Master Thesis
By
Tessa Henze – s1999710
University of Groningen Faculty of Economics and Business MSc BA - Small Business & Entrepreneurship
Supervisor: Dr. A.J. Rauch – Co-assessor: Prof. D. P.S. Zwart June 2015 – Word Count: 13532
ARE FAMILY FIRMS MAKING THE RIGHT SUCCESSION CHOICE?
How education influences the choice of succession in medium-sized firms in The Netherlands and its
implications for firm performance
1
Table of Content
Abstract ... 2
Key Words ... 2
1. Introduction ... 3
1.1 Problem Formulation ... 5
1.2 Structural Approach ... 6
2. Theoretical Background ... 7
2.1 Theory Foundation and Research Objective ... 7
2.1 Development of Constructs and Definitions ... 8
SMEs ... 8
FFs and NFFs ... 10
Level of Education of the Predecessor ... 11
Performance ... 11
2.2 Relations between Variables ... 12
FF and NFF Comparison ... 12
Succession Choice Differences ... 12
Education and Succession Choice ... 13
Succession Choice and Performance ... 15
3. Methodology ... 17
3.1 Sample ... 17
3.2 Data Preparation ... 19
Clearing Dataset ... 19
Operationalization of Variables ... 19
3.3 Method of Analysis ... 25
Findings ... 26
4. Discussion of Results and Conclusion ... 29
4.1 Practical and Theoretical Implications ... 31
2
4.2 Limitations and Future Research ... 32
5. References ... 33
6. Appendices ... 40
6.1 Appendix I: Questionnaire in Dutch ... 40
6.2 Appendix II: SIC Industry Codes ... 45
Abstract
Contradictory practical and theoretical views need to reach a consensus. Exactly this is the case for the opposing views on performance outcomes due to family involvement, family firm status and ultimately, succession choice. Furthermore, it remains unclear what exactly influences this succession choice of family firms compared to non-family firms.
This study aims to come to a consensus on how the education level of the predecessor influences the choice of succession in medium sized FFs and NFFs in The Netherlands and its implications for firm performance. The results suggest that even though the level of education is not the determinant of succession choice, experience plays an important role of intergenerational versus external succession. Furthermore, the succession choices specific for family firms and non- family firms do indeed lead to different financial performance outcomes.
Key Words
Family Firms, SMEs, The Netherlands, Education, Succession Choice, Performance
Master Thesis 2015, Family Firms in The Netherlands
3
1. Introduction
Family firm performance and business success remain important areas of interest for researchers due to their worldwide economic importance (Schulze & Gedailjolic, 2010). Family firms (FFs) are defined by the GEEF (European Group of Owner Managed and Family Enterprises) as firms where “The majority of ownership (directly or indirectly) rests in hands of a natural person and/or family; and at least one representative of the family or kin is involved in management or administration of the firm” (Flören et al., 2010; p. 2). Throughout the whole world, FFs are responsible for 70% to 90% of the annual GDP (Global Data Points, n.d., para.
5) and make up a large part of all firms (Perrini et al., 2007). This demonstrates that FFs are of great economic importance as they are the dominating type of firm and consequently account for a large part of the global employment and income (Perrini et al., 2007). However, research on performance influencing factors specific to FFs as well as on what sets them apart from non- family firms (NFFs) has not reached a useful consensus yet.
Previous research on team theory in FFs by Pearson et al. (2014) suggests that there are a lot of positive characteristics such as an exceptionally unified team, a shared vision and employee satisfaction associated with family involvement. The authors argue that the very nature of family relationships is essential for the uniqueness of FFs compared to other organizational forms. These typical FF characteristics can be proposed as performance- increasing factors which in practice might also be important for NFFs in order to obtain a more successful strategy and be more profitable (Kotlar, Fang, De Massis & Frattini, 2014). However, in recent years, many opposing views have emerged, challenging the previous consensus that FFs are good for firm performance. More and more research indicates that family involvement is actually negatively related to firm performance (Bloom & Van Reenen, 2007; Bennedsen et al., 2007). According to Lippi and Schivardi (2014), FFs specifically have worse executive selection in terms of succession due to nepotism, which accounts for a 6% productivity loss compared to NFFs. This FF specific intention to shape and pursue the vision of the business held by a dominant coalition in a manner that is potentially sustainable across generations of the family (Chua, Chrisman & Sharma, 1999), meaning intergenerational succession (IS) opposed to external succession (ES), was reported by various studies to have high failure rates (Davis & Harveston, 1998; Kets de Vries, 1996; Ward, 1987; Handler, 1992; Morris et al., 1997).
In order to discover the reason behind the above stated opposing views on family involvement and firm performance, academics as well as practitioners have been interested in
4
the factors that contribute to team effectiveness within top management teams (TMTs) and the influence on firm performance (Pearson et al., 2014) as FFs and NFFs differ in terms of TMT characteristics (Jorissen et al., 2005). According to Hambrick and Mason (1984), the upper echelon theory (UET) defines that organizational outcomes such as strategic choices and performance levels are determined by team specific managerial background characteristics such as age, gender, education and experience, which in turn influence the choice of succession, namely IS versus ES (Ganter et al., 2014).
After establishing that FFs and NFFs differ on TMT characteristics as well as on the major performance influencing dimension of succession choice, it is reasonable to be interested in what TMT characteristic specifically influences this choice between IS and ES. A recent meta- analysis by Tom van Esch (2015) has investigated the antecedents influencing the choice of succession in FFs. The research has shown that the level of education of the predecessor negatively influences IS, which is said to have a negative effect on firm performance (Bennedsen et al., 2007). According to Thornton et al. (2012), a predecessor should attempt to find the most capable successor, not taking family ties, thus IS, into account. Consequently, the higher educated the predecessor, the more likely he will see that ES is the most reasonable choice since there is a larger pool of highly educated potential candidates external to the family (Fairclough & Micelotta, 2013). Thus, it can be assumed that the higher the level of education of the predecessor, the higher is the likelihood of ES, which is shown to have positive performance outcomes (Chung & Luo, 2013). On the contrary, the lower the level of education of the predecessor, the higher the likelihood of IS, the lower the performance outcome (Bennedsen et al., 2007).
Concluding this introductory section on contradictory views of family involvement, succession choices and how the education level of the predecessor influences them, it remains unclear which form of business and thus succession choice is the one that leads to the most favourable performance outcome. To summarize, as firstly shown by global facts, the family involvement is said to have a positive influence on firm performance due to FF specific characteristics, however, the succession choice associated with keeping the business in the family is said to negatively influence performance. On the other hand, NFFs do not have the performance increasing advantage of FF specific performance increasing characteristics, but they rely on ES which boosts performance outcomes. Either way, the level of education of the predecessor has been determined to influence succession choice, as more informed decisions can be expected, leading to an increase in ES.
5
However, next to the theoretical reasoning, there is also a practical side to consider which will be discussed in the problem formulation.
1.1 Problem Formulation
After the short theoretical introduction, one major practical problem comes to mind, namely, are FFs making the right succession choice? Since there are opposing views on performance outcomes due to family involvement, family firm status and ultimately, succession choice, it is very important to firstly determine which approach applies to the area of interest of this study, namely The Netherlands; and secondly, to determine how family firms can make a more informed choice on the succession matter in order to improve firm performance.
In line with the global view, FFs are essential for the economy in The Netherlands as they are less influenced by economic unrest as they take fewer risks and are more long-term oriented compared to NFFs (Flören & Jansen, 2010). In addition, FFs are more innovative and of great importance to Dutch employment statistics (bvdinfo, 2015). Since family firms are usually small to medium sized enterprises (SMEs) (Bach, 2009), the above stated theory and the practical problem is most applicable in that firm size sector. Even though firms in the SME sector individually generate less financial turnover than larger firms, they still make a great financial contribution to the economy due to the fact that in The Netherlands, SMEs account for approximately 99% of active companies; they create work for over 3.5 million employees, which is more than half of the total labour force, and together make up almost 50% of the domestic gross value added (European Fact Sheet, 2014; bvdinfo, 2015). This shows the great level of economic importance and makes a point for the practical relevance and contribution of this paper which aims to provide an overview of IS versus ES and their influence on performance, especially focusing on the differences between FFs and NFFs.
Since FFs clearly contribute so much to today’s business world, it is very disconcerting that there are still opposing views on the implications of FF business status on performance. It should be clear why FFs are still doing so well if they rely on IS even though that is said to negatively influence firm performance, thus determining if FFs are still a feasible choice of status in today’s business world. Furthermore, research should come to a consensus to determine if FFs should actively make that choice towards staying a FF by finding an internal successor or if they should strive for external succession in order to boost performance. Since the level of education of the predecessor has been shown to influence the succession choice, it is important to investigate if there are differences between FFs and NFFs predecessors’ level of
6
education in The Netherlands and how the education level influences the presumably performance enhancing or hindering succession choice.
Consequently, this study aims to test how the education level of the predecessor influences the choice of succession in medium sized FFs and NFFs in The Netherlands and its implications for firm performance. Furthermore, it contributes to the literature and to the decision making power of FF predecessors on succession choice by reaching a consensus about which firm type (FF vs. NFF) reaches better performance outcomes due to distinct succession choices.
1.2 Structural Approach
The research will be tackled by firstly looking at the theory foundation and the research objective which lay the groundwork for the rest of the paper. The construct development section will describe the various definitions used in this paper and set the framework for the research.
It is necessary to delve into the importance of SMEs and make a distinction between the two categories of FFs and NFFs by providing definitions, clearly drawing a line within this paper where other literature comes to no consensus. Defining sections on level of education, succession choices and performance will also be given. Furthermore, the relations between the variables will be discussed, thus developing the hypotheses. The differentiating factors for FFs and NFFs, as well as the related different succession choices will be looked at, followed by a detailed description of why education plays an important role in making the succession choice and what implications this has for firm performance and success. Then, the methodology will describe the processes of data collection, data preparation, analysis and findings in detail, followed by a discussion of the results which links them back to the theory foundation, the problem foundation and ultimately the research objective. Lastly, the conclusion section will summarize all findings and implications to give an overview of how the educational level of the predecessor influences the succession choice and its implications for firm performance in medium-sized family and non-family firms in The Netherlands.
7
2. Theoretical Background
2.1 Theory Foundation and Research Objective
The family involvement has positive influence on firm performance due to FF specific characteristics (Pearson et al., 2014), however, the succession choice associated with keeping the business in the family is said to negatively influence performance (Bennedsen et al., 2007).
NFFs do not have the performance improving advantage of FF specific performance increasing characteristics, but they rely on ES which boosts performance outcomes (Chung & Luo, 2013).
Either way, the level of education of the predecessor has been determined to influence succession choice (Ganter et al., 2014), as more informed decisions associated with higher level education can be expected to be leading to an increase in ES (Fairclough & Micelotta, 2013).
Consequently, it remains unclear which form of business and thus succession choice is the one that leads to the most favourable performance outcome. This study aims to test how the education level of the predecessor influences the choice of succession in medium-sized FFs and NFFs in The Netherlands and its implications for firm performance.
The following theory can be derived from the theory foundation and the practical problem formulation: Medium-sized FFs in The Netherlands are facing a choice between either staying a FF, keeping the positive performance increasing characteristics associated with that status, but opting for performance decreasing IS; or choosing to give up the FF status and the performance increasing characteristics, but gaining a performance advantage due to opting for ES.
The research objective of this thesis needs to be split into two parts due to the shape and flow of the conceptual model which can be found below, and the type of analysis which will be comparative between the succession choices IS and ES, and the firm types of FFs and NFFs.
The first part looks at how the level of education of the predecessor influences the choice of succession, namely intergenerational or external succession. The second part focuses on the outcome of the succession choice, namely the influence on firm performance.
The following research questions are: How does the level of education of the predecessors influence the succession choice, differentiating between FFs and NFFs?; How does the succession choice influence firm performance, differentiating between FFs and NFFs?
8 Figure 1. Conceptual Model
2.1 Development of Constructs and Definitions
Since the definitions of the constructs used have not reached a consensus in previous literature, the following part will describe and define each item according to how it will be defined within this thesis.
SMEs
The focus of this study lies on FFs in The Netherlands. Even though individually seen, firms in the SME sector generate less financial turnover than larger firms, they still make a great overall financial contribution to the economy (Flören, 1998). This is backed up by the fact that the SME sector accounts for approximately 99.79% of all firms in the world (Perrini et al., 2007). Furthermore, the European Comission (2014) reports that more than 99% of all EU businesses are SMEs. “They provide two out of three of the private sector jobs and contribute to more than half of the total value-added created by businesses in the EU. Moreover, SMEs are the true back-bone of the European economy, being primarily responsible for wealth and economic growth, next to their key role in innovation and R&D” (European Comission, European SMEs, 2014; p.2). However, statistical numbers for the SME sector should be treated with caution since they are highly dependent on the definition of SME used, which can vary greatly.
Choice of Succession
Industry Firm Size Firm Age Gender
Age Experience
H3 i)
H3 ii) H2 ii)
H2 i)
9
Both the Centraal Bureau voor de Statistiek (CBS) and the EU define firm size in a quantitative manner, yet, they use different definitions for SMEs as shown in Tables T1 and T2. The CBS solely relies on employee count, whereas the EU definition additionally includes financial information on turnover and balance sheet total.
Table T1. Centraal Bureau voor de Statistiek criteria for dutch firm sizes
Firm size Employee count
Small firms <10 employees
Medium-sized firms 10-100 employees
Large firms >100 employees
Source: Centraal Bureau voor de Statistiek, http://www.cbs.nl/nl-NL/menu/home/default.html
Table T2. EU definition for small and medium sized firms
Firm size Employee count Turnover OR Balance Sheet Total
Medium < 250 ≤ € 50 m ≤ € 43 m
Small < 50 ≤ € 10 m ≤ € 10 m
Micro < 10 ≤ € 2 m ≤ € 2 m
Source: European Commission (2014), http://ec.europa.eu/enterprise/policies/sme/facts- figures-analysis/sme-definition/index_en.html
Alternatively to firm size, SMEs can also be defined qualitatively. According to Storey and Greene (2010), small businesses are defined as being “owned and managed by the same individual(s); have to be legally independent; and have a small share of the marketplace”
(p.32).
As there are various definitions on SMEs and even though the first intent of this study was to look at SMEs, small firms were excluded in order to narrow down the sample size. It was assumed that many small firms would not have a TMT, thus many small firms would not be eligible for this study. Even though the focus lies on The Netherlands, in terms of a consensus, it was chosen against the CBS definition because the EU definition additionally defines financial boundaries to distinguish between micro, small, and medium-sized firms.
Ultimately, the focus of the sample was on the total number of employees of medium sized companies. A minimum of 50 and a maximum 249 employees were used, based on the European definition for medium sized firms. Furthermore, the financial criteria imposed by the EU definition were used, namely firms that have to have a turnover between 10m€ and 50m€
or total assets between 10m€ and 43m€.
10 FFs and NFFs
Since it was established that there is need to differentiate between FFs and NFFs in the SME sector due to differing performance outcomes which should be further investigated, a consensus should be reached in terms of a FF definition. However, similar to the SME definition, the FF literature discusses many varieties of definitions as to what a FF actually is.
Still, no consensus has been reached about this definition, thus it is critical to reach a definition that combines existing definitions and will be used throughout this paper.
A truly perception based definition is given by Johannison and Huse (2000), who state that a family owner-managed company is considered to be a family business if perceived as such by the owners. Since this definition proves to be rather non-scientific and not measurable, many researchers have developed more quantitative definitions for family businesses. Flören (1998), considers a firm to be a FF if it meets at least one of the following criteria: More than 50% of shares are owned by a single family; A single family can exercise considerable influence; A significant proportion of the members of the board of directors are from one family. Alternatively, the European Group of Owner Managed and Family Enterprises (GEEF) defines FFs as follows: “The majority of ownership (directly or indirectly) rests in hands of a natural person and/or family; and at least one representative of the family or kin is involved in management or administration of the firm” (Flören et al., 2010; p. 2). Additionally, succession is also a determinant of family involvement in many definitions for FFs (Giarmarco, 2012).
While reviewing various FF definitions, Chua, et al. (1999) found that there is a general agreement that FFs involve family in the ownership and management. Even with this agreement on family involvement in ownership, researchers differ on the degree of involvement. Smyrnios et al. (1998) state that 50 percent or more ownership must be held by members of a number of families, whereas La-Porta et al. (1999) recommend a 20 % family stock ownership.
To come to a conclusion, the previously stated GEEF criteria combine the most important facts of FFs in a comprehensive manner and will be used throughout this paper in terms of defining a FF. As an addition, the majority of ownership described in the GEEF definition was substituted by 20% stock ownership as suggested by La-Porta et al. (1999) in order to differentiate on a numerical basis.
11 Level of Education of the Predecessor
Empirically, the relationship between education and succession is discussed in several publications, however, no clear definition for higher and lower education was formulated in exact years (Glauben et al., 2009; Kimhi & Nachlieli, 2001; Pérez-González, 2006; Stiglbauer
& Weiss, 2000; Sifei, 2013; Zacher et al., 2012), and thus the standard education levels available in The Netherlands were used to account for level of education. These are according to the Vereniging Hogescholen NL (2015): Primary school, VMBO, HAVO and VWO and MBO, HBO and WO Bachelor, HBO and WO Master and Doctor. Furthermore, as this study is based on medium-sized firms in The Netherlands, it was assumed that each firm has a TMT which acts as a team. However, no information on decision making power was included in the questionnaire, thus the level of education is team based since it cannot be determined which TMT member is making the decision on succession choice, thus equal decision making power is assumed.
Performance
Business management studies often aim to gain knowledge on business success as success is the ultimate performance indication, even though it is difficult to measure as different determinants have to be evaluated. Success is an elusive concept as each person and firm perceives it differently, and has principles and definitions that determine success for them.
However, the predominant form for firms to measure their performance is by the balanced scorecard approach developed by Kaplan and Norton: “a performance measurement framework that adds strategic non-financial performance measures to traditional financial metrics to give managers and executives a more 'balanced' view of organizational performance” (Balance Scorecard Institute, 2015). But as this master thesis focuses on medium-sized firms, a rather pragmatic approach to performance measures has to be taken since when looking at this category. Researchers have assessed the implementation of the Balanced Scorecard in SMEs, but concluded that this model is not suitable for such businesses (Hvolby & Thorstenson 2000;
McAdam, 20003) as even if the model was applied correctly, it would be inadequate for the particular characteristics of SMEs. In previous research, a variety of hypotheses and theories on how this business success can be obtained have been developed; the most common being financial performance as stated by Jasra (2011).
However, no perfect performance measure for SMEs has been determines yet (Hudson et al., 2001), thus this study was restricted to the data gathered and additionally obtained within
12
the Orbis database. Various authors propose industry, firm size and firm age as performance influencing factors (Murphy et al., 1996; Powell, T.C., 1996). Additionally, Powell (1996) showed that industry can account for more than 10% variance in firm performance.
Consequently, two performance measures arising out of the above stated factors were chosen, namely asset performance, which looks at the firms total assets compared to industry averages on total assets, and growth performance, which looks at the company age divided by the company size in employee numbers. This way, both a financial, as well as a non-financial measure were used to define firm performance.
2.2 Relations between Variables FF and NFF Comparison
Since FFs and NFFs differ in terms of top management teams (TMTs) (Jorissen et al., 2005), academics as well as practitioners have been interested in the role of TMTs on firm performance, and ultimately business success (Pearson et al., 2014). More specifically, Pearson et al. (2014) have been trying to determine which factors contribute to team effectiveness and firm performance. According to Hambrick and Mason (1984), the upper echelon theory (UET) defines that organizational outcomes such as strategic choices and performance levels are determined by managerial background characteristics. The UET focuses on the fact that organizational outcomes are directly influenced by the knowledge, experiences and expertise of the individuals in managerial roles in the organization (Hambrick & Mason, 1984). Thus, it can be assumed that each TMT has its own set of distinguishing characteristics determined by the characteristics of the TMT members, which influence its performance outcome and business success.
Succession Choice Differences
Due to the opposing views on the comparison between FFs and NFFs, it is important to highlight what exactly differentiates FFs and NFFs other than TMT characteristics. FFs can be simply distinguished from NFFs by the intention to shape and pursue the vision of the business held by a dominant coalition in a manner that is potentially sustainable across generations of the family (Chua, Chrisman & Sharma, 1999). In order to keep an enterprise within the family and to maintain activity and management through succession, an internal successor must be found. This is underlined by a recent study by Tom van Esch (2015) showed that IS appeared to be the preferred form of succession in FFs opposed to ES. On the contrary, NFFs succession
13
choices do not include IS, but within the scope of this paper, only ES if the company wants to stay a NFF. At some stage of the life cycle (Jones, 2010), all firms have to be either transferred to the next generation, internally or externally chosen, be sold, or go out of business (Martinson, 2012). Since this paper is concerned with the continuity of FFs and NFFs and their performance outcomes, both the sale as well as death will not be regarded further. As Pitcher, Cherim and Kisfalvi (2000) stated, succession planning is essential regarding the success and continuity of businesses.
The first hypothesis serves as a control of the findings of the Master Thesis by Tom van Esch (2015) and its applicability on this studies sample of medium-sized enterprises in The Netherlands. Furthermore, it adds to van Eschs finding by including NFFs predominant succession choice. This hypothesis is not included in the conceptual model and it will be tested on the firm level.
Hypothesis 1: FFs are predominantly associated with IS; NFFs are predominantly associated with ES.
Education and Succession Choice
After establishing that FFs and NFFs differ on the major performance influencing dimension of succession choice, it is important to look at what TMT characteristic may specifically influence this choice between IS and ES since performance outcomes vary between them. According to Hambrick and Mason (1984), the upper echelon theory (UET) defines that organizational outcomes such as strategic choices and performance levels are determined by team specific managerial background characteristics such as age, gender, education and experience, which in turn influence the choice of succession, namely IS versus ES (Ganter et al., 2014). A recent meta-analysis by Tom van Esch (2015) has investigated the antecedents influencing the choice of succession in FFs. The research has shown that the level of education of the predecessor negatively influences IS. According to Thornton et al. (2012) a predecessor should attempt to find the most capable successor, not taking family ties, thus IS, into account.
Consequently, ES is the most reasonable choice since there is a larger pool of highly educated potential successors external to the family (Fairclough & Micelotta, 2013). Thus, it can be expected that the higher the level of education of the predecessor, the higher is the likelihood
14
of ES. On the contrary, the lower the level of education of the predecessor, the higher the likelihood of IS.
Ganter et al. (2014) found evidence for the described relationships between education and the chosen form of succession. They confirmed the expectation that preference for IS decreases as the level of education of the predecessor increases. The authors describe that the reasoning of this finding is based on logics (Ganter et al., 2014). Horn (1983) defines logics as underlying assumptions which form a framework within which reasoning takes place. These assumptions are divided into corporate logics and family logics. Thornton et al. (2012) found that individuals base their role identities and goals on either corporate logics or family logics.
On the one hand, the corporate logic is defined by the predecessor treating succession in the role of a CEO (Thornton et al., 2012) that attempts to find the most capable successor by disregarding nepotism (Fairclough & Micelotta, 2013). Additionally, higher educated predecessors are more likely to be able to judge successor’s abilities in a more informed way (Barach et al., 1988). Following the corporate logic, ES is more likely as there is a larger pool of highly appropriate potential successors external to the family. On the other hand, the family logic is defined by the predecessor treating succession as a provider of the family (Miller et al., 2011). Following a family logic, IS is more likely as the control over the business stays within the family (Cennamo, Berrone, Cruz & Gomez-Meija, 2012) and future job opportunities for relatives are secured (Miller et al., 2011).
Ganter et al. (2014) concluded that the higher educated the predecessor, the more likely it is for him to act to a greater extent like a CEO following the corporate logic. They state that the corporate logic is mostly associated with NFF succession choice, whereas the family logic is mostly associated with FF succession choice, meaning that the level of education of the predecessor is less high in FFs compared to NFFs. Ultimately, the search for a family-related successor shifts to the search for the most capable successor, meaning that the preference for IS decreases and the level of ES increases as the level of education of the predecessor increases.
Research by Klyver (2007) underlines that statement, as the author found evidence that when predecessors have between three and four years of higher education, the odds of involving a family member decreases by 43%.
15
The second hypothesis focuses on the level of education of the predecessors and its influence on the choice of succession. As the level of education of the predecessors increases, the likelihood of IS decreases and the likelihood of ES increases. This hypothesis will be tested on a team level as the described TMT characteristics are a mean of characteristics of each TMT member.
Hypothesis 2:
i) The level of education of the predecessor negatively influences the likelihood of IS.
ii) The level of education of the predecessor positively influences the likelihood of ES.
Succession Choice and Performance
Family firm performance remains an important area of interest for researchers due to its worldwide economic importance (Schulze & Gedailjolic, 2010). Throughout the whole world, family firms (FFs) are responsible for 70% to 90% of the annual GDP (Global Data Points, n.d., para. 5) and make up a large part of all firms (Perrini et al., 2007). In the United States, FFs comprise 80% to 90% of all businesses (Astrachan & Shanker, 2003); In Germany, FFs make up over 90% of all businesses (Haunschild & Wolter, 2010); In the Gulf Cooperation Council (Saudi Arabia, Kuweit and other of the Gulf states) approximately 98% of commercial activities are held by FFs (Fadhel, 2004). These numbers impressively demonstrate that FFs are indeed of great economic importance as they are the dominating type of firm and consequently account for a large part of the global employment and income.
Coming back to individual firm performance, interestingly, opposing views on the influence of family involvement on firm performance have emerged as studies have found that family involvement can be both positively as well as negatively related to firm performance.
On the one hand, FFs have been reported to be more profitable than NFFs, which can be accounted to their unique characteristics (Chua, Chrisman & Sharma, 1999). Chua et al. (1999) also determined that the family component shapes the business in a way that executives of NFFs are unable to achieve. In addition, Jasra et al. (2011) found that family involvement might positively affect performance of firms in the SME sector, as well as they have been found to be more profitable than NFFs (Kotlar et al., 2014). Pearson et al. (2014) argue that the very nature of family relationships is essential for the uniqueness and profitability of FFs compared to other
16
organizational forms. Consequently, these typical FF characteristics can be proposed as performance-increasing factors which in practice might also be important for NFFs in order to obtain a more successful strategy (Pearson et al., 2014). However, in recent years, many opposing views have emerged, challenging the consensus that FFs are good for firm performance. More and more research indicates that family involvement is actually negatively related to firm performance (Bloom & Van Reenen, 2007). Bennedsen et al. (2007) investigated Danish FFs and estimated a 4% profitability loss for firms having a family manager rather than a professional one. According to Lippi and Schivardi (2014), FFs have worse executive selection due to nepotism which accounts for a 6% productivity loss compared to NFFs.
Consequently, and opposing to previously stated numbers, family involvement could also be negatively influencing firm performance. Thus, it can be assumed that the higher the level of education of the predecessor, the higher is the likelihood of ES, which is shown to have positive performance outcomes (Chung & Luo, 2013). On the contrary, the lower the level of education of the predecessor, the higher the likelihood of IS, the lower the performance outcome (Bennedsen et al., 2007).
The third hypothesis looks at the influence of the choice of succession on firm performance.
Since firm performance is measured on the firm level, this hypothesis will be tested at firm level as well.
Hypothesis 3:
i) IS negatively influences firm performance.
ii) ES positively influences firm performance.
17
3. Methodology
3.1 Sample
To assess the research objective, thus the research question itself and the following hypotheses on the relations between the education of the predecessor, succession choice and firm performance in FFs and NFFs, data were gathered in two different ways.
Firstly, an extensive literature review was conducted which led to the actual problem formulation, the theory foundation and the theoretical background information already stated.
Secondly, the sample for the statistical analysis was decided on, namely the resulting SPSS dataset and Excel sheets from the Business Diagnosis and Design course by Maryse Brand (2015). The BD&D study was part of a larger PhD research that focused on the differences between FFs and NFFs in The Netherlands. The sample was gathered from the Orbis database that provides information on nearly 150 million companies worldwide (bvdinfo, 2015). It was narrowed down to fit the selection criteria imposed by the study in terms of the medium-sized firm definition by the EU, size determined by employee count, and financial criteria to arrive at a total sample of 1485 firms.
Selection criteria in Orbis were applied in order to reject firms that could not be used for the purpose of this research. First, all active companies in 2013 in the country region of The Netherlands were selected in order to rule out inactive firms. The year of 2013 was used because the information of 2014 was not yet available at that moment of time. The ownership status was looked at together with the exclusion of foreign subsidiaries. Since the focus of the study was on medium-sized businesses in The Netherlands, firms such as subsidiaries were left out.
Furthermore, no firms with foreign TMTs or owners in Dutch companies (above 50.1 % by foreign owner) were accepted, thus ruling out high percentages of foreign involvement. Even though the first intent was to look at small and medium sized firms, it has been decided to exclude the small firms in order to narrow down the sample size. The reason why small firms were excluded instead of medium-sized firms was that it was assumed that many small firms would not have a TMT. Therefore, many small firms would not be eligible for this study.
Ultimately, the focus of the sample selection was on the total number of employees of medium- sized companies. A minimum of 50 and a maximum 249 employees were used, based on EU definition for medium-sized firms. Furthermore, the financial criteria imposed by the EU definition were included, namely firms that have to have a turnover between 10m€ and 50m€
or total assets between 10m€ and 43m€. Financial services like insurance, reinsurance and pension funding and public authorities/States/Governments were excluded because they are not
18
part of the private sector. Finally, the national legal forms of The Netherlands, such as general partnerships (VOF, NV, BV) were included because the other legal forms were considered to be too small to have a TMTs. A summary of each selection criteria with its justification is shown in table M1. After these criteria were selected for, a total of 1485 firms were left that complied with the criteria selected for this research project.
Table M1. Sample Selection Criteria
Selection Criteria Justification
Active companies in 2013 in The Netherlands Region of focus
Number of employees: 2013, min=50, max=249 EU definition for medium sized firms Operating revenue (Turnover) (th EUR): 2013,
min=10,000, max=50,001
Total assets ( th EUR): 2013, min=10,000, max=43,001
EU definition for medium sized firms
Exclusion of companies with no recent financial data and Public authorities/States/Governments
Not privately owned
National legal form: The Netherlands, General partnership - V.o.f. (The Netherlands), Private limited liability company - BV (The Netherlands), Public limited liability company - NV (The Netherlands)
EU definition for medium sized firms
Foreign subsidiaries: Def. of the UO: min. path of 50.01%, known or unknown shareholder
Ruling out high percentages of foreign involvement.
A subsample was randomly chosen and the questionnaire was distributed amongst 921 FFs and NFFs both per email and telephone. The selection process by numbers is shown in table M2. Even though the scope of this study is on the firm level, the data collection occurred on the individual level. Respondents were TMT members who were expected to provide the most accurate information regarding the interview.
Table M2. Sample Selection Process by Numbers
Population All medium-sized Dutch firms 4,211,325 Sample Frame List of all these firms
Sample Firms selected (1485) and approached (521 telephone, 400 email) Net Responses Number of complete and valid responses 96
19
The research objective of the BD&D course study was to compare TMT compositions of medium sized FFs and NFFs in The Netherlands based on demographics such as age, gender, education and tenure of TMT members. However, further variables such as succession, ownership status and family influence were also included, leading to an initial response dataset of 148 cases, of which several had to be excluded due to incompleteness. In terms of this master thesis research, the dataset was very valuable as it included both the education and the succession variables, as well as industry classifications, ownership status, TMT counts and employee counts, which were used to calculate the final variables. The need for them, as well as the preparation of the variables will be discussed in the following section. Additional data on firm performance measures were collected using the Orbis database for secondary data.
3.2 Data Preparation
Before the actual statistical analysis could be conducted, the database was cleared and the variables were operationalized and relevant new ones were computed and added. Secondly, the analysis was conducted by means of chi-squared test, logistic regression and multiple regression. These steps will now be discussed in detail, followed by the resulting findings.
Clearing Dataset
The initial data set consisted of 148 cases of which some were incomplete due to an automatic stop in the questionnaire that occurred when firms were part of a large concern, thus not included according to the selection criteria. After discarding these cases, 101 cases were left for further investigation. Furthermore, cases with no information on ownership status, succession, education and financial data were excluded, leaving a sample of 96 cases to be prepared and analyzed further.
Operationalization of Variables
The following sections show how the variables were derived by explicitly stating which definitions were used, as well as the processes in SPSS of computing them.
FFs and NFFs
The GEEF definition, which states that “the majority of ownership (directly or indirectly) rests in hands of a natural person and/or family; and at least one representative of the family or kin is involved in management or administration of the firm” (Flören et al., 2010;
p. 2), combines the most important facts of FFs in a comprehensive manner. This definition was
20
used to operationalize the FF_NFF variable. As an addition, the majority of ownership described in the GEEF definition was substituted by 20% stock ownership as suggested by La- Porta et al. (1999) in order to differentiate on a numerical basis.
The aim of this variable was to differentiate between FFs and NFFs in the most definitive way, meaning a dichotomous operationalization. The data were ordered into FFs and NFFs by means of the creation of a new variable FF_NFF. The previously discussed GEEF definition for FFs used in this paper relies both on majority of ownership and TMT participation, which both were stated by the firms. This new variable was computed to determine which firms are FFs and which firms are NFFs according to the definition. Tables M3 and M4 show the computing and recoding procedure.
Table M3. FF_NFF
Condition Variable (Question) Category (Answers)
Majority ownership in the hands of one family
Q7_a_ownship No = 0
Yes = 1 Unknown = 3 At least one family member in
the TMT
Q8_a_TMTfam No = 0
Yes = 1
Table M4. Computing FF_NFF
Q7_a_ownship Q8_a_TMTfam Total value
No = 0 N/A 0
Yes = 1 No = 0 1
Yes = 1 Yes = 1 2
This indicates that a FF, according to the definitions used in this paper, would have a total value of 2 as shown in the last line of Table M2, meaning that both the majority of the ownership, as well as a member of the TMT is provided by the family. The new variable was computed through the following procedure:
Transform > Compute new variable > FF_NFF = (Q7a_own =1) + (Q8a_TMTfam = 1) Because only two groups were needed, FF and NFF respectively, all cases with values of 0 and 1 were combined into a composite group NFF, meaning that all values of 0 were recoded into 1. Furthermore, both groups were recoded into FF=0, NFF=1.
21 Education
The education level of members of the TMT had to be recoded first as shown in Table M5, and then computed into a new variable, EDU_mean, by using the following procedure:
Transform > Compute new variable > EDU_mean = Mean (q12_d_educ, q13_ d_educ, …, q17_ d_educ)
Table M5. Recoding Variables Education
Variable Variable (Question) Category Old → New Values Education Q12 to Q17_d_educ Primary school 1 → 1
VMBO 2 → 2
HAVO, VWO, MBO 4 → 3
HBO, WO Bachelor 5 → 4 HBO, WO Master, Doctor 6 → 5
Empirically, no clear definition for levels of education was formulated, thus the standard education levels available in The Netherlands were used to measure level of education. As this study is based on medium-sized firms in The Netherlands, it was assumed that each firm has a TMT which acts as a team. However, no information on decision making power was included in the questionnaire, thus the level of education is team based and was calculated as a mean across the TMT, assuming equal decision making power. Even though it is not usual protocol to calculate a mean for an ordinal variable, in this case it is preferred as the objective is to show an increase and decrease in education of the predecessor or TMT where there are more members involved in the TMT. A median would not explicitly show the level of combined intelligence or level of education as the result would be the middle category. Since it cannot be determined which TMT member is making the decision on succession choice, equal decision making power is supposed, thus calculating the mean shows the average level of education that can be expected of the TMT in one firm and can be compared to others. Furthermore, it was checked whether the TMT count was corresponding with the filled in specifics for each TMT member. No missing values were found.
Succession
In terms of succession, the questionnaire contained five questions on the topic, of which four were used to deduct which firm chooses for which succession strategy. The left out question 9 e about how many times the firm was already transferred is sufficiently answered by question 9 b that asks for past intergenerational transfers. Firstly, the questions were recoded of
22
which the procedure can be seen in Table M6 underneath. Succession intent consisted of a 7- point Likert scale which was coded into two groups, namely low intergenerational intent and high intergenerational intent. The “undecided” category was coded into 1 (low IS intent) as well because the further calculations of the new variables SUC_choice and IS_ES shown in Tables M6 and M7 still gave undecided firms the chance to fall into the IS choice category. The other variables were also recoded in the same fashion as succession intent, meaning no = 1, yes = 2.
Table M6. Recoding Variables Succession
Variable Variable (Question) Category Old → New Values Succession Intent Q9_a_genintent Very unlikely 13 → 1 (low)
Unlikely 14 → 1 (low)
Somewhat unlikely 15 → 1 (low)
Undecided 16 → 1 (low)
Somewhat likely Likely
Verly likely
17 → 2 (high) 18 → 2 (high) 19 → 2 (high)
Succession Past Q9_b_pastgen Yes 9 → 2
No 10 → 1
Succession Future Q9_c_futgen Yes 9 → 2
No 10 → 1
Succession Current Q9_d_currtransf Yes 9 → 2
No 10 → 1
Table M7. SUC_choice
SUC_intent SUC_past SUC_future SUC_current Total Value
Low = 1 No = 1 No = 1 No = 1 =< 4 = Choice ES
High = 2 Yes = 2 Yes = 2 Yes = 2 > 4 = Choice IS
As a next step, the new variable SUC_choice was computed as a sum of SUC_intent, SUC_past, SUC_future and SUC_current. The total value and thus the classification into the succession choice by either internal or external means can be represented by the following equation: Choice ES =< 4 > Choice IS.
Table M8 below shows the categorization of SUC_choice into either IS or ES (variable IS_ES). Once a 2 (yes) was chosen, the firms were automatically part of the IS group as the total value of SUC_choice was above 4. Missing systems were coded to ES as
23
they were NFFs that did not answer this question as they always rely on ES. As previously discussed, there are more options for firms other than IS and ES, such as closing the company, however, this was not taken into account in this research as that would lead to no valuable insight on firm performance.
Table M8. IS_ES
Value SUC_choice Choice Succession New value
> 4 IS 0
=< 4 ES 1
Missing Systems ES 1
Performance
In terms of financial performance measures, the collected data was from the Orbis database for the year 2013. The assets per company were stated, but since companies differ on asset value depending on their industry, the variable asset performance compared to industry average was calculated. This included computing industry averages in a second SPSS database of all medium sized companies in The Netherlands. The five digit industry code was shortened down to a two numbered code, and recoded into industry groups according to the SIC classification which can be seen in Appendix II. Of each of them, the industry mean was calculated and transferred to the original datasheet. There, each company was already coded in the same industry classification, and the company assets were subtracted from the industry asset average, leading to the PER_Asset variable. Furthermore, a second performance variable not depending on financial measures was computed, namely growth performance. Both firm age and firm size were given, leading to the variable PER_Growth being computed by dividing firm age by firm size.
Control Variables
As the conceptual model is a two-step model, where H2 is on team level and H3 is on firm level, the control variables also differ accordingly. According to the UET, organizational outcomes such as strategic choices and performance levels are determined by team specific managerial background characteristics such as age, gender, education and experience; Thus, the control variables for H2 (EDU->SUC) are gender, age and experience. As the performance is measured at firm level, the controls influencing the succession choice performance relation were chosen as such as well. Industry, firm size, firm age were previously shown to affect
24
objective firm performance, thus, the control variables for H3 (SUC->PER) are industry, firm size and firm age. As previously stated, asset performance was calculated using the industry average. This means that for the test including the asset performance variable, the industry control will be left out. Similarly, as the growth performance was calculated using firm age and firm size, those two control variables will not be used in the test that measures growth performance.
Team Level Control Variables
The gender ratio, the age and the experience variables first had to be recoded as shown in Table M9 below. Both the age and the experience were further calculated in the same fashion as the education variable, namely a mean calculation for experience in the TMT, in order to have comparable means. In terms of gender ratio, the 0-1 coding was chosen in order to get percentages for females and males within teams. For example, if a team scores a 0.089, this indicated a percentage of 8.9% females and 91.1% males in the TMT.
Table M9. Recoding Team Level Control Variables
Variable Variable Range Category Old → New values
Gender Q12 to Q21_a_gender male 9 → 0
female 10 → 1
Age Q12 to Q21_b_age < 26 10 → 1
26-35 11 → 2
36-45 12 → 3
46-55 13 → 4
56-65 14 → 5
>65 15 → 6
Years of TM experience Q12 to Q21_c_TMexp 0-2 10 → 1
3-5 11 → 2
6-10 12 → 3
11-15 13 → 4
16-20
>20
14 → 5 15 → 6
25 Firm Level Control Variables
The industry codes were already recoded according to SIC standards shown in Appendix II while calculating the asset performance measure. Firm size was given by the firms in the form of employee count, and firm age was determined by subtracting the year of establishment from 2015.
3.3 Method of Analysis
The collected dataset was imported to SPSS and the raw data was processed, re-coded and prepared as described previously. Data analyses were performed using SPSS. This paragraph will provide a description of the analyses and which tests were used, followed by the findings of the analyses. In order to investigate the hypotheses and the research statement, firstly descriptive statistics were run, namely means, numbers, standard deviations and checks for normality and linearity. In order to investigate the relations between the variables, thus their independence, they were additionally tested using correlations between the FF_NFF and SUC variables, as well as between the EDU, SUC and PER variables and the controls, analysing the association between two variables at a time and determining if the control variables can be included in further analysis.
In terms of hypothesis 1, a chi-square test was performed to compare the observed frequencies with the expected frequencies derived from the literature. Hypothesis 2 was tested by means of a logistic regression analysis, in order to determine the odds probability occurring for the succession choice, namely IS and ES, as the values for level of education change. Model one included both the independent variable (EDU) as well as the controls, whereas model two only included the control variables in order to determine the strength for the relation between EDU and succession choice while accounting for the strength between the control and succession choice. For the last hypothesis 3, two multiple regression analyses were run, taking into account the two performance measures and their specific control variables as, due to the calculations, not all control variables were usable for both performance measures. For all calculations, no reliability measure was needed as all calculations stem from indices.
