1
FACULTY OF ECONOMICS AND
BUSINESS
Master Thesis IB&M 2019/20
The effect of family firm heterogeneity on the preferred level of control in
international acquisitions of large family firms.
Pascal Klingenmeier
14/06/2020Word count
(without tables and appendices) 14.824
Student number: S4051319
Mail: p.klingenmeier@student.rug.nl
pascal.klingenmeier@web.de
Thesis supervisor: Dr. A. Kuiken
The effect of family firm heterogeneity on the preferred level of control in
international acquisitions of large family firms.
Abstract:
Family firm internationalization is a topic of major interest. Thereby, it is essential to differentiate between key concepts of the family firm and internationalization. This thesis uses a sample of acquisitions of some of the largest and most successful family firms worldwide to clarify the role that family firm heterogeneity has on the level of control chosen in international acquisitions of large family firms. Based on 336 acquisitions from the years 2016 to 2019, Tobit regression analysis is used to explain variance in the level of control of their international acquisitions. The thesis uses the resource-based view and resource-dependence theory to argue that governance structures in the hand of the family are drivers of high levels of control. Furthermore, this thesis investigates on interactions when certain governance structures are in the hand of the same family simultaneously. Overall, findings do not support the hypotheses and a family CEO is even negatively associated with the level of control choice. Still, the thesis allows some exciting insights and opens the way for prospective further research.
Keywords: family firm internationalization, family firm heterogeneity, family firm acquisition,
3
Acknowledgments
At this point, I would like to express my gratitude to the people who supported me during the process of writing this thesis. First, my supervisor, Dr. Andrea Kuiken who always had an open ear for me and improved my work with her valuable feedback and positive attitude. Furthermore, I want to thank my family for the continuous support over the process of writing this thesis and my entire studies.
I. Content
II. List of Tables 5
III. List of Figures 5
IV. List of Abbreviations 6
1. INTRODUCTION 7
2. THEORETICAL FRAMEWORK 10
2.1 Family business internationalization 10
2.1.1 Definition and heterogeneity 10
2.1.2 Factors influencing family business internationalization 11
2.1.3 Internationalization of family and non-family firms 12
2.1.4 Family firm heterogeneity and internationalization 13
2.1.5 The role of company type and type of internationalization 14
2.1.6 Family businesses and M&As 15
2.2 Resource-based view and resource-dependence theory 18
2.3 Hypothesis development 19
2.3.1 Family ownership and choice of level of control 19
2.3.2 Family CEO and choice of level of control 22
2.3.3 Family board and choice of level of control 24
2.3.4 Interaction effects 26
3. METHODOLOGY 29
3.1 Data sources and sampling 29
3.2 Dependent Variable 30
3.3 Independent Variables 30
3.4 Control Variables 30
3.5 Method description and analysis 32
4. FINDINGS 35
4.1 Descriptive statistics 35
4.2 Results 36
4.3 Robustness Check 37
5. DISCUSSION & CONCLUSION 39
5.1 Discussion 39
5.2 Future research 41
5.3 Limitations 42
5.4 Contributions to theory 43
5
5.6 Concluding remarks 43
6. REFERENCES 45
7. APPENDICES 54
II. List of Tables
Table 1: Pairwise correlation matrix 33
Table 2: Year and industry sector of family business acquisitions 35
Table 3: Descriptive Statistics 35
Table 4: Results of the random-effects Tobit regression 37
Table 5: Descriptive statistics of the dichotomous DV 38
Table 6: Variance inflation factors 55
Table 7: Acquisitions per home country 57
Table 8: Tobit regression results following the literature approach 58
Table 9: Random-effects Logistic regression output 59
III. List of Figures
Figure 1: Conceptual framework 28
Figure 2: Histogram of the level of control in acquisition targets 34
Figure 3: Share of large family companies across regions 54
Figure 4: Number and value of M&As worldwide 54
IV. List of Abbreviations
CEO Chief executive officer
DV Dependent variable
FDI Foreign direct investment
GFBI Global Family Business Index
LR Likelihood-ratio
M&A Merger and acquisitions
MNE Multinational enterprise
RBV Resource-based view
RDT Resource-dependence theory
RoA Return on assets
R&D Research and development
SEW Socio-emotional wealth
TMT Telecommunication, media and technology
VIF Variance-Inflation-factor
7
1. INTRODUCTION
Although family businesses have been overlooked for a long time by research, they play an important role in the global economy and account for up to 90% of all businesses (Gedajlovic et al., 2012; Heck & Trent, 1999). While they differ considerably from other types of companies (Duran et al., 2016), research of how that distinctiveness would manifest in their internationalization is comparably young (Gallo & Sveen, 1991; Swinth & Vinton, 1993). Meanwhile, the literature of family firm internationalization is quite diverse and, in some topics, somewhat inconsistent. So, in some studies, it was found that family firms are more hesitant towards internationalization (Fernández & Nieto, 2006; Graves & Thomas, 2004). Then again, other results suggest being a family firm is advantageous for internationalization (Carr & Bateman, 2009; Zahra, 2003). Last, Mitter et al. (2014) and Sciascia et al. (2010) find evidence for a curvilinear relationship suggesting that companies with mid-level family ownership are most internationalized. These findings demonstrate that our understanding of the internationalizing family business is still limited. Recent progress in acknowledging the prevailing heterogeneity among family businesses helps to shed further light in the field of family businesses and their internationalization (Chua et al., 2012; Melin & Nordqvist, 2007). Accordingly, family firms differ in terms of family ownership and involvement but also in their values, objectives and resources (Chua et al., 2012). Furthermore, as not all companies do have the same means to internationalize available it is also necessary to differentiate between company types. For example, one set of studies focuses on small and medium sized enterprises (SMEs) and their preferred use of export when penetrating international markets (e.g. Arregle et al., 2012; Fernández & Nieto, 2006; Graves & Thomas, 2004, 2008; Lin, 2012). Yet, exports compared to equity-involving entry modes differ significantly in terms of risk, commitment and control. Since such factors are pivotal in family business decision-making (Fernandez & Nieto, 2014), it is also important to distinguish between different modes of internationalization.
number and volume of mergers and acquisitions (M&As) in the past years (Appendix 2), it comes as no surprise that also these family businesses use acquisitions as a mean to internationalize (Family Capital, 2015). Exemplary evidence can be drawn from the $1.5 billion takeover of the Norwegian aquaculture company EWOS by the family firm Cargill (Bray, 2015). Cargill expected to add EWOS’s unique capabilities and sustainable business practices to its resource portfolio. Furthermore, Cargill pursued long-term growth objectives expecting a strong rise in demand for farmed fish in the coming decades (Cargill, 2015). This exemplifies already some of the distinctive features of family firm acquisitions that will be discussed in more detail throughout the thesis. As acquisitions offer a unique interplay in terms of commitment and risk on the one side and control and potential returns on the other side (Fernandez & Nieto, 2014; Hussinger & Issah, 2019), it comes as no surprise that acquisitions are evaluated differently by family firms. In agreement with their greater risk aversion (Naldi et al., 2007), support for a reluctance towards M&As is quite established in the literature (Bauguess & Stegemoller, 2008; Caprio et al., 2011; Geppert et al., 2013; Miller et al., 2010; Sraer & Thesmar, 2007). Yet, acquisitions by family firms do occur. Additionally, strategic considerations of acquisitions do not end after the choice to acquire or not. Rather, in case of acquisitions, the choice of the first stake influences how factors like risk and control ultimately play out. Entering by lower stakes could offer the option for inside screening and leave the firm with an exit option (Chari & Chang, 2007; Choi, 1991; Folta, 1998). Or, family firms could prefer to acquire higher percentages of ownership to limit or rule out the influence of other owners. Research on stake differences of acquisitions is in general rather scarce but has been completely overlooked for family businesses. Moreover, since variables of family firm heterogeneity have been found influential for family business internationalization (e.g. Arregle et al., 2012; Calabrò et al., 2016; Kraus et al., 2015), I expect them to also have an impact on the level of control chosen in an acquisition abroad. In this thesis, family firm heterogeneity is captured by governance structures such as ownership, the CEO and the board. Therefore, I postulate the following research question in this thesis:
How does family firm heterogeneity affect the preferred level of control in international acquisitions of large family firms?
9
constraints of the family business (RBV) but also mind its dependence on the environment (RDT). Besides, both theories offer motivations for acquisitions that can be combined meaningfully in the context of family firm internationalization. By basing my analysis on a dataset of the 500 largest family businesses from the ‘Global Family Business Index’ (GFBI) (EY & University of St. Gallen, 2019), I ensure that the organizations in the sample have the necessary resources to conduct international acquisitions. The proposed impact of family governance structures and their respective interactions are tested using a random-effects Tobit model.
This thesis contributes to two streams of literature. First, by expanding the literature of family business acquisitions to equity choices, it adds a situational measure for family preferences. Also, by incorporating interaction terms between family heterogeneity variables it contributes to advance the field. Then, this thesis contributes to the literature on equity choices in international acquisitions by showing the effect different governance structures can have. Most of the results are insignificant. However, the presence of a family CEO seems to lower the level of control in an international acquisition.
2. THEORETICAL FRAMEWORK
In this section, I review the literature on family firm internationalization. Starting from a broader point of view, the topic will be narrowed down towards what we know about family firms and acquisitions while also acknowledging family firm heterogeneity. Then, I draw from arguments of the RBV of the firm and RDT to develop my hypotheses.
2.1 Family business internationalization
2.1.1 Definition and heterogeneity
The internationalization of family firms has been a topic of growing interest for the past three decades (Gallo & Sveen, 1991; Swinth & Vinton, 1993). Despite the undisputed progress we have made in understanding this specific type of company, there is still a lot we do not know. For instance, until today no definition of family firms has found general acceptance (Kontinen & Ojala, 2010; Sharma, 2004). Yet, it is acknowledged that family firms differ from other types of companies. Sharma (2004: 9) states that “the intertwinement and reciprocal relationships between
the family and business systems is being recognized as the key feature distinguishing this field of study from others “. From this, three features can be deduced that are unique to family businesses.
First, the owning family has a substantial level of control over the firm, e.g. through voting rights. Then, the wealth of the family is usually concentrated in the family business. Lastly, non-economic goals play a more pronounced role compared to other types of companies (Duran et al., 2016). The last feature has even inspired a new literature stream focusing on family firms' so-called
socio-emotional wealth (SEW) (Gomez-Mejia et al., 2007). Additionally, family business researchers
11
management either as an executive or through board membership. By choosing this definition, I align with a considerable part of the literature (Miller et al., 2007).
2.1.2 Factors influencing family business internationalization
Before investigating further on family business internationalization, it is worthy to recall some of the factors that make family business internationalization distinct from the internationalization of other companies. This is because the only existence of such factors warrants research that focuses on family business internationalization separately (Rau, 2014).
Given that internationalization is a risky activity (see e.g. Reeb et al., 1998; Zaheer, 1995; Zaheer & Kostova, 1999), also the unique risk perceptions of family firms are influencing family business internationalization. Most studies describe family firms to be more risk-averse than their non-family peers (Hiebl, 2012). One reason behind this is that families are often only invested in their company and therefore cannot protect their wealth using diversification (Bianco et al., 2013). However, when the business itself is threatened, family firms often pursue rather daring strategies to ensure their survival (Hiebl, 2012). Apart from financial concerns, family businesses also consider threats to their SEW (Berrone et al., 2012). SEW captures all non-financial benefits and
“affective needs” the owning family derives from its business (Gómez-Mejía et al., 2007:106). It
follows that these unique risk perceptions are also considered when taking the company abroad. Since the business satisfies both the financial and affective needs of the family, their desire to keep control within the family is also observable when family businesses internationalize (Claver et al., 2007; Kraus et al., 2016; Okoroafo, 1999). Partly, this is because family firms want to preserve the company for future generations (Berrone et al., 2012; Gallo & Pont, 1996; Kontinen & Ojala, 2010). Consequently, they avoid raising external involvement in their business when internationalizing and have the ability to pursue much longer time horizons in their international ventures.
have disadvantages. This is because they often place family members in top executive positions although they might have lower capabilities than professionally trained managers (Gallo & Pont, 1996; Lin & Hu, 2007). Moreover, lower career opportunities in family firms might also daunt ambitious and capable managers from joining family firms in the first place what further constrains the pool of capable candidates (Fernández & Nieto, 2006).
On the other hand, family firms do have some beneficial resources for internationalization. Factors like patient financial capital, superior human and social capital (Sirmon & Hitt, 2003) and reduced potential for agency conflicts (Schulze et al., 2001) have been argued to be assets in internationalization. On the benefits, I fall back on in the hypothesis development section. Altogether, it becomes clear that there are unique family firm attributes that also influence their internationalization and therefore justify research on family firm internationalization.
2.1.3 Internationalization of family and non-family firms
Despite the early acknowledgments of considerable differences among family firms (Astrachan & Shanker, 2003; Melin & Nordqvist, 2007), the field of family business internationalization started off around the question of how the internationalization of family firms differs from the internationalization of non-family firms. Many studies draw from a randomly selected group of companies, sometimes based in a specific country (e.g. Fernández & Nieto, 2006; Graves & Thomas, 2006) or industry (Fernández & Nieto, 2005), and distinguish between family and non-family firms using a binary variable. The criteria deployed for the categorization of non-family businesses turns positive when there is family involvement in management (Fernández & Nieto, 2005, 2006), ownership (Pinho, 2007) or both (Abdellatif et al., 2010; Gomez-Mejia et al., 2010; Graves & Thomas, 2004, 2006).
13
2.1.4 Family firm heterogeneity and internationalization
In a different approach, researchers choose from samples of family firms to incorporate their differences rather than simply dividing companies into family and non-family firms. Governance structures shape a firm’s ability to handle complexity arising from internationalization (Sanders & Carpenter, 1998) so that it comes as no surprise that differences in ownership and management also affect family firm internationalization (Mitter et al., 2014). As heterogeneity in governance structures is easily observable from the outside and the functions of these roles have different implications, governance variables have been used widely to account for family firm heterogeneity. The owners of a firm can wield influence e.g. via voting rights. Their influence in matters of internationalization has been captured through measures such as the family as majority owner (Bhaumik et al., 2010), the relative ownership of the family (Sciascia et al., 2010; Zahra, 2003, 2005) or of other owners (Arregle et al., 2012; Kraus et al., 2016). Besides, the presence of a family CEO (Claver et al., 2009; Kraus et al., 2015; Zahra, 2003, 2005) and of family board members (Arregle et al., 2012; Kraus et al., 2015) are prominent measures as individuals in these roles are more involved in day-to-day operations.
Astonishingly, also the inclusion of more family variables has failed to produce clear-cut results of whether family ownership and influence in management will contribute to internationalization. For starters, studies investigating on family ownership are inconclusive. Zahra (2003, 2005) uses stewardship and agency theory respectively to argue for the benefits of being a family-owned firm. In contrast, Arregle et al. (2012) stretch the importance of external ownership from an RDT perspective finding support for their arguments as well. Also, Bhaumik et al. (2010) conclude that family majority-ownership is negatively related to internationalization. Then again, two studies find evidence for a curvilinear relationship between family ownership and internationalization in which companies with mid-level family ownership are the most internationalized ones (Mitter et al., 2014; Sciascia et al., 2010). Both studies acknowledge the advantages and disadvantages of family firms in matters of internationalization arguing that in early stages of internationalization the family as good stewards dominate while later the lack of internal resources causes agency conflicts with newly added external partners (Mitter et al., 2014) or leads to stagnation (Sciascia et al., 2010).
internationalization scope. Similarly, Alessandri et al. (2018) find that family managers promote internationalization. In contrast, outside managers compared to family managers increase the commitment towards international markets (Claver et al., 2009). Then again, Arregle et al. (2012) find only partial support for their argument that external board members will contribute positively towards internationalization. Neither is support found for the benefits of nonfamily managers in comparison to managing family members (Mitter et al., 2014). Recently, the idea that there is more than just a single optimal way for family firms to achieve high levels of internationalization has been raised to explain such inconclusive findings (Kraus et al., 2015).
This review suggests that while family firm heterogeneity surely improves our understanding of the internationalizing family firm. Yet, the incorporation of heterogeneity is no magic bullet or panacea that can overcome all inconsistencies in the literature. This suggests that it is also important to pay closer attention to the type of company and what is meant by the generic term internationalization.
2.1.5 The role of company type and type of internationalization
Many studies in both, the comparative (e.g. Fernández & Nieto, 2005, 2006; Graves & Thomas, 2004, 2006) and in the heterogeneity set of studies (Arregle et al., 2012; Mitter et al., 2014; Sciascia et al., 2010) are using samples of SMEs. This is because many SMEs are family firms. Often, they internationalize using exports and do not have the resources and capabilities available for equity involving modes of internationalization. Therefore, many studies that aim to investigate on family firm internationalization only look into the export behavior of family SMEs. Commonly these studies rely on export or foreign sales density (e.g. Arregle et al., 2012; Fernández & Nieto, 2005; 2006; Graves & Thomas, 2004; 2006; Sciascia et al., 2010; Zahra, 2003) or export scope in terms of countries (Arregle et al., 2012; Sciascia et al., 2010; Zahra, 2003). Some researchers use more comprehensive approaches of internationalization, yet, they also incorporate exports in some way (Kraus et al., 2015; Muñoz-Bullón & Sánchez-Bueno, 2012). Since considerations in larger firms can differ and every mode of internationalization comes with different implications e.g. regarding risk and control, it can be deceptive to apply knowledge gained from above studies to larger firms and other types of internationalization.
15
65 non-family firms in terms of internationalization. In contrast to some of the findings for smaller companies, they find that family firms are more internationally configurated than the comparison group. The international orientation of these firms was even found to rise with family ownership and family involvement. While both studies compare family firms with non-family firms, they also differentiate among family firms. The former study distinguishes ownership and management, while Carr & Bateman (2009) deploy a more comprehensive classification of family governance. In terms of internationalization, large companies have more choices at hand. Particularly, equity involving modes of foreign market entry are available to them. Here, literature becomes quite diverse in terms of measures and ways to capture equity-involving internationalization. Pinho (2007) for example finds no difference in preference for equity involving market entries between family and non-family firms. Yet. he equalizes joint ventures and wholly-owned subsidiaries (WOS). Another study ranks foreign market entry modes according to their commitment to the international market from exports to WOS (Claver et al., 2009). Then, Bhaumik et al. (2010) use the relative share of assets held abroad to capture equity involving internationalization of firms and find that family involvement is obstructive towards outward FDI. However, since the differences among equity-involving types of internationalization are relevant for family firms, they should be analyzed individually (Fernandez & Nieto, 2014). Therefore, I review specifically the literature that covers M&As of family firms in the next section.
2.1.6 Family businesses and M&As
firms, entering new markets, reducing competition (Grimpe & Hussinger, 2008; Vermeule & Harry, 2001) or managerial hubris (Nguyen et al., 2012). Cassiman & Veugelers (2006) propose in their work that internal R&D and external knowledge acquisitions are complementary. In other words, a company that does forgo any acquisitions might miss out on opportunities for technological improvement and resulting differentiation from its competitors, which would make acquisitions critical for the long-term survival of the firm and help explain family firm involvement. While M&As are often analyzed simultaneously, they have very different implications for family firms and their desire to be in charge. In case of an acquisition, the family remains in control over the entity, while in a merger control must be shared with another party (Worek et al., 2018). Therefore, most family firms prefer acquisitions and mergers do remain an exception. This can also be observed in the literature so that a study focusing on Japanese firms’ merger activity is quite a rare exception. As expected, the study in question concludes that family firms are more reluctant towards mergers than other types of companies (Shim & Okamuro, 2011). Besides the classic motives for M&As, family firms also have unique interests when it comes to acquisitions. They stem from earlier introduced unique characteristics of the family firm, namely the preservation of high levels of control, the wealth concentration of family members in a single asset and their non-economic goals (Duran et al., 2016; Worek et al., 2018). Based on an extensive qualitative evaluation of press releases, Worek et al. (2018: 260) elaborate on the different goals of family firms. Compared to non-family firms, they “disclose more goals related to stakeholders and
market competitiveness (…) and fewer financial and innovation goals”. Particularly, they are more
risk-averse and have a long-term horizon in their acquisition behavior while also taking stakeholder demands into consideration (Geppert et al., 2013; Hussinger & Issah, 2019). Another distinguishing factor in the M&A behavior of family firms is their preference for cash over equity-involving acquisition deals caused by an aversion to relinquish control with investors (Caprio et al., 2011). In a comparative case study from the brewery industry, Geppert et al. (2013) observe that the acquisitions of family firms can be considered less risky and that the raising of funds is done without relying on capital markets too much. In the long run, the authors also note that family breweries created superior synergies from their acquisitions.
17
S&P 500 companies (Bauguess & Stegemoller, 2008), studies find that family firms execute fewer M&As than nonfamily companies. Additionally, some researchers also employ heterogeneity measures by differentiating family ownership and family management. Miller et al. (2010), for example, find that higher degrees of family ownership in Fortune 1000 companies are negatively related to the number and value of acquisition deals. This was found to be more pronounced when the family has relatively lower levels of ownership due to a perceived threat of losing control (Caprio et al., 2011). Family involvement in management is also found to contribute towards a reluctance against M&As. This has been observed for CEOs (Caprio et al., 2011; Sraer & Thesmar, 2007) and family board directors (Requejo et al., 2018). Naturally, the type of ownership is not the only influencing factor on the M&A behavior of acquiring firms. For example, the institutional context can play a role as well (Geppert et al., 2013; Requejo et al., 2018). While it has become a common understanding in the literature that family firms are more reluctant to acquire other companies, most studies base their analysis on rather limited acquisition measures. They either count the acquisitions made by the company per year (Miller et al., 2010; Sraer & Thesmar, 2007) or whether in a certain year an acquisition was performed (Bauguess & Stegemoller, 2008; Caprio et al., 2011; Hussinger & Issah, 2019; Requejo et al., 2018). By relying on such measures, these studies aggregate the variance in the data into a dichotomous or counting variable and thereby reduce their informative value. Additionally, some studies employ criteria on acquisition deals that exclude minority deals completely from the analysis (Bauguess & Stegemoller, 2008; Hussinger & Issah, 2019) or rely heavily on domestic deals (Miller et al., 2010).
as it determines the extent of risk and control (Lahiri et al., 2014) – factors particularly relevant to the internationalization of family firms (Fernandez & Nieto, 2014). Entering by lower levels of control could offer family firms the option for inside screening and leave them with a cheap exit option (Chari & Chang, 2007; Choi, 1991; Folta, 1998). Alternatively, it can be argued that issues of control are prioritized in the decision and that consequently, family firms prefer to be majority owners in international acquisitions. Since family firm heterogeneity is found to be impactful in matters of internationalization, it is worthwhile to investigate whether it will also influence equity and control choices of acquisitions.
2.2 Resource-based view and resource-dependence theory
In this thesis, I apply a combination of arguments from the resource-based view and resource
dependency theory. Both theories have been used in the family business literature before, but not
in the context of level of control choices of international acquisitions (Arregle et al., 2012; Fernandez & Nieto, 2014; Habbershon & Williams, 1999; Naldi & Nordqvist, 2008; Sirmon & Hitt, 2003).
19
While the RBV focuses on resources within companies, resource-dependence theory considers resources outside the firms’ direct reach. The foundation of this reasoning is that a firm is not in control of all resources it needs for its survival. Rather, firms depend on other actors to provide them with vital resources. The availability of these resources, for instance by suppliers, lies outside a firm’s control and might fluctuate with changing circumstances. The threat of dependence of other organizations is managed in a risk-mitigating way by the firm and its management. The organization has several tools available in order to do so. Among others in particular M&As offer the option to impose power over organizations that are critical to a firm’s survival (Hillman et al., 2009; Pfeffer & Salancik, 1978). RDT offers three reasons why companies undertake M&As: The first option is that the firm absorbs a competitor and by doing so reduces the uncertainty in its market. Second, the organization absorbs an upstream or downstream partner to decrease uncertainty in its supply chain or its revenue streams. Last, the organization acquires an unrelated target so that its current dependencies are relatively less important (Hillman et al., 2009; Pfeffer, 1976). In the literature, the arguments are postulated for mergers. Yet, as the goal is taking control and managing previous external dependencies internally, they also hold for the acquisition case. For family firms who are reluctant towards sharing control with external parties, it can be argued from an RDT perspective that a merger is an inadequate measure of internalizing firm interdependencies. This is because in such a scenario ownership and control would be shared with another organization and dilute family control in the new entity. In the context of family firms, RDT focuses mainly on the board and how external involvement can facilitate the access to external resources (Arregle et al., 2012; Voordeckers et al., 2007).Additionally, external dependencies and the distinctive risk perception of family companies can be combined meaningfully.
2.3 Hypothesis development
In this section, I use arguments of RBV and RDT successively to postulate hypotheses of how family heterogeneity might affect the level of control in international acquisitions.
2.3.1 Family ownership and choice of level of control
Following the RBV, some of the familiness attributes apply specifically to family ownership and are positively associated with internationalization. These arguments can also be implemented to the level of control choice. RDT, on the other hand, sees acquisitions as a way to mitigate external risks so that according to both theories, family ownership is expected to raise the level of control chosen in international acquisitions.
family firms are usually argued to face restraints raising the necessary financial resources for internationalization (see e.g. Bauguess & Stegemoller, 2008; Fernández & Nieto, 2006; Sirmon & Hitt, 2003), it is necessary to dispel this concern from the outset. While financial constraints might hold true for many family firms, in this thesis I investigate on large and successful family firms. Therefore, it can be excepted that these companies have overcome financial restraints by entering the capital market themselves and by issuing their own bonds. Exemplarily support is provided by the German car manufacturer Volkswagen who is also part of the sample (Volkswagen AG, 2020). On the contrary, family-owned companies possess a superior financial structure called patient capital that is likely to influence international acquisitions. Patient capital stems from the family’s wish to preserve the company for future generations, resulting in a long-term investment horizon. Accordingly, family firms can pursue a long-term strategy in their international acquisitions without an underlying pressure of having good financial results in the regular financial reports (Dreux, 1990). Lehrer & Celo (2016: 730) describe patient capital as ownership patterns that protect “the firm from short-term performance pressures”. This implies that family firms can pursue expedient acquisitions even if the success is only expected in the mid or long-term. Therefore, as was observed by Kang (2000) family firms are particularly well suited to pursue long-term growth strategies (as cited by Issah, 2019). International acquisitions will either be executed to appropriate external resources from the target (Ahuja & Katila, 2001), the creation of synergy effects (Cassiman & Veugelers, 2006) or the use of external resources to facilitate access to a foreign market (Grimpe & Hussinger, 2008). In all those scenarios, family firms will pursue a long-term strategic goal. Yet, as the status quo is likely to change over time and family firms will aim to fulfill their long-term strategies, they will try to buy out other owners as their behavior is too uncertain in the long run. Additionally, in case of a combination of resources from the parent and the target company (Cassiman & Veugelers, 2006), it might become necessary to push both companies strategically in the same direction, for example with regard to R&D. Such adjustments can better be implemented if the acquirer has higher levels of ownership. Therefore, I argue that family-owned firms will try to safeguard their long-term strategy by entering with high levels of control in the acquisition target to limit third party influence on the acquired resources from the outset. By doing so, they can exert their influence to alter the strategy of the target firm the way it contributes most to the whole business. Thus, family ownership is associated with higher levels of control in acquisitions.
21
dependencies of whom one is the integration of external resources via acquisition (Pfeffer & Salancik, 1978). Since family firms are risk-averse (Claver et al., 2007; Hiebl, 2012) they will perceive the threat posed by external resource-dependence very pronouncedly. Additionally, the international context involving a new environment, different customers and competitors and a new institutional context (Fernandez & Nieto, 2014) are likely to increase the perceived risk even further. Therefore, the integration is conducted for the purpose of reducing the risk posed to the family firm from the uncertainty about the environment’s behavior and consequences on resources necessary to the survival of the firm. Following this logic, the suggestion that family firms will acquire higher levels of control is intuitive for two reasons. First, the involvement of other owners in the target company creates uncertainty about their behavior in strategic decisions within the firm and is therefore undesirable from a family perspective. Second, the integration of other companies attributes power to the acquiring company which strengthens its position in the environment. This creates the advantageous situation that the family business can exercise power over other players instead of other firms misusing their power on them (Hillman et al., 2009; Pfeffer & Salancik, 1978). Moreover, the recognition by the environment as a powerful organization rests on the assumption that the family firm is the most powerful entity within the acquired company. Therefore, it follows that the family firms can only leverage their power over the environment when they have higher levels of control so that family-owned firms will prefer to acquire high levels of corporate control. Additional evidence for such reasoning can be drawn from the Naldi & Nordqvist (2008) study. Their findings reveal that external ownership in family firms is only beneficial for internationalization scope but not for its scale. As an explanation, they offer that external owners might prefer the entry in multiple markets simultaneously over a committed market entry in just one country. Hence, this supports the idea that family owners are more inclined to an acquisition with higher levels of control.
By combining the arguments of RBV and RDT, namely the appropriation of external resources and the need to safeguard them and the family business itself, I argue that family-owned firms prefer to acquire higher levels of control. The owners' influence on decision-making (Goodstein & Boeker, 1991) will be higher with higher levels of family ownership. Therefore, I postulate the following first hypothesis:
2.3.2 Family CEO and choice of level of control
As in all companies also in family firms managers at the top level have a strong influence on the strategic direction of the firm (Boeker, 1997). From an RBV perspective having a family member appointed as CEO will be impactful due to his or her human and social capital resources (Sirmon & Hitt, 2003) as they can influence firm outcomes (Hitt et al., 2001). Similarly, RDT suggests that a family CEO is more involved in reducing risks associated with resource-intensive takeovers so that following both theories, I expect the presence of a family CEO to positively impact the level of control choice.
23
et al., 2014). Thus, I argue that when a family CEO reaches the conclusion that an acquisition is the right strategy, he will go through with it and chose to acquire high levels of control in order to safeguard and make best use of the target’s resources for the company.
The second familiness resource that can be leveraged in an international acquisition is the superior social capital of managing family members. Social capital is a resource that is important in building and maintaining relationships. Family members can use their social capital when doing business and dealing with stakeholders (Sirmon & Hitt, 2003). Greater social capital of family members seems plausible as members of the founding family are likely to be perceived as more authentic and trustworthy to external parties. Exemplarily, Lounsbury & Glynn (2001) argue for the importance of relational capital (here: storytelling capacity) when securing external resources for a firm. I argue that also in case of an international acquisition, a family CEO will make use of his or her social capital and reach out to the executives of the target firm to assure mutual understanding about the process of the acquisition, the strategic role of the target in the future, etc. This can entail a reduction of uncertainty about the acquisition target and the inherent risk of the acquisition. The reduced uncertainty due to family resources manifested in the CEO lowers concerns about higher levels of corporate control in the target and paves the way for a committed takeover.
The last resource I address is not specifically human or social capital but has also been brought forward for family firms. Namely, it is their ability for fast and uncomplicated decision-making (Dyer, 1988; Tagiuri & Davis, 1996). At first glance, this seems contradictory to the argument that family CEOs commit more effort to their decision-making process. Yet, when the decision has been taken, a family CEO will have more freedom to invest compared to a non-family CEO due to his authority as an owner-manager. This ability is particularly important when entering insecure market environments abroad. Therefore, a family CEO is less restricted to attribute resources to the acquisition and can commit to it once the belief to do the right thing has emerged (Carney, 2005).
RDT, on the other hand, sees the basic role of the manager as a symbolic one and that outcomes are independent of the individual manager. Yet, Pfeffer & Salancik (1978) pay tribute to the manager's role in resolving organizational constraints and in guiding through the social context. The latter contains also to manage demands posed from the environment to the organization and to manage inter-organizational interdependencies via social coordination. They also make the point that social interaction is common upfront of large contracts with the goal of “stabilizing the
transactions of organizations” (p.144). I argue that due to the increased risk aversion of the family
of the family business in a risk-mitigating way. Thereby, it is likely that he will reach out to the executives and owners of the target company to ensure that the planned process of integration and the perceived goals are well understood and accepted in the target firm. This idea is supported by Worek et al. (2018) who find that family firms often pursue acquisition goals relatable to stability and good reputation of the family firm. The effects of socializing are likely to reduce the risk associated with the acquisition and therefore pave the way for an acquisition with high levels of control.
A second RDT argument follows the idea of Naldi & Nordqvist (2008). They propose that external CEOs can be an impetus for international expansion. A family CEO on the other hand is likely to have similar sensibilities as the owning family and therefore is reluctant when it comes to relinquishing control with other parties (e.g. Duran et al., 2016). This argument was put forward for the family-owned company but is also applicable to their acquisition targets. From these points, it can be deduced that while nonfamily CEOs are likely to pursue an internationalization strategy that allows for multiple destinations simultaneously, family CEOs commit to fewer destinations with relatively more resources. Due to their interest in limiting external influence in the target, they are likely to eliminate other parties' ownership by acquiring the highest possible level of control in their targets.
Due to the high impact, a CEO has on the strategic orientation of the company (Boeker, 1997), I argue that there will be a difference in the choice of the level of control depending on whether there is a family CEO or not. Based on the arguments of RBV and RDT, I propose the following second hypothesis:
H2: Having a family CEO is positively related to higher levels of control in the acquisition target.
2.3.3 Family board and choice of level of control
By relying on arguments of RBV and RDT, I expect that a higher share of family board members is associated with higher levels of control in an international acquisition due to the specific resources of family board members and their risk perceptions as family representatives.
25
firm with deep expert knowledge about the firm and therefore strengthen its decision-making processes (Fama & Jensen, 1983). This effect is likely to be reinforced when the board member is also a family member due to his or her specific resource bundle of human capital like high firm-specific tacit knowledge (Sirmon & Hitt, 2003). External board members can also bring a unique set of resources to the firm that may not be available internally. For instance, this could be expertise and specific knowledge about capital markets or technologies (Fama & Jensen, 1983; Kakabadse et al., 2001). Most studies argue that external board members are beneficial to the internationalization process of family firms (e.g. Arregle et al., 2012) - a view that that is also empirically supported (Calabrò & Mussolino, 2013; Naldi & Nordqvist, 2008). However, these findings are only partially applicable to the expected level of control in single acquisition cases. I argue that the human and social resources associated with family CEOs will also come into play with family board members: A family board member will have an above-average commitment to the family business and its important strategic decisions such as international acquisitions. Coherently, family board directors are well-suited to assess the fit between acquirer and target due to their tacit knowledge about the company (Sirmon & Hitt, 2003). Therefore, when reaching the conclusion to acquire, family board directors will prefer high levels of control and accordingly give advice to the management. More importantly, family board directors will also use their social capital in relation-building (Sirmon & Hitt, 2003) to reduce uncertainty about the target. Therefore, family board members, once favorable of the acquisition, will prefer acquiring higher levels of corporate control.
members on the board increases the number of people involved in smoothing the process of the acquisition and therefore reduces the risk associated with it.
Alternatively, RDT allows one to argue that boards consisting of more family members lack or are endowed with only limited resources necessary for internationalization (Arregle et al., 2012; Naldi & Nordqvist, 2008; Pfeffer & Salancik, 1978). Hence, these companies are disadvantaged and possibly overwhelmed by the international context and arising uncertainty. For those firms, such an environment is posing a real threat to the organization. The result will be an incentive for the board to pursue a risk-mitigating strategy that internalizes priorly external threats using the vehicle of an acquisition. Such an effort to push back the environment requires high levels of control in the acquired company to limit the access of third parties to the newly acquired resources. This is consistent with the observation that family firms pursue risky strategies when they are under threat (Hiebl, 2012).
Last, I argue that external board directors who are experts in matters of internationalization might be more inclined towards other strategies of internationalization than a resource-intensive acquisition. They might prefer lower acquisition stakes as they can be sufficient for the exertion of influence in the target firm and fulfill the strategic purpose but involve fewer financial resources that can then be leveraged differently. A similar notion was raised by Naldi & Nordqvist (2008) who find that external directors are only beneficial for the scope but not to the scale of family business internationalization.
Because family members can intervene with decisions of management due to their control rights (Carney, 2005), family board members will make use of this influence to raise stakes in the acquisition for the above-stated reasons of RBV and RDT. When the share of family members on the board is higher, their voice gets more pronounced so that management will perceive a stronger pressure to satisfy their demands. Therefore, it seems reasonable to postulate the following third hypothesis:
H3: Having a higher share of family board members is positively related to higher levels of control in acquisition targets.
2.3.4 Interaction effects
27
ownership and a family CEO as well as family ownership and family board membership interact positively.
As argued before, according to the RBV, family ownership is associated with patient capital and a long-term investment horizon so that higher shares of family ownership are expected to raise resource commitment for international acquisitions to safeguard and leverage them in the long run. Here, I argue that when a family member serves as a CEO the patient capital effect is reinforced due to the higher level of trust family owners place in a family CEO (Tagiuri & Davis, 1996). The family CEO can then go through with the acquisition relying on the support of the family as an important equity holder of the firm. The degree to which trust comes into play depends on the equity share of the family as non-family owners are unlikely to share the family-specific long-term horizon. Noteworthy is also that a non-family CEO is very unlikely to be given the same amount of trust than a family CEO.
Following RDT, the role of the CEO in international acquisitions is to smoothen the process and thereby reduce the risk associated with it (Pfeffer & Salancik, 1978). According to Carney (2005: 255), the “unification of ownership and control concentrates and incorporates organizational
authority in the person of an owner-manager or family” so that a family CEO can act firm and
binding when negotiating with representatives of potential acquisition targets. Thus, possible concerns are allayed and disagreements between the involved parties can be avoided from the outset. Yet, the authority of a family CEO is higher when the family possesses higher levels of equity in the firm. Using authority, a family CEO can then better influence deals according to his wishes as his word has more weight externally. Consequently, the risk of the acquisition can be reduced more easily, and higher levels of control can be purchased.
Bringing together both arguments I expect that higher levels of family ownership and the presence of a family CEO reinforce each other positively:
H4a: Higher levels of family ownership and a family CEO will reinforce each other and contribute to the preference of high levels of control in the acquisition target.
they reach the conclusion to acquire, they will use their influence in the board room to lobby for the long-term family interests of the acquisition and assert higher levels of control in the target. Consequently, a higher share of family board members will lead ceteris paribus to a greater influence of the family as the family board members can bring forward their point of view together. Vice versa, higher levels of family ownership equip family board members with authority (Carney, 2005) so that their say has more weight in the board room even when the family might proportionately be underrepresented on the board.
RDTs arguments are centered around the idea that threats of the environment can be internalized via acquisition (Hillman et al., 2009). Higher levels of family ownership will equip family board members with authority in the board room (Carney, 2005). Therefore, when the family and family board members perceive the threat of external risks more pronounced and want to internalize them by high levels of control, this will have a stronger impact on the company's decision.
Based on these considerations, I postulate the following last hypothesis:
H4b: The influence of the share of family ownership and family members on the board will strengthen each other and contribute to the preference of high levels of control in the acquisition target.
The theoretical framework is summarized in the conceptual model in Figure 1.
29
3. METHODOLOGY
3.1 Data sources and sampling
3.2 Dependent Variable
The dependent variable in this analysis will be the percent of ownership the acquiring family business chose in the target firm - Acquisition stake (Cuypers et al., 2015; Malhotra & Gaur, 2014). As was stated by Goodstein & Boeker (1991) higher degrees of ownership are likely to increase the influence in the target firm. On the other hand, higher levels of control are associated with a higher resource commitment and putting more resources at risk (Folta, 1998). Therefore, this dependent variable is well suited to measure context-specific preferences of family firms and extend measures beyond the sheer presence of acquisition(s) per year which is often used in the family context (see Bauguess & Stegemoller, 2008; Caprio et al., 2011; Hussinger & Issah, 2019; Requejo et al., 2018).
3.3 Independent Variables
In the analysis, I use three different variables of family heterogeneity that have been applied in the context of family firm internationalization:
Family ownership represents the share of equity and voting rights held by the family. Due to the
definition of the GFBI, the variable runs continuously between 0.32 and 1. The ownership measure has been prominently used in family business studies (Miller et al., 2010; Sciascia et al., 2010; Zahra, 2003). Studies arguing for the benefits of external ownership have used the opposite measure, share of equity held by external owners (Arregle et al., 2012; Kraus et al., 2015). Accordingly, the ownership structure has been acknowledged by the literature as a source of heterogeneity that is influential on family business internationalization.
Family CEO is a binary variable that takes the value 1 when the CEO of the company is a family
member and 0 when the CEO is a hired manager. The same variable is commonly employed in studies that incorporate the CEO in family business internationalization (Alessandri et al., 2018; Kraus et al., 2015; Naldi & Nordqvist, 2008; Zahra, 2005).
Family board is the third independent variable. It captures the involvement of family members in
the board room by using the ratio number of family board members to the number of total board members. The use of this ratio ensures comparability despite different board sizes. The share of family or respectively non-family board members have also been used in earlier studies in this field (Arregle et al., 2012; Naldi & Nordqvist, 2008) to capture the voice of the family on the board.
3.4
Control Variables
31 Firm age is the difference between the year of the acquisition and the year of incorporation of the
family firm. The year of incorporation is available from the GFBI. As the resulting firm age in extreme cases differs for several centuries, I use the natural logarithm of firm age in the analysis (Anderson & Reeb, 2003).
Firm size has also been argued to be influential for the internationalization of family firms. A
common proxy to measure firm size is the number of employees. This value was derived from the GFBI and is therefore given for the year 2019. I logarithm firm size because it is right-skewed (mean larger median) and influenced by Walmart as an extreme outlier.
Industry effects are captured via a categorical variable. The GFBI categorizes the companies in
eight sectors. For this thesis, I follow its classification. The sectors are Consumer, Advanced manufacturing and mobility, Financial services, Telecommunication, Media and Technology (TMT), Energy, Health Sciences and Wellness, Smart Infrastructure, Energy and Diversified Industrial Products. However, the final sample does not contain acquisitions from the last sector.
Performance also is influential in internationalization so that almost all studies reviewed so far
have incorporated a performance measure. Good financial results facilitate internationalization as the necessary resources are more accessible. In the following, I will use the accounting-based performance measure Return on Assets (RoA) defined as the ratio net income to total assets. As accounting figures will only become available to the management in the following year, I lag RoA by one year. I downloaded the values for net income and total assets for the years 2015 to 2018 from Orbis. RoA has already been used by several studies in the field of large family businesses and internationalization (e.g. Anderson & Reeb, 2003; Hussinger & Issah, 2019).
Organizational slack: Bourgeois (1981) identified three types of slack: absorbed, unabsorbed and
potential slack. I choose potential slack because data for its calculation, which is the ratio equity to current liabilities. is widely available. The yearly data points were also downloaded from the Orbis database. This is in line with Iyer & Miller (2008) who state that apart from unabsorbed slack especially potential slack should be considered as an acquisition initiating factor. The measure is also controlled for in the study of Hussinger & Issah (2019). Because this calculation of organizational slack is also dependent on accounting-based measures, I lag the slack variable by one year.
Subsidiaries abroad: The international experience and resulting capabilities of the firm are also
2020). Therefore, I include a structural indicator of internationalization, the number of subsidiaries abroad as a control variable (Dörrenbächer, 2000). I argue that the number of subsidiaries is meaningful as it captures the experience gained with every new subsidiary. Other measures of internationalization e.g. export or number of host countries do not capture these effects as good. The variable is also logarithmized.
Acquisition year: To control for year effects, I include the year of the acquisition as a categorical
variable (e.g. Chari & Chang, 2009).
3.5 Method description and analysis
The dependent variable (DV) of the analysis is the level of control in international acquisitions. By definition, this variable can run only between 0% (here 5%) and 100% which makes it a censored variable. For the analysis of data including a censored DV, James Tobin introduced the Tobit regression model (Tobin, 1958). Due to the nature of the DV, I will rely on this method in the analysis. Moreover, the Tobit model has priorly been used in the equity analysis of international acquisitions (Chari & Chang, 2009; Cuypers et al., 2015; Malhotra & Gaur, 2014; Ragozzino, 2009). Formula (1) and (2) illustrate models 2 and 3, respectively with (3) being the censoring conditions. 𝑦∗ = 𝛽 0 + 𝛽1( 𝐹𝑎𝑚𝑖𝑙𝑦 𝑜𝑤𝑛𝑒𝑟𝑠ℎ𝑖𝑝 𝐹𝑎𝑚𝑖𝑙𝑦 𝐶𝐸𝑂 𝐹𝑎𝑚𝑖𝑙𝑦 𝑏𝑜𝑎𝑟𝑑 ) + 𝛽2 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 + 𝜖 (1) 𝑦∗ = 𝛽 0+ 𝛽1 ( 𝐹𝑎𝑚𝑖𝑙𝑦 𝑜𝑤𝑛𝑒𝑟𝑠ℎ𝑖𝑝 𝐹𝑎𝑚𝑖𝑙𝑦 𝐶𝐸𝑂 𝐹𝑎𝑚𝑖𝑙𝑦 𝑏𝑜𝑎𝑟𝑑 ) + 𝛽2( 𝐹𝑂 × 𝐹𝑎𝑚𝑖𝑙𝑦 𝐶𝐸𝑂 𝐹𝑂 × 𝐹𝑎𝑚𝑖𝑙𝑦 𝑏𝑜𝑎𝑟𝑑) + 𝛽3 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 + 𝜖 (2) 𝑦 = { 𝑦 ∗ 𝑖𝑓 𝑦∗ ≥ 0.05 0.05 𝑖𝑓 𝑦∗ < 0.05 and 𝑦 = { 𝑦∗ 𝑖𝑓 𝑦∗≤ 1 1 𝑖𝑓 𝑦∗> 1 (3) 𝑦∗ = underlying latent variable
𝑦 = censored dependent variable
33
Table 1: Pairwise correlation matrix distribution. As the data situation is outside
the influence of the student, it remains to analyze them as well as possible. Prior to the analysis, I address multicollinearity, heteroscedasticity, and the choice of the right Tobit regression model:
The correlation matrix is displayed in Table
1. Certain variables seem to have some relation. To address the potential issue of multicollinearity the Variance-Inflation-Test (VIF) is conducted using standard OLS procedures (Ragozzino, 2009) (see Appendix 3). The mean VIF is only 1.32 and all variables are below the value of 2. As the VIF is nowhere near the critical value of 10 (Lin, 2008), multicollinearity is not an issue and I include all independent variables in model 2. However, when adding interaction terms in model 3 the VIF of the family variables rises and slightly surpasses the threshold of 10 (Appendix 3). Therefore, model 3 uses mean-centered values for these variables (Shieh, 2011). As the study period runs from 2016 to 2019,
but not all companies performed
acquisitions on a yearly basis, the question of whether to use a regular Tobit or a random-effects Tobit Regression had to be answered. In Stata, this refers to the tobit and xttobit commands. The output of the
xttobit model when the tobit option is
included can be used to answer this question: The rho in the output is larger than 0 indicating that there is a difference
between both models. Additionally, the likelihood-ratio (LR) test compares the random-effects regression and the pooled Tobit regression with the null hypothesis being the absence of panel effects. The chi-bar-square of 13.97 (p-value < 0.01) indicates the null hypothesis can be rejected so that I conduct the random-effects Tobit regression (xttobit). As some companies also perform multiple acquisitions per year, the acquisition year could not be used as a panel identifier. Hence, only acquisitions of the same company are grouped together. To still take account of year effects, I include the acquisition year as a categorical variable.
As can be seen from the scatter plots in Appendix 4, heteroscedasticity seems to be present in the data. Heteroscedasticity assumes a dependency of the variance of error terms on the independent variable and can be adjusted for via the standard errors (Wooldridge, 2016). Ignoring heteroscedasticity in a Tobit regression can bias its estimates severely (Brown, 1983). Therefore, I make use of bootstrapped standard errors (robust standard errors cannot be used with xttobit). Bootstrapping is a resampling method that treats the sample as the entire population and is used when standard errors are hard to obtain or possibly incorrect (Wooldridge, 2016).
35
4. FINDINGS
4.1 Descriptive statistics
The final sample consists of 336 acquisitions from the years 2016 to 2019 ranging over seven different industry sectors. An overview of the yearly distribution per industry is given in Table
2. Most acquisitions were conducted by family firms in the Consumer and Advanced
Manufacturing and Mobility sectors. Descriptive statistics of the continuous variables are
provided in Table 3. Figure 2 shows the histogram of the dependent variable Acquisition stake: Around 80% of the observations are takeovers where the acquirer bought 100% of the target firm. The mean level of control in the sample was 89.8% (see Table 3). This seems high but is only slightly above the mean of other studies that investigate on equity choices in international acquisitions (Chari & Chang, 2009: 89%; Cuypers et al., 2015: 87%; Gaffney et al., 2016: 83%; Malhotra & Gaur, 2014: 85%), which enhances the credibility of the data. Average family ownership is 63.2% and in only 23.8% of the deals the family firm is run by a family CEO. The average share of family board members is 18.2%. In total, family firms from 30 different countries are part of the analysis. Interestingly, family firms from Germany and Switzerland conducted most acquisitions (Appendix 5).
Table 2: Year and industry sector of family business acquisitions
Industry sector Year
2016 2017 2018 2019 Total
1 - Consumer 29 31 21 17 98
2 – Advanced Manufacturing & Mobility 29 26 20 19 94
3 – Financial Services 10 13 5 4 32 4 – TMT 12 9 14 11 46 5 – Energy 4 5 6 6 21 6 – Smart Infrastructure 5 5 9 7 26 7 – Health Sciences 7 5 4 3 19 Total 96 94 79 67 336
Table 3: Descriptive Statistics
Variable Obs Mean SD Min Max
Acquisition stake 336 .898 .252 .05 1 Family ownership 336 .626 .203 .32 1 Family CEO 336 .232 .423 0 1 Family board 336 .182 .15 0 1 Firm age 336 4.244 .662 2.485 5.861 Firm size 336 10.48 1.166 7.709 14.648 Performance 336 .067 .053 -.136 .306 Organizational slack 336 1.309 6.115 -.139 112.149 Subsidiaries abroad 336 4.623 1.355 0 7.228
4.2 Results
The purpose of this study is to investigate how different variables of ‘the family’ affect the level of control chosen in international acquisitions. The results of the random-effects Tobit regression can be found in Table 4. Model 1 runs the basic model. Model 2 includes all three variables of family firm heterogeneity simultaneously. Then, model 3 also includes the interaction terms of family ownership and management introduced in hypothesis 4a and 4b. The log-likelihood of all three models lies very closely together. However, the Wald chi² suggests that model 2 and model 3 are significant. Coefficients of a Tobit regression can similarly be interpreted as OLS coefficients. Yet, the estimates give the linear effect on the uncensored latent variable underlying the model and not the observed dependent variable (Wooldridge, 2016). However, as bootstrapping only leads to convergence towards the true parameters and therefore corrects the bias but does not entirely overcome it (Guan, 2003), the coefficients of this Tobit should only be interpreted with great caution so that I will focus on the direction of the effect.
Family ownership was expected to positively influence the level of control in the acquisition
target. The coefficients are indeed positive yet not statistically significant. Moreover, they have large standard errors. The coefficient of family ownership is 0.199 in model 2 and 0.050 in model 3. Thus, hypothesis 1 is not supported.
I hypothesized that having a Family CEO will positively impact the level of control in an acquisition. Here, both models point in the opposite direction and are even statistically significant on the 5% level. In model 2, the coefficient of -0.440 and in model 3 it is -0.428 respectively. These results suggest that hypothesis 2 can be rejected.
For Family board, I also expected a positive relationship with the level of control in acquisitions. Results support the direction of the effect with a coefficient of 0.183 in model 2 and 0.214 in model 3 yet these results are not significant. Therefore, also hypothesis 3 is not supported.
Last, the interaction terms are also statistically insignificant with large standard errors. The interaction Family ownership and Family CEO is positive with an estimator of 0.474 and the interaction Family ownership and Family board is negative with an estimator of -0.707 so that also no support for hypothesis 4a and 4b is found.
It is worth noting that the control variables have consistent signs across all models and Firm
age is statistically significant (p-value < 0.1) with a coefficient of 0.253 in the basic model 1
37
Table 4: Results of the random-effects Tobit regression
Model (1) Model (2) Model (3)
DV: Acquisition Stake Family ownership - 0.199 0.050 (0.519) (0.547) Family CEO - -0.440** -0.428** (0.222) (0.218) Family board - 0.183 0.214 (0.634) (0.957)
Family ownership # Family CEO - 0.475
(1.229)
Family ownership # Family board - - -0.707
(3.317) Firm size -0.0212 -0.0624 -0.063 (0.105) (0.102) (0.085) Performance 3.251 2.953 2.913* (2.047) (2.335) (1.627) Organizational slack 0.0373 0.0539 0.0587 (0.0979) (0.118) (0.117) Subsidiaries abroad -0.0670 -0.0558 -0.0544 (0.0928) (0.0943) (0.1061) Firm age 0.253* 0.187 0.193 (0.147) (0.151) (0.161) _cons 1.054 1.697 1.769 (1.106) (1.413) (1.097) sigma_u 0.604*** 0.580*** 0.576*** (0.109) (0.183) (0.158) sigma_e 0.661*** 0.659*** 0.659*** (0.0822) (0.0757) (0.0757) N 336 336 336 Log likelihood -151.91291 -149.30422 -149.11047 Wald chi² 16.70 36.47 30.88 P value 0.2725 0.0040 0.0416
Bootstrapped standard errors in parentheses
* p < 0.10, ** p < 0.05, *** p < 0.01
In model (3), Family ownership and family board are mean-centered. The interaction was conducted in model (3) after mean centering. For reasons of clarity, the coefficients for the years and sectors are not indicated. There were no significant results.
4.3 Robustness Check
To verify the robustness of the results above, I execute two robustness checks that consist of three different models. The structure of the models 1 to 3 is as above.
coefficient. Across all three models, firm age and performance have a positive and significant coefficient.
The second robustness check is conducted using a binary Logistic regression to verify the signs of the Tobit models. As homoscedasticity is no assumption of binary Logistic regressions (Osborne, 2017), potential problems caused by heteroscedasticity can thereby be avoided. The DV is dichotomous and takes the value of 1 for 100% acquisitions and 0 for all other acquisition stakes (Lahiri et al., 2014). Before the analysis, I checked for a sufficient sample size (van Smeden et al., 2019) and multicollinearity. Descriptive statistics for the dichotomous dependent variable are provided in Table 5. Following the LR test, I use a random-effect Logistic regression. Results in the form of logits are shown in Appendix 7. As with logits only the sign can be used for interpretation (Mood, 2010), this robustness check is suitable to verify the direction of the effects. Results overall confirm the directions and significances of the Tobit regression results.
Table 5: Descriptive statistics of the dichotomous DV
Acquisition Freq. Percent Cum.
0 59 17.56 17.56
1 277 82.44 100.00
Total 336 100
