authority and a basis of trust in decision-making (Miller & Le Breton-Miller, 2005;Weidenbaum, 1996;Whitley, 1992).In Chinese societies, social relationships are structured in concentric circles, with family members in the innermost circle and strangers in the outermost circle (Fei, 1992;Tsui and Farh, 1997).For Taiwanese business groups, family members distrust people outside their family and kinship group (Fukuyama, 1995); thus, they intend to occupy leadership positions in various group firms to keep the ultimate control.That is, Chinese family business groups ensure that coordination and control is appropriate by placing family members in key managerial positions (Hamilton, 1997;Luo & Chung, 2005).
However, with the expansion of business group, the family faces the problem of maintaining a controlling stake in the business group.Changing institutional and market conditions led to family members never fully occupied the inner circle.The higher degree of family members occupied in the managerial positions indicates that the family members can obtain higher control power over the group.If there are more family members occupy the key managerial positions within the family business groups, the family can keep and entrench the control over the group, and then there is no emerging requirement to let the succession happened.Therefore, we propose that the higher the percentage of family members as presidents of affiliates within the business group, the lower the possibility of the succession happened.

H2: The higher proportion of family members as the
presidents, the lower possibility that the family business group have the succession.

2.3.The moderating effect of previous performance in family business group
One of the reasons that scholars have long been interested in top executives succession is they seek to answer the question of whether the performance of CEO does matter?If the top management possesses merely symbolic value, CEO turnover would not relate to the firm performance.But if the top management has instrumental value, the past performance of corporation should be the important factor of changing firm leader.CEOs are typically blamed for poor performance, removed for unsatisfactory performance, and replaced by top managers who are expected to improve performance (Boeker, 1992).Indeed, executive turnover has been found to be more likely as corporate performance falls by several empirical studies (Coughlan & Schmidt, 1985;Weisbach, 1988;Murphy & Zimmerman, 1993).To replace a CEO with poor performance seems to be a common rule for all economy systems.Kaplan (1994) finds that, given the large differences in the corporate governance systems, the probabilities of CEO turnover in Germany, Japan and the United States are similar; the increase in the likelihood of turnover due to poor returns and income losses is almost the same in the three countries.As, Kesner & Sebora (1994: 356) in their review of this literature conclude: ‗In instances where succession frequency was treated as the dependent variable, the findings were consistentsuccession rates were higher in low performing firms'.
Although the general finding that poor prior performance is associated with succession continues to be robust, the relationship between CEO succession and organizational performance is complicated.To advance our knowledge, the field has begun to emphasize performance as a moderator (Giambatista, Rowe, & Riaz, 2005).The arguments presented above show that the higher degree of overlapping investment between owner-managers and the higher proportion of family members as presidents of group firms are expected to have a lower chance on CEO succession.However, if group performance is poor, other members in the inner circle may contest the incumbent leader's power.Poor group performance feeds power struggles and political fighting at the top of business group (Eisenhardt & Bourgeois, 1988) that increases the possibility of CEO succession.Following this line of reasoning, we assert that the effects of overlapping investment and family members as the presidents of group affiliates on the likelihood of CEO succession will be moderated by prior group performance.H3a: The negative relationship between overlapping investment and succession will be weaker at lower levels of prior group performance.H3b: The negative relationship between family members as the presidents and succession will be weaker at lower levels of prior group performance.

3.Research Method 3.1.Sample Selection
The  1990,1992,1994,1996,1998,2000,2001,2002,2003,2004), the current research tests the above mentioned hypotheses in the context of Taiwan.Taiwan is one of the 15 largest economies in the world (IMF, 2007).Furthermore, in the Asia pacific markets, Taiwan is one representative model in the newly industrialized economies (NIEs, South Korea, Taiwan, Hong Kong and Singapore) (Lasserre and Schütte, 2006; Leung, 2006).The focus here is on the top-100 business groups since they play such a critical economic role in the Taiwan economy; accounting for above 70% the island's GNP during the time period examined (China Credit Information Service, 2004).The business group in Taiwan dictionary is published per two years before 2000.Therefore, there are 10 data-points during the examined 14 years.Employing a longitudinal sample can result in business groups moving in or dropping out of the top-100 specification for business groups.To ensure that the impact of the enterprise is not a temporary fluke we require that the family business group be part of the top-100 business groups for at least 8-years during the 14-years time-periods to be examined.The result was a sample of 35 family business groups.Furthermore, this study examines these family business groups over the period of 1988 to 2002.It is determined to examine this time period since there are succession events happened among family business groups after 1990.

3.2.Variables
Dependent Variable (Y): The CEO Succession in Family Business Group.The dependent variable is measured the CEO succession events in a family business group over the 14 years.The succession affair is defined as the change of CEO position in the core company of business group.In addition, the successor of the CEO position in the core company of business group can be either the family member or the non-family professional manager.Therefore, this variable can be coded from 0, 1 and 2 numerical to identify that the successor in the CEO position is a family person or not.If the CEO successor is a family member, it is coded -1‖; if he (or her) is a non-family professional manager, it is coded -2‖.Furthermore, if there is no CEO succession affair in a specific year, it is coded -0‖.The identification of the successor is family one or not is by identifying that the CEO in the core company has the family tie with the founder or not, such as the son in the next generation or the younger brother in the same generation.Independent Variable 1 (X1) (Overlapping Investment): The Degree of Overlapping Investment between Owner-Managers in Family Business Group.The overlapping investment between owner-managers indicates the degree of interlocking shareholders among the various affiliates in a business group.By definition, a business group is a conglomerate of enterprises (or affiliates) interlocked through overlapping investments by a small number of core owner-managers (Numazaki, 1986).The overlapping investment established by the core owner-managers is the foundation of interlocking ties among the affiliates within the group.However, as there is a -borrowing name‖ phenomenon across firms in Taiwan, the core owner-managers within a business group will possibly borrow each other's name to represent the board members.The use of -degree of interlocking shareholders between each affiliate‖ tries to capture the concept of overlapping investment for each business group.
For each affiliate in a business group, the names of the key shareholders are listed, and hence it is possible to identify the existence of common key shareholders between different affiliates.In this study, this variable was measured by computing for each affiliate the number of its key shareholders who were also key shareholders in other affiliates.This number was then accumulated for all of the affiliates in the business group and then weighted by the total number of affiliates in the business group.For example, if there were 30 affiliates in a business group (including the core company), this variable was computed by accumulating the number of common key shareholders for each of the 30 affiliates, and then divided this numerical value by 30 (the total number of affiliates).The aim of the weighting process is to take account of the fact that the total number of common shareholders in a business group is a function of the number of affiliates within that business group.The numerical value, if not scaled by the number of group affiliates, will be larger when there are more affiliates within a group.Therefore, the number of affiliates must be weighted to enable a meaningful comparison to be made between business groups that vary in their number of affiliates.

Independent Variable 2 (X2) (PFO): The Family Business Groups' Use of Family Ties in
Affiliates.The identification of the family business groups' use of particularistic ties in affiliates is based on determining whether the CEOs of their total affiliates have family ties.If the CEO is a family member, he or she is coded as the one having particularistic ties with the members of the inner circle.We identify the affiliate's CEO's family tie by using multiple sources of secondary databases.For example, if the affiliate's CEO is described as -having kinship relationships with the founding family members in the group,‖ this affiliate's CEO is identified as possessing particularistic ties; in that, family tie in this study.In family business groups, the business groups' use of family ties in its affiliates is measured by the percentage of the CEO's having family ties divided by the total number of affiliates.This index can represent the relative importance of family ties in managing the affiliates for specific family business groups.
Interaction Variables.This research utilizes two kinds of interaction term to identify the possible interactive influence of two independent variables with the business group's prior performance on the CEO succession events.The business group's prior performance is measured as average firm ROA over the last two years prior to the year of the succession (See Khanna and Rivkin, 2001).
Control Variables.This research controls the possible influences-such as business group is belonging to manufacturing or not, business group's size, age, the revenue percentage of core company, and its year effect-on the relationships among the overlapping investment, the use of family ties, and the CEO succession issue in Taiwan's family business groups.
C1: The Year Effects.The use of -period effects‖ or -year effects‖ is a possible method to diminish the contemporaneous correlation that may occur in the panel data; that is, when the residuals of units observed in each time period are correlated (e.g., See Khanna and Rivkin, 2001).In this current study, the year effect is a dummy-coded variable to indicate the five years, i.e., y1988, y1990, y1992, y1994, y1996, y1998, y1999, y2000, y2001, and y2002 in the time frame.
C2: Industrial Type of Business Group's Core Company.The core area of the business group plays a key role in determining its resource characteristics and has a significant effect on its strategic decision (e.g., Markides and Williamson, 1994).This variable is dummy-coded to indicate whether the core area of the family business group belongs to the manufacturing sector or not.
C3 & C4: Business Group Size and Age.This study controls business group's size and age that can potential impact succession decision (e.g., Lee, Lim, and Lim, 2003).The group's size in the timeframe is defined as the total assets of the group taken over the years 1988 to 2002, respectively, and is employed by the natural logarithm transformation.Meanwhile, the group's age is computed by subtracting the date of the group's founding from the years 1988 to 2002, respectively.
C5: Core Company's Revenue Percentage in Business Group.As discussed above, the business group's core area plays a key role in determining its resource characteristics, and then its relatedness between its strategic assets, or its resource utilization.Therefore, the revenue percentage of the business group's core company will influence its resource allocation, and its decision in succession issue.This influential effect must be controlled when considering the determinants of the CEO succession in the core company.

3.3.Data Analysis and Data Structure
In this study, there are 35 family business groups that are on the top-100 business group list at least 8 years in Taiwan over the period.Since the current research utilizes panel data over a 14-year period, and the dependent variable is a count variable, the research test the proposed hypothesis using Poisson GEE population-averaged regression model implemented using STATA software.The use of panel data can overcome the limitations of the cross-sectional data, particularly the broadened time frame (Beck and Katz, 1995).Furthermore, the use of panel data over 10 years is required for observing the succession issue in family enterprise.
As shown in Table 1, among the 35 family business groups, there are 25 manufacturing ones.Those family business groups engaged in the financial service sector are the largest among the sample in this research.And the family business groups engaged in the food sector have the highest number of affiliates during the time period.In dealing with the succession events, during the time period, the average CEO succession events over the 35 business groups is more than one.Furthermore, in addressing the issue of using family ties in affiliates, the table indicates that the average percentage of the affiliate's CEOs possessed family ties is -0.30‖.Additionally, the average degree of overlapping investment between the owner-managers is -5.98‖ over the time period.

-----------------------------------------INSERT TABLE 2 ABOUT HERE ------------------------------------------Three
Poisson GEE population-averaged models (M1 to M3) were used to investigate the causal effects of the control variables and the independent variables.The Wald Chi 2 values were found to be significant for all models.Therefore, both the model-of-fitness and the model settings were reasonably well satisfied in all cases.Regarding the effect of the group's degree of overlapping investment on its CEO succession, Table 3 shows that the family business group's degree of overlapping investment appears to have significantly negative influence on the group's CEO succession probability.Specifically, the higher the degree of family business group's overlapping investment between owner-managers, the lower the possibility that the family business group has the CEO succession, as proposed by hypothesis 1 (H1) argument.Our result indicates that in family business enterprises of Taiwan, when there are a greater number of common shareholders in the business group, the control power of original inner circle members is stronger, so that decreases the likelihood of succession.
Furthermore, in addressing the causal relationship by business group's use of family ties in affiliates, Table 3 reveals that the higher proportion of family members as the CEO positions within the family business group, the lower possibility that the family business group has the CEO succession happened, as proposed by hypothesis 2 (H2).Family business enterprises are characterized by utilizing family ties in business operation to keep the family control over the enterprises.In addition, family business enterprises are also prominent by their distinctive family goal and family resources.The negative influence by family business group's use of family ties on the succession indicates that when there are more family members occupy the key managerial positions within the family business groups, the family can keep and entrench the control over the group, and then decreases the likelihood of the succession.
Hypothesis 3 predicts that overlapping investment and family ties will interact with prior performance of the business group to positively affect the possibility of succession.With the inclusion of the interaction in model 3 of Table 3, the coefficients for the interaction are positive.However, the two interaction terms do not have significant influence on the family business group's CEO succession.The results do not support the proposed arguments in hypothesis 3 (H3a and H3b).The non-significant influence of the interaction term indicates that the group's prior performance does not mediate the relationship among the overlapping investment, the use of family ties in affiliates, and the CEO succession in family business group.The prior performance is an index to identify the group's external performance in the market.In addition, the overlapping investment and the use of family ties in affiliates indicate the internal control power within the family business groups.The non-significance of the interaction effect indicates that, in the context of Taiwan, the succession decision in family business group is a possible reaction to its internal control status, and not a reaction to its external performance.
Regarding the control variables, the core company's revenue percentage in family business group is significantly influential on the business group's CEO succession possibility.The negative influence of the core company's revenue percentage in family business group indicates the relative importance of revenue contribution from the core company may lower the possibility of CEO succession in the core company.

5.Conclusion and Discussion
The definition of family business contains three major constructs: family, ownership, and control.Ward (1987) defines the family business as -a business that will be passed on for the family's next generation to manage and control.‖The interaction of family involvement and the business operation may bring the intra-group dynamics between the organizational social capital and family capital, and then arise the discussion about the possible implications of family value (Arregle, Hitt, Sirmon, & Very, 2007; Miller & Miller, 2005).Our research has examined the possible determinants in predicting the enterprises' succession happened.According to the empirical results, the higher degree of overlapping investment between owner-managers and the group's use of family ties, the lower the possibility that the succession happened in the family business group.However, the group's previous performance do not have significant influence upon the succession happened in the Taiwan's context.This study highlights the importance of alternative control choices within the family business by investigating the relationships among performance, overlapping investment, use of particularistic ties and succession.It is also provide a good start point to discuss the succession issue under the Chinese institutional context.Although the previous studies indicates the importance of succession issue in family business (e.g., Gersick et al., 1997;Lee, Lim, & Lim, 2003), how the family businesses deal with the succession issue still lacks of empirical evidence.Furthermore, how can the family businesses transit from family to professional managed enterprises, from generation to generation, is still a puzzle.In the future, it needs more discussion to argue that there has any difference in the strategy and performance implication by comparing between Taiwan and other East Asia region's family businesses.
Taiwan is a distinctive economy entity within the Great China societies.Current research in Taiwan's family business has found the importance of alternative family control mechanisms in the embedded institutional environments.Furthermore, they also indicate the possible future challenges for family businesses running business in this area.With the globalization, business environment is now much more competitive and highly dynamic.How the family operation can add value to the enterprises is a very challenging mission.More systemic researches in Taiwan's family business is not only meaningful in its own right, it also provides a good comparing start-point for researchers interested in understanding the family value issue in Great China and Asia.Note: *All data calculated at year-end 1988 to 2002.We calculate the data by averaging the within-group data over 14-years, and then calculate the average data within a specific industrial sector.Industrial sector definitions of business groups based upon nature of core company business.If the core company of the business group is both engaged in manufacturing and service sectors, this business group will be classified to other category.

Table 1 .
Summary of Business Group Data Analyzed by Industrial Sector*

Table 2 .
Spearman Rank Correlation Table