Master Thesis, Msc BA, specialization Small Business & Entrepreneurship
University of Groningen, Faculty of Economics and Business
Top Management Team diversity of Dutch SMES: making a
difference by being different?
Rick de Lange
S3209075
University of Groningen
Faculty of Economics and Business
March 2021
Supervisor: dr. M.J. Brand
Co-assessor: M. Hanisch
Abstract
Demographic diversity of corporate top management teams (TMT) and its impact on firm performance has been a topic of considerable attention. Interestingly, the bulk of scientific investigations into this relationship focus on large enterprises and fail to address small to medium enterprises (SMEs). As the majority of businesses are SMEs, and the modus operandi of large enterprises is dissimilar compared to SMEs, a focused empirical investigation is warranted to address the effect of TMT diversity on firm performance in these companies. The current thesis reports on a survey-based investigation on the TMT members of 90 SMEs in 2018 and 2020 by inquiring the diversity of age, gender, nationality and tenure. Based on the Upper Echelons theory and socio- psychological theories, we initially hypothesized a positive correlation between TMT diversity and firm performance. In addition, we expected a negative interaction in the case of family firms but no effect was found. Based on our findings we found no significant relation for age, gender, nationality and tenure diversity on firm performance. Several explanations for these results will be discussed. Contrary to findings of similar studies in large companies, we report no statistical association between firm performance and TMT diversity. We discuss the implications of our findings and develop avenues for further research.
Keywords: Diversity, Top management team, Small and medium-sized enterprises, Upper echelons theory, Firm performance, Socio-psychological theories.
Table of content
1.Introduction 4
2.
Literature review 6
2.1 Diversity and firm performance 6
2.2 Theories 7
2.2.1 Upper echelons theory 7
2.2.2 Socio-psychological theories 9
2.3 Firm characteristics 11
2.3.1 Firm size 11
2.3.2 Family firms 12
2.4 Conceptual framework and hypotheses development 13
2.4.1 Conceptual framework 13
2.4.2 Hypotheses development 13
3.
Methodology 19
3.1 Data collection 19
3.1.1 Target population 19
3.1.2 Aggregated data 19
3.1.3 Data 2020 20
3.1.4 Survey 20
3.1.5 Collection procedure 21
3.2 Definition and measurement of variables 22
3.2.1 Top Management Team 22
3.2.2 Dependent variable 22
3.2.3 Independent variables 23
3.2.4 Moderator 25
3.2.5 Control variables 25
3.3. Analysis plan 26
4.
Results 28
4.1 Descriptives 28
4.2 Correlations 29
4.3 Multivariate logistic regression 31
4
Discussion and conclusions 34
5.1 Discussion and conclusions 34
5.2 Limitations and suggestions for further research 36
5.3 Implications for practice 38
References 39
Appendices 45
1.0 Overview data 2013 45
1.1 Overview data 2018 45
1.2 Search Strategy 2020 – Orbis 47
2.0 Survey 48
2.1 Introduction research by phone call 51
2.2 Template e-mail send to firm after a phone call. 51
3.0 Summary of additional tests and regression analyses 52
1. Introduction
Make a difference by being differentPeople are unique, in both observable and unobservable manners. Cultural backgrounds, social experiences and life experience form the backbone for our being and the decisions we make. As companies are led by individuals forming a top management team (TMT), the characteristics of team members are decisive in the way a company is governed. Traits of TMT members are determinants of a company’s strategic and by that of a firm’s performance (Crossland et al., 2014; Yohannes & Ayako, 2016). The diversity of characteristics in a TMT could thus have specific implications for the performance of a company. Therefore, the relationship between top management team (TMT) diversity and firm performance has been of major interest on academic and social grounds over the last decades. The current thesis will further discuss the relationship between TMT diversity and firm performance supported by an original investigation into the firm performance of small and medium-sized enterprises (SME) and the effect of TMT diversity.
Hambrick and Mason (1984) state that organizations are a reflection of their top managers. They established the ‘Upper Echelons’ theory, which investigates the demographics of team members and hypothesizes how that affects behavior and organizational outcomes. This theory formed the basis for many investigations into the importance of the managerial role and the features and the interaction of TMT members.Previous studies in the field of Upper Echelons Theory provided evidence that TMT diversity can improve firm performance (e.g. Carpenter, 2002), while other studies investigating TMT diversity have shown the contrary, namely that diversity impeded firm performance (Kochan et al., 2003; Nielsen, 2010). These inconsistent results make ‘TMT diversity’ an interesting and on-going subject of research.
Over the past years, companies have been looking to improve firm performance by diversifying their TMT (e.g. equal male/female ratio), whereby TMT diversity is defined as the degree of variety inherent in the TMTs composition (Hoffmann & Meusburger, 2018). TMT diversity is an interesting topic because organizations and TMTs are increasingly more heterogeneous due to the changing population demographics, as well government-imposed diversity regulations (Triandis, Kurowski, & Gelfand, 1994). Thereby, the selection for leadership positions is subject to different corporate laws, regulations and codes of EU countries. Civil rights gains made by woman and racial/ethnic minorities and the increasing female workforce are crucial to this change (Ansari et al., 2016). Furthermore, most studies into TMT diversity and firm performance focus on large firms (Chen, Kang, & Butler, 2019; Crossland & Hambrick, 2011), whileSMEs give rise to 60 to 70% of jobs in most countries (OECD, 2020) and there is limited knowledge on the relationship between SME TMTs and firm performance
(Faems, Sels, De Winne, & Maes, 2005). Human- and organizational processes within SMEs are essentially different from large firms (Koch & De Kok, 1999).Managers in SMEs seem more motivated (Nooteboom, 1994) and managers are pivotal for the success of a business as there is a small room for error. This means it is reasonable to assume that the impact of TMT diversity on firm performance within a SME differs from TMT diversity within larger firms.
The inherent differences between SMEs and larger firms could have strong implications on the effect of TMT diversity on firm performance. The results of investigations in larger firms might not be generalizable, given that there could be other legislation influencing the (diversity within) TMTs.
The goal of this thesis is to investigate the relationships of TMT diversity on the performance of SMEs. The managerial implications from this study are both practical and applicable in the selection of TMT members. The theoretical contributions of this issue are as follows: firstly, it adds to the on-going discussion on the effect of TMT diversity. Secondly, this thesis adds to literature by contributing newly garnered theoretical and empirical knowledge on TMT diversity and its impact on company performance. Thirdly, it contributes to the upper echelons theory by building on the current literature and combining this with additional socio-psychological theories. This study focuses on the relation between top management team diversity and firm performance of Dutch SMEs, which has led to the following research question
RQ: What is the impact of Top Management Team diversity on the performance of small and medium-sized enterprises?
This thesis consists of five chapters. This introduction is followed by a literature review into the key and core concepts foundational to the investigation. Based on the newly attained knowledge a conceptual framework and model is established to create a clear overview. The third chapter covers the methodology of this study. Chapter four presents the empirical results based on statistical analysis. Lastly, chapter five, concludes the paper and also present the limitations of the investigation and suggestions for further research.
2. Literature review
In this chapter, the area of research and its accompanying theoretical concepts will be described. The first section will give an overview of diversity and firm performance. Secondly, the upper echelons theory and three additional ‘social’ theories are discussed, as these are foundational to this study. In the third section, the firm characteristics will be outlined with two subsections: small and medium-sized enterprises and family firms. This chapter closes with presenting the conceptual framework and formulating relevant hypotheses.
2.1 Diversity and firm performance
Diversity is a concept that features in a variety of disparate disciplines and can be referred to situations in which the actors of interest are not alike with respect to some facets or to respect in which things differ. The definition of diversity is discussed by several scholars. While some debate that it should encompass the variation based on race, cultural categories, gender and other visible characteristics (Cross et al., 1994; Morrison, 1992), others argue for a definition enveloping all characteristics in which a team can vary (Jackson et al., 1995; Thomas & Ely, 1996). This thesis will not touch upon this academic debate and the term diversity will be used to define variation in demographic factors of individuals within a group (Van Knippenberg, De Dreu, & Homan, 2004). Demographic factors are defined as the characteristics of individuals which are considered as readily available and observable. This approach is both broad and applicable to any group and ties in with research to TMT. There are two main categories of diversity factors within a group, namely the social category and the
informational/functional category. Social diversity factors would include gender, age, and ethnicity,
while the informational/functional category would include factors such as tenure, educational background and functional background (Williams & O'Reilly, 1998).
As mentioned in the introduction, companies are looking to improve firm performance by diversifying their TMT. TMTs exists to serve the company’s goals and to increase the performance of the firm and play a key role in organizational outcomes (Umans & Smith, 2013). Venkatraman and Ramanujam (1986) state that firm performance has multiple perspectives, namely, organizational effectiveness, financial performance and operational performance. Most studies predominantly discuss financial performance which is directed towards fulfillment of economic goals (e.g. Campbell & Mínguez-Vera, 2008; Joecks, Pull, & Vetter, 2013). In these studies, accounting based measures like sales, return on assets, return on investment are often used. Operational performance indicates the non-financial measures like market share, flexibility and other measures of efficiency. Organizational effectiveness, relates to using the right tools and strategies most ‘effective’, for example, the use of human resources and focus on education and growth of employees. Most papers confine their focus to financial and operational performance of the firm (Venkatraman & Ramanujam, 1986). The charted
course of an organization is ultimately determined by the strategic and decision-making of the TMT, and thus the TMT should be a strong determinant for the firm’s performance (Huse, 2007).
2.2 Theories
2.2.1 Upper echelons theory
TMT diversity is often related to the upper echelons theory. The central idea of upper echelons theory is based on the following statement: ‘If we want to understand why organizations do the things they do,
or why they perform the way they do, we must consider the biases and dispositions of their most powerful actors—their top executives’ (Hambrick, 2007, p. 343). In the article of Hambrick and Mason (1984)
there is an emphasis on the ‘dominant coalition’ and a prime focus on ‘top managers’. Based on this, Hambrick and Mason (1984) introduced two ideas which both stimulated considerable streams of research. The first idea suggests that in order to investigate the relationship between diversity characteristics of company governance on firm performance, the analysis should not be limited to solely the supreme executive (e.g. CEO). A broader scope of evaluation is required as the strategic behavior of organizations is governed by the collective cognitions and capabilities of the entire TMT (Hambrick, 2007). The second idea that Hambrick and Mason (1984) presented, purports that demographic characteristics can be used to evaluate the TMT characteristics. Based on the organizational demography approach, they criticize the use of ‘cognitive’ constructs as attitudes, values and cognitions, since these constructs are difficult to access and measure reliably, and there is no certain method to measure this directly (Pfeffer, 1983). To approach these cognitive constructs, Wiersema and Bantel (1992) argue that demographic characteristics are a viable proxy. They debate that certain demographic characteristics are related to distinct values and beliefs and may be used to measure personalities of managers. This integral idea is echoed by several investigators in the field. Demographic factors are considered to be determinants of behaviors of an individual, which will affect decision-making and information processing in a group, which will finally affect the organizational outcomes (e.g. Groeneveld & Verbeek, 2012; Haleblian & Finkelstein, 1993; Sahu & Manna, 2013). Upper echelons theory investigates the demographics of team members and hypothesizes how that affects behavior and organizational outcomes. Furthermore, the upper echelon theory suggests that group interaction, which plays a major role in organizational outcomes, is also greatly dependent on demographic factors of its constituents. So, the upper echelons theory attempts to explain the motivations behind business strategy decision-making according to behavioral characteristics and the variation between individuals in the power groups (Hambrick, 2007). This is based upon the assumption that organizations or businesses are a reflection of the characteristics of those power groups, and that the evaluation of the TMT will give meaningful insight about the organization and its performance. Figure 1 presents a simple graphical depiction of the integral premise of the upper echelons theory.
There have been many investigations into the relationship between TMT diversity and firm performance, however, study results have shown to be inconsistent. The findings vary from a positive (e.g. Carpenter 2002) to a negative correlation (e.g. Michel and Hambrick, 1992). For example, Ruiz-Jiménez and del Mar Fuentes-Fuentes (2016) argue that gender diversity in the TMT seems to encourage an improved working climate that could positively influence firm performance, while other studies reported neutral or negative results (Kochan et al, 2003; Pelled et al., 1999). Furthermore, the believe among scholars, social scientist and laypeople that diversity will lead to synergistic knowledge and hence to greater creativity is criticized by studies that shows that diversity can cause negative effects on communication and create conflicts in groups (e.g. Chatman et al., 1998; Kochan et al., 2003). Various hypotheses have been put forward attempting to explain these discrepancies.
Firstly, variation in defining ‘dominant coalition’ could have resulted in discrepant results. Early studies considered TMT to refer to executives serving functions on the board of directors. Later studies have broadened this scope to include other management and directorial positions that can still have quite an effect on day-to-day decision-making and strategic implementation. According to Kor (2003), all inside top-level executives, business unit heads and vice presidents ought to be included in the definition of TMT. The exact definition of TMT is discussed by several scholars but there seems to be no consensus (Jackson, 1992; Carpenter, Geletkanycz, & Sanders, 2004).
Secondly, Nielsen (2010) puts forward that not all TMT diversity aspects are equally studied. Diversity aspects were evaluated in widely varying contexts using a multitude of analyses while lacking the appropriate control.
Thirdly, in most of the studies the interaction between different aspects of TMT diversity and environmental uncertainty were not accounted for. These aspects can be considered to be crucial in all TMT research and results could have been significantly impacted (Nielsen, 2010; Homberg & Bui, 2013). The organizational context of a firm should be considered when studying the impact of TMT diversity. The context is sometimes disregarded during research while it does affect the results of a study. This may explain some of the inconsistent results. Organizational contexts might include: environmental uncertainty (e.g. Homberg & Bui, 2013), organizational culture (e.g. Pelled et al., 1999), and even temporal issues. An example of the effect of organizational context is the study by Hambrick et al. (1996), in which they found empirical results for better performing heterogeneous TMT’s under high environmental uncertainty, while more homogeneous TMT’s perform better in stable contexts. Furthermore, diversity research employs distinct methods of analyses and diversity is measured
societal contexts (Jackson, Joshi, & Erhardt, 2003). It seems as if this is omitted when comparing conflicting research results.
2.2.2 Socio-psychological theories
Besides the potential reasons for discrepant research results on the relationship between TMT diversity and firm performance mentioned in 2.1.3, other considerations can be made. Many authors neglected the call of Hambrick and Mason (1984) that research of the upper echelons theory needs to draw upon literature in other fields such as sociology and psychology. More than two decades later, Hambrick (2007) repeated that there is a need to apply theories in combination with the upper echelons theory in order to back up the research. While a number of excellent reviews of upper echelons theory studies are available (e.g. Carpenter et al., 2004; Finkelstein & Hambrick, 1996), little has been done in reviewing what theories were used to supplement the upper echelons theory. To investigate this, Nielsen (2010) conducted a comprehensive review of contemporary research on TMT diversity across multiple. She reviewed articles published in the most influential management research journals (Tahai & Meyer, 1999). In most of the studies, the upper echelons theory was combined with social psychological theories, in line with the comment of Hambrick and Mason (1984). Based on the review of Nielsen (2010) and the call by Hambrick and Mason (1984), social psychological theories seem to be the most fitting theory to investigate simultaneously with upper echelons theory. The three most commonly applied socio-psychological theories, namely, the information processing theory, the similarity-attraction paradigm and the self- and social categorization theory will be utilized in this study and are discussed below
2.2.2.1 The positive view
The central ideological view of diversity, claims that diversity creates value and benefit for the team, is rooted in the study of Hoffman (1959) on heterogeneity in small teams. He suggested that diverse teams generally should have a broader range of perspectives and knowledge than homogeneous teams. This is in line with the information processing perspective which tends to focus on the benefits of diversity in terms of information sharing among team members. Information sharing within the TMT is essential for doing business. According to the information processing theory, diverse teams will have ‘improved’ information because of the variety of perspectives and problem-solving approaches, as well as diverse sources and expertise. Furthermore, diversity in teams may increase decision-making quality and enhance collective creativity required for out-of-the-box challenges (Hoffman, 1959; Triandis et al., 1965). These favorable features of TMT diversity may be influenced by ‘conflicts’ which acts as a mediating factor (Hoffman, 1959; Damon, 1991). On the other hand, diversity could cause coordination and integration problems for the team due to non-work related conflicts, these conflicts should be managed in order for the benefits of TMT diversity to manifest (Van Knippenberg et al., 2004). Where most of the studies, in the field of information processing focus on variables that are argued to directly
impact the cognitive processes of a group (e.g. expertise, educational or functional background) there is a lack of research measuring the demographic characteristics such as gender, age or ethnicity (Pitcher & Smith, 2001). We can conclude that the ‘positive’ view of information processing theory depends on managing the coordination and integration problems within a team.
2.2.2.2 The pessimistic view
The other two perspectives present a more pessimistic view of diversity. Both perspectives underline that diversity can create social division and poor social integration resulting in negative outcomes (Pfeffer, 1983). The similarity-attraction theory is based on Newcomb’s theory of social attraction. The predictions of this theory can be defined as: ‘Similarity on attributes such as attitudes, values, and beliefs will facilitate interpersonal attraction and liking, and vice versa. Liking and similarity reinforce one another and create a strain toward symmetry. People will avoid communicating with those they dislike or with those who hold opinions or views differing from their own as a means of reducing the strain produced by the disagreement (Mannix & Neale, 2005). Where the similarity-attraction theory describes the dyadic relationships within teams, it may not account for all demographic effects. The self- and social categorization theory argues that individuals can have preferences for membership of a certain group even when there has not been interaction with members of that group (Pfeffer, 1983; Tsui et al., 1992). Individuals tend to evaluate themselves and categorize themselves as to belonging to certain groups which often underly visible properties such as nationality and gender. This can cause social division within a team, which in turn could impede the performance of the team. Both perspectives offer theoretical insight into group processes. Individuals define their self-concept in particular in terms of membership of certain ‘groups’. In these theories, diversity will lead to negative social processes which result in poor outcomes for the group, while in homogeneous groups individuals will experience more cohesion (O’Reilly et al., 1989) and less relational conflict (Tsui et al., 1992). Both theories highlight problems with differences in groups which should be taken into account when investigating TMT diversity.
2.2.2.3 Summary
While there is no consensus about the importance of each of the three theoretical approaches, to understand the effect of TMT diversity on firm performance it is necessary to integrate all three approaches. The information processing approach helps us understand how diversity can result in enhanced performance through interaction and the valuable exchange of information among team members, while the other perspectives help us to explain why individuals need validation of homogeneity and the comfort of belonging which are potentially destructive in heterogeneous TMT’s (Mannix & Neale, 2005). Diverse teams will only succeed if they adequately overcome the disruptive effects of their differences.
2.3 Firm characteristics
2.3.1 Firm size
Given the fact that information processing and ‘right’ interaction within TMTs are decisive for the success of an organization, it is notable that there is a lack of research on TMT diversity in SMEs (e.g. Neill & Dulewicz, 2010). As SMEs give rise to 60 to 70% of jobs, new research is certainly warranted (OECD, 2020). As mentioned in the introduction, SMEs differ from large firms in the human and organizational processes (Koch & De Kok, 1999). Because of resource limitations, less developed administrative support systems for decision-making of SMEs and fewer external influences (e.g. capital markets, stakeholders), the capabilities of management are pivotal within the organization and room for error is small. The lack of research and the supposed stronger relationship between the TMTs of SMEs and firm performance present a promising research opportunity.
In order to give insight in the differences between a SME and a large firm, the article by Nooteboom (1994) defines three core characteristics of small firms compared to large firms, namely: small scale, personality and independence.
- Small scale: The lack of economies of scale are not only in production and management but also in marketing and transaction costs.
- Personality: There is an intertwining of private and business affairs. (e.g. using private capital) - Independence: small firms are relatively ‘free’, allowing for more idiosyncratic goals and
conduct.
Those core characteristics lead to several ‘derived characteristics’ which specify certain strengths and weaknesses of SMEs. As the advantages of SMEs are generally the disadvantages of large firms and vice versa. The advantages of SMEs and large firms are listed in Table 1.
Table 1
Differences and relative advantages of SMEs and large firms
Small and medium-sized enterprises Large firms
Little to no bureaucracy (informal) Formal management skills
Rapid decision-making & risk taking Ability controlling complex organizations Motivated labour & management More specialized labour
Effective internal communication (faster decision-making) Access to external capital & better able to fund diversification (synergy)
R&D efficiency Economies of scale and scope (R&D) Capacity for customization High market power (creating entry barriers)
Fast learning and adapting routines and strategy Comprehensive distribution and servicing (Multiple products)
Source: Nooteboom, (1994)
According to Escribá-Esteve et al. (2009) the greater administrative flexibility and participation of the TMT is a great benefit of SMEs. For instance, particular departments for R&D and marketing are less common in SMEs, and if they exist, they will rarely be influenced by the TMT (Brunninge et al., 2007).
Fewer departments in a firm may result in fewer individual interests, which in turn might reduce conflicts and improve team cohesion within the TMT. The fact that employees and managers in SMEs seem more motivated (Nooteboom, 1994) implies that conflicts are resolved as quickly as possible or avoided in advance. Given the limited resources of SMEs, we argue that the impact of the TMT on a firm’s performance is larger than in large firms. (e.g. due to lack of resources the TMT of a SME should make well considered choices in doing business while in a large firm some of these issues will be outsourced and thus their impact will decrease). This, combined with stronger team cohesion within SMEs, might increase the positive impact of TMT diversity on SMEs performance.
2.3.2 Family firms
Many of the small and medium-sized enterprises are also family firms (Brunninge et al., 2007). The mix of ‘business matters’ with concerns about the ‘family welfare’ may cause inertia in these firms, which can result in postponing important business decisions (Schulze et al., 2002). Schulze et al. (2002) argue that family ownership may hinder strategic change activities as a result of risk aversion and humanitarian motives or internal family conflicts. Further, the commitment of the family to the ‘traditional’ strategy and culture can result in unwillingness to change or avoidance of risks, while change may be necessary for the continuity of the firm (Goodstein & Boeker, 1991). As Goodstein & Boeker (1991, p. 312) state in their article: ‘‘Over time, owners may become insulated from environmental and performance changes and fail to perceive and react to critical environmental and organizational changes’’. Because, in general, the opinion or influence of owner-managers outweighs that of non-family within a TMT, the interaction could be negatively affected. In firms, it is important to create a TMT where members feel confident and safe to share valuable input. Within family firms, this can be even more important, because of the prevailing impact of owner-managers. Brunninge et al. (2007) report that ‘larger TMT size has a positive effect on ‘strategic change’ in which ‘strategic change’ can be seen as a positive outcome for TMT. The input (e.g. knowledge, experience and creativity) within a team increases as the team grows in size. In addition, more input could provide more pressure to convince the owner-manager about a certain issue and hence decrease the prevailing impact of owner-managers.
2.4 Conceptual framework and hypotheses development
2.4.1 Conceptual framework
In order to create a clear overview of the aforementioned knowledge, a conceptual framework is included (figure 2). In this framework the underlying theories of the relationships between ‘demographic characteristics’ and ‘firm performance’ are summarized. As described in chapter 2.1.3, we assume that the demographic characteristics of TMT members affect the cognitive constructs of TMT members. Those cognitive constructs converge in the team process because team members ‘interact’ with each other. By studying the relation between demographic characteristics and firm performance, the psychological and social processes within a TMT still remain largely a mystery, the so-called ‘Black box problem’ (Lawrence, 1997). This warrants further discussion, but this study will not contribute to this debate. Contextual factors play a major role in investigating the impact of diversity. Issues like environmental uncertainty and organizational culture may both have impact on a firm’s performance.
2.4.2 Hypotheses development
In this section, the conceptual model will be presented. Subsequently demographic diversity of TMT members in small and medium-sized enterprises will be discussed and lastly, supportive hypotheses will be developed.
Figure 3 – Conceptual model.
The figure presents a graphical depiction of the expected relations within this study. The conceptual model is to be read as follows: firm performance is affected by four characteristics of demographic diversity, whose effects are expected to be negatively influenced by whether it is a family firm. The relations will be further explained in the following section.
2.4.2.1 Demographic diversity
Demographic factors are based on individual biological and socioeconomic backgrounds. These often are considered visible characteristics as they manifest in the outside appearance and the most unique characteristics of a group (Forbes & Milliken 1999; Williams & O’Reilly, 1998). In this study we opted for a selection of traits in order to define diversity, namely: age, gender, nationality and tenure. Literature, as discussed in chapter 2.2, highlighted that diversity can have beneficial and detrimental effects on firm performance. The effect of diversity for all four characteristics are similar and therefore we will briefly explain each variable with its corresponding references.
1. Age diversity
The relation between age diversity in TMTs and firm performance has been investigated substantially (e.g. Ali et al., 2014; Nielsen, 2010). According to a small selection of studies, age diversity is positively correlated to firm performance (e.g. Jackson et al., 1991; Urzander & Larsson 2015) while other researchers report a negative relationship (e.g. Pegels et al., 2000; Pelled et al., 1999). Research has demonstrated that, until a certain point, older individuals have greater human and social capital acquired over their career and are more risk averse (Hambrick & Mason, 1984). Younger individuals contrarily seem to have greater capacity for innovative and strategic changes (Richard & Shelor, 2002). Age
diversity in TMTs can cause conflicts. The main reason for these conflicts is the age bias towards team members, which is in line with the self- and social categorization theory discussed earlier. Those conflicts are mostly not work related and for this reason negatively disturb TMT performance (Lau & Murnighan, 1998; Pelled et al., 1999). As we explained before, conflicts are generally not beneficial since it possibly stops the information sharing within a team or slow down the decision-making process, although conflicts are sometimes a key event to enhance problem solving. Non-work-related conflicts prove to be largely detrimental for TMT performance; these conflicts take up valuable time which is not used to the benefit of the company. Despite the possible conflicts, TMT age diversity will enhance firm performance given the fact that a diverse group of older and younger team members will perform better due to a mix of human and social capital (older) and capacity for innovative and strategic changes (younger).
2. Gender diversity
According to Dwyer et al. (2003), gender diversity in TMT may be beneficial to firms. Further, Wood (1987) suggests that men and women have qualitatively different caches of knowledge such that gender diversity within a TMT may spark creativity and innovation. While earlier TMT research report that gender diversity in a group is a direct cause of conflict which can inhibit group decision-making and therefore organizational outcomes (Pelled, 1996; Pelled et al., 1999). In line with the similarity-attraction paradigm, it is possible that subconscious categorization based on visible characteristics such as gender leads to tensions between those social groups, and in doing so can inhibit successful cooperation (Williams & O'Reilly, 1998). According to the information processing theory, gender diversity can contribute to a variety of perspectives and problem-solving approaches, as well as diverse sources and expertise within TMT. Despite the possible conflicts, TMT gender
diversity could enhance firm performance due to the fact that men and woman working together may enhance problem solving and decision making within the TMT because of the increased creativity.
3. Nationality diversity
In the field of upper echelons theory, empirical studies investigating the impact of TMT nationality are still limited (Nielsen & Nielsen, 2013). One study performed by Staples (2007) found evidence that the number of non-nationals on TMTs of large firms is increasing because of accelerated
globalization. Besides this, to our knowledge, little to no studies were performed related to SMEs and nationality diversity. Nationality diversity examines the ratio between the number of local and foreign members of TMTs, which brings different cultural heritages and experiences together. Hambrick et al. (1998, p. 187) argue that ‘a person's nationality (which may be amplified or muted by professional and
personal experiences) affects his or her values, cognitive schema, demeanor, and language. These attributes, in turn, shape the person's behavior in response to task stimuli’. National culture plays an
rooted within a person’s mind and therefore have impact on the manager’s mindset, interpretation and response to strategic issues (Geletkanycz, 1997; Schneider & DeMeyer, 1991). Furthermore, formal institutions constrain and regulate economic behavior. This is because growing up in a society with a particular configuration of (formal and informal) institutions affects information processing and problem solving of managers (Scott, 2008). Consistent with the information processing theory, TMT nationality diversity bring a wide range of experiences and knowledge of diverse institutional environments which enhance in-depth discussions and consideration of several option in decision-making (Hambrick et al., 1998). Nevertheless, the more pessimistic theories suggest that TMT nationality diversity may result in affective conflicts and slower decision-making because of the different communication patters and interaction styles (Earley & Mosakowski, 2000). Despite the possible conflicts, TMT nationality diversity will enhance firm performance given that a diverse team will have a wide range of experiences and knowledge of diverse institutional environments.
4. Tenure diversity
Tenure diversity is investigated by scholars in several contexts. It is a demographic characteristic that according to Finkelstein et al. (2009) is associated with: commitment to the status quo, risk aversion and information processing. Therefore, tenure diversity has implications for the length of time the TMT members have been on a position. This could affect decision-making and exploration activities within the TMT. Long term TMT members focus on the status quo and are more risk averse (exploitation), while relatively new TMT members are expected to focus on innovative and new projects (exploration) (McClelland et al., 2012). Studies found both positive (e.g. West & Anderson, 1996), and negative (e.g. Finkelstein & Hambrick, 1990) effects of TMT tenure diversity on firm performance. In line with the information processing theory, we can hypothesize that tenure diversity could have positive impact of a firm’s performance due to balance in focus on exploitation and exploration activities and insights. Nevertheless, according to the pessimistic theories, these differences can negatively influence the decision-making process due to conflicts among the ‘older’ and ‘younger’ members. Carpenter (2002) argued that short-tenured TMTs perform best but it is only temporary. Because of ‘informal effective communication’ and ‘fast learning and adapting routines’ behavioral integration will come about even sooner, which could diminish the effect of TMT tenure diversity. Despite the possible conflicts, TMT tenure diversity will enhance firm performance given that a diverse team will have a certain balance in focus on exploitation and exploration activities and insights.
2.4.2.2 Hypotheses
As with all four characteristics above, both positive and negative effects of diversity are discussed. While the bulk of research within the field of upper echelons theory investigated the impact of demographic diversity on firm performance by studying large firms, this study focusses on SMEs. Literature, as discussed in chapter 2.3.1, highlighted that certain features of SMEs might reduce conflicts and improve team cohesion within the TMT (e.g. motivated managers, more idiosyncratic goals). TMT diversity induced conflicts that arise in SMEs are fewer and more easily resolved than those in large firms. Based on these differences, we expect the positive effects of diversity within this study. Given that previous research found significant effects of TMT diversity on firm performance and the positive differences of SMEs relative to large firms, we expect that demographic diversity within TMTs of SME’s will have positive impact on the performance of a firm. All four characteristics might be taken to mark variety, in which variety broadens the cognitive and behavioral repertoire of the team (Harrison & Klein, 2007). In line with the information processing theory, it is expected that demographic diversity within the TMT will enhance the ‘information’, because of the variety of perspectives and problem-solving approaches, as well as diverse sources and expertise.
Therefore, we hypothesize:
H1: TMT age diversity is positively correlated to firm performance of SME’s H2: TMT gender diversity is positively correlated to firm performance of SME’s. H3: TMT nationality diversity is positive correlated to firm performance of SME’s H4: TMT tenure diversity is positive correlated to firm performance of SME’s
Based on theories discussed earlier, we argue above that TMT diversity will have a positive impact on SME’s performance. An important aspect which we discussed while reviewing literature of TMT diversity is the ownership of the firm, and in particular whether it is a family firm (Brunninge et al., 2007). As stated, scholars argue that within family firms, decision making and openness to change differ compared to non-family firms (Gomez-Meija et al., 2014). The goals of owners are mostly related to the survival of the firm (exploitation), while other TMT members prefer more explorative activities (exploration) (Block & Spiegel, 2013). So far, few studies on TMT diversity have been concerned with family firms and therefore investigating the interaction effect is relevant. This ‘family’ influence will disrupt the process within TMT’s negatively and thus weakens the effect of diversity on firm performance. Vice-versa, the environment presented by non-family firms could strengthen the positive effect of diversity on firm performance due to the fact that there is no influence which interrupt the interaction within the TMT. Given that the opinion or influence of owner-managers outweighs that of non-family within a TMT, the interaction could be negatively affected. This impact of family firms negatively affects the enhanced ‘information’ within a TMT as explained by hypotheses 1 till 4.
Therefore, we expect that ‘family firm’ will have a negative impact on the relation between TMT diversity and firm performance.
Thus, we hypothesize:
H5a: The positive effect of TMT age diversity on firm performance is weakened when the firm is a family firm.
H5b: The positive effect of TMT gender diversity on firm performance is weakened when the firm is a family firm.
H5c: The positive effect of TMT nationality diversity on firm performance is weakened when the firm is a family firm.
H5d: The positive effect of TMT tenure diversity on firm performance is weakened when the firm is a family firm.
3. Methodology
This chapter describes the quantitative methods used to model the effects of TMT diversity on firm performance. The data used in this study is an aggregate of data by a similar study and newly collected data for this investigation. Firstly, the data collection and aggregation procedure is clarified. Secondly, the candidate variables for the model are defined. The chapter ends with a statistical analysis plan.
3.1 Data collection
3.1.1 Target population
For measuring the effect of TMT diversity within an SME, the firm should have a certain size of employees, which was set at a minimum of 50 but not more than 250 persons and required a governance structure with at least two members in the TMT. The minimum of 50 employees is used to omit small companies with one director or without a governance structure. Furthermore, the company must be independent in making ‘strategic’ decisions, for example not limited by regulations or a parent company. The EU defines SMEs as: enterprises which employ < 250 individuals and which have a turnover not exceeding €50 million per year, or a balance sheet total not exceeding €43 million per year (EU Europa, 2020). See table 2.
Table 2
EU norms for SME
Category Num. of employees Turnover (Annual) Total Assets Medium < 250 < € 50.000.000 < € 43.000.000
Small < 50 < € 10.000.000 < €10.000.000 Micro < 10 < € 2.000.000 < € 2.000.000
Source: EU Europa, (2020)
3.1.2 Aggregated data
The data used in this study was gathered via primary (survey) and secondary (Orbis) sources and contains data gathered over several years, namely 2013, 2018 & 2020. In 2013 & 2018, groups of students from the University of Groningen conducted research into the composition of TMTs within Dutch SME’s on behalf of dr. M.J. Brand. This data has been released by the University of Groningen for this thesis. Initially, the data from both years seemed appropriate for this study, but the data of 2013 proved inadequate for this study because of missing data related to firm performance and tenure diversity. This study attempts to ensure consistency in data collection, therefore the survey questions and search strategy of 2018 were used as a basis for this study to ensure the reliability and validity of the data. See Appendices 1.0 and 1.1 for an overview of the data of 2013 and 2018.
3.1.3 Data 2020
The data contains responses from Dutch firms in every sector except for the banking and insurance, as these companies are not able to make independent decisions. Banking and insurance companies are required to adhere to strict laws and regulations to ensure safety and reliability, limiting the independence of the TMT. Candidate firms included in this investigation were identified through the online database Orbis and were obtained on November 27th of 2020 (See appendix 1.2). Orbis, provided by bureau van Dijk, draws information from firm’s registrations from the Kamer van Koophandel (Dutch Chamber of Commerce). The initial sample included 310 firms, selected by following the criteria of the EU norms and criteria used in 2013 & 2018. See appendices 1.0, 1.1 & table 3.
Table 3
Search strategy Orbis 2020
The criteria ‘companies with a phone number’ made the sample drop from 315 to 310 and was added because of the convenience while running the questionnaires. The criteria ‘companies with an email address’ was omitted because it left a sample of only 28 companies.
3.1.4 Survey
For collecting data in research studies, mail surveys are one of the most used methods (Bourque & Fielder, 2003). According to Bourque and Fielder (2003), response rates of 10-20% are common for mail surveys, we see that almost two decades later this is much lower, namely around 4% (e.g. Graafland, 2020). In 2018 the survey was conducted by phone with a response rate of 32%. Given that the telephone approach had a higher response rate comparing to the mail survey in 2013 and the high response rate of 2018, the telephone approach is used in this study. Another reason for conducting the survey by phone is timing. This study took place at the time of the corona crisis (Covid-19). It was expected that many companies had other priorities than to participate in a survey. Thus, the response rate would possibly be even lower. This was also confirmed later during the phone calls; many companies did not participate in the study because of other priorities. In order to exclude the influence of Covid-19 on the firms from consideration within this study, respondents were asked to answer the questions as if it were January 1th 2020. A survey (see Appendix 2.0) was developed which was largely based on the survey conducted in 2018, most questions were copied and rewritten. Four additional
- Dutch Company
- Number of employees: Minimum of employees = 50 & Maximum of employees = 249
- National legal form: Private limited liability company (BV) & Public limited liability company (NV) - Exclusion of companies operating in financial services, insurance & public administration and defense. - Exclusion of subsidiaries with foreign shareholders with ownership of 51% +
- Maximum operating revenue of €50.000.000 - Maximum total assets of €43.000.000 - Companies with a phone number
questions were added, these related to the sector, the overall diversity of the firm, performance and impact of Covid-19 on the firm. In order to improve the response rate, the survey was designed to ensure that an executive secretary or employee of the human resource department was able to answer the questions. This is in contrast to the surveys in 2013 and 2018, where the survey focused only on a member of the TMT. Language has been simplified and the explanation and the explanatory notes to the survey were improved. When the respondent was unable to answer the questions, the investigator requested a member of the TMT to continue the survey. Before conducting the research, several experts1 have been asked to review and test the survey. Qualtrics Survey Software was used to document survey answers during the calling of respondents and to email the survey to the respondent if they preferred to answer by email.
3.1.5 Collection procedure
To ensure the validity of the study, a telephone call script and a template for sending e-mails were established to make sure that every participant was introduced and informed in the same manner (see appendices 2.1 & 2.2). More than 400 calls were made across 192 firms between 2th and 10th of December 2020. All communication was conducted during business days and hours. The standardization in the process reduces the possibility of reliability biases and process variation. Because of the corona crisis, it was government policy to work from home as much as possible. Therefore, most of the employees worked at home and the internal communication (transferring calls) did not go smoothly. That may explain why there were so many callback requests. See table 4 for a general overview of the collection procedure. There were 22 companies (11,5%) that were able/willing to complete a survey during the first call. Because of the aggregated data we were able to call firms that participated in the earlier studies. 23 firms were called that participated in both 2013 and 2018 and another 65 firms were called that participated only in 2018. Due to a lack of time, not all the firms of the 2020 sample (N=310) were called. The 104 firms called from the sample of 2020 are randomly chosen by using the system ‘case number +4’. The three most given reasons by firms for not participating to this study are: The firm
has other priorities or no time at all, sending a mail was not necessary (n=34), because of privacy issues (n=12) and the company never participate in studies (n=8). In total there were 61 respondents, of which
4 were deleted because they were insufficiently completed. Ultimately, the 2020 sample consisted of 57 completed surveys, resulting in a gross response rate of 29,7%. This was combined with the sample of 2018 (N=88) to form a final sample size of 145 completed company surveys. As you can see in table 4, 38 firms participated only in 2020 which means that 66,7 % of the 2020 sample were new firms.
Table 4
Overview collection procedure
Total firms called: 192 Final sample: 145
Participated in 2013, 2018 & called in 2020 23 Participated in 2013, 2018 and 2020 6 Participated in 2018 & called 2020 65 Participated in 2018 and 2020 13 Firm called in 2020 (new firms) 104 Participated in 2020 38
Firms mailed after phone call 54 Participated in 2018 (data 2018) 88
Firms participated at first call 22
Firms called > 2 times 170
Respondents by phone 2020 53
Respondents by mail 2020 8
All information of cases is included (also 2013) to give a complete overview.
3.2 Definition and measurement of variables
3.2.1 Top Management Team
To investigate TMT issues within the context of Dutch SMEs, the legal governance requirements of the Netherlands first has to be considered. In the Netherlands both an one- and two-tier board is entitled. The two-tier model separates executive director’s management tasks from the supervisory directors monitoring tasks (Van Gils, 2005). For this study, the executive directors are considered TMT because they are the most powerful actors and top executives within the firm. Thus, firms with a two-tier board were asked about both board of directors and supervisory board. Information of the board of directors is used for thus study, this is in line with Hambrick (2007).
To measure whether a TMT operates independent and is not limited by a parent company, the following questions are added to the survey. Firstly, the firm is asked whether the firm is a subsidiary of a company with less or more than 250 employees. (0: no, 1: yes with < 250 employees and 2: yes with > 250 employees), when it turned out to be a subsidiary with over 250 employees we asked: 'to what extent does the parent company have an influence on the operation of the business?’. This question was measured on a 5 point Likert scales (1: very small, 5: very strong). Firms that answered 5 (very strong effect) and 4 (strong effect) were not included in this study.
3.2.2 Dependent variable
Firm Performance
.
This study uses a measure of organizational growth as a proxy of firm performance. Growth measures intend to reflect the growth of organizations and are applied by researchers since organizational growth is considered a major driver of the economic value of a firm (Brealey et al., 2012; Carton & Hofer, 2006). Given that there was no other data available for measuring performance, we followed Frank et al. (2012) by using the number of employees as performance measure to indicate firm growth. Sales and the number of employees are the mostincrease in the number of employees represents an appropriate growth measure given the fact that employees are the key elements in conducting the firm activities which affects firm performance. Furthermore, the proxy within this study seems more accessible to ask a firm because the ‘financial’ results are not known by the outside world. SMEs are not required to share their results, for example in public annual reports.The survey included two questions about firm performance. One question was about whether there was an increase of employees within the firm over the past 5 years. (FirmPerf._Growth) This question was both in 2018 as 2020 part of the survey (0: decreased or remained the same, 1: increased). In the survey of 2020 a second question was added following Dess and Robinson (1984), it concerns a self-reported performance measure measured with a three-item scale: ‘how does the firm perform in comparison to your direct competitors?’ (FirmPerf._Competitors, 0: worse, 1:similar and 2:better). Due to the large heterogeneity among SMEs, these measures will better reflect an increase in performance for SME’s than the well-known financial proxies as ROA and profit. According to Vij and Bedi (2016) these objective financial measures are losing their
significance because these are static and mainly internal focused. Besides, they take no account of competitors and customers. For this study, the variable FirmPerf.Growth is used as dependent
variable. Given the answers of 2020 on both questions differ substantially (see table 12, Appendix 3.0) and FirmPerf.Competitors is possibly biased due to the subjective measure. Furthermore,
FirmPerf.Growth is applicable to all cases for both 2018 and 2020.
3.2.3 Independent variables
The independent variables within this study are age, gender, nationality and tenure diversity of the TMTs. According to Nielsen (2010), most studies in the field on group diversity research apply a single item measure for each diversity dimension. The Blau index (Blau, 1977) is often used for categorical constructs such as gender, nationality, education, etc. The index is widely used in ecology, genetics, linguistics and economics and quantifies the diversity of a group with regard to nominal features. The Blau index is measured as follows: " = 1 − ) &'(
'*+ = 1 − (&-.&+)(− (&-.&…)(− (&-.&))(. Where H is the Blau index, and p is the proportion of TMT members in k categories. For instance, the variable gender has two categories (k = 2). If the proportion of women is 0.70 and the proportion of men is 0.30, then Blau index is 1 – 0.302 – 0.702 = 0.42. The higher the value of H for a specific variable, the greater the degree of diversity in the TMT for that variable. The maximum value of the Blau index depends on the amount of k categories. The maximum Blau index for a variable is calculated as: (k – 1)/k, where k again refers to the number of categories (Biemann & Kearney, 2010). However, all categorical independent variables in this study have two categories, which have a maximum Blau index of 0.50. Maximum score for diversity means equal distribution for each category.
For continuous variables, such as age, the coefficient of variation is most commonly used. Both measures are used in this research and are accepted as a norm in the field (Nielsen, 2010). In addition, proportions and dummy variables were used to analyze the effect of diversity (table 5). Proportions range between 0 and 1. In general, the further the number deviates from 0.50, the less diverse the team. Table 5 gives an overview of which measures are used for the independent variables, explanation can be found below the table.
Table 5
Summary independent variables Coefficient
of variation
Proportion category Dummy - mixed teams Blau index
Age diversity √ √
Gender diversity % female | % male (0) only men or women | (1) both sexes √ Nationality diversity % non-Dutch | % Dutch (0) 1 nationality | (1) >1 nationality √ Tenure diversity % shorter | % longer (0) not mixed | (1) mixed team √ Previous function % no | % yes (0) not mixed | (1) mixed team √ First of all, for the variables of gender, nationality and tenure, heterogeneity across the TMT was calculated by the Blau Index (Blau, 1977). Resulting in the following variables:
- Blau_Index_Gender - Blau_Index_Nationality - Blau_Index_Tenure - Blau_Index_PreFunc
Age diversity within the TMT. The age (measured in years) of every team member was asked in both
the survey of 2018 and 2020. In order to measure TMT age diversity, we follow Ali, Ng and Kulik (2014) by using the coefficient of variation to measure age diversity. It is calculated by dividing the st. deviation of the TMT member’s age by the mean of their age (Age_Variation). Age is the only continuous variable in this study. In addition, the median age of the team is also measured (Age_Median). The median is often used as a measure of central tendency because it is less effected by outliers and useful for measuring spread with quartiles what will be used for the analyses.
Gender diversity within the TMT. The gender (0: female, 1: male) of every team member was asked in
both the survey of 2018 and 2020. Having this data makes it possible to calculate the proportion of gender within the firm’s TMT (Prop_Female & Prop_Male). Based on the study of Joecks et al. (2013), this research used a dummy for a mixed team (0: only men or women, 1: both sexes).
Nationality diversity within the TMT. The nationality of every team member was asked (0: non-Dutch,
1: Dutch). The proportion ‘Dutch’ members is measured (Prop_Dutch). The maximum value of Prop_Dutch is 1, indicating that the team consists completely of Dutch members and is therefore not
diverse. In this case, a Prop_Dutch closer to 0 indicates increased diversity, as the proportion of non-Dutch members of multiple nationalities increases. A dummy (Divers_Team_Nationality) is used for measuring whether a team is mixed (0: only Dutch or Dutch members, 1: Both Dutch and non-Dutch members).
Tenure diversity within the TMT. In order to measure the independent variable TMT tenure diversity,
the survey has two questions about tenure. One question was whether the team member had been appointed before or after a certain date. The date differs between the surveys of 2018 and 2020, but the essence of the question was whether the team member had been appointed for more or less than two years (0: shorter, 1: longer). The second question was if the member has had a previous function at the firm before being appointed to the TMT (0: no, 1: yes). For both questions, the proportions were measured (Prop_Previous_Function & Prop_Before_2014_2018).
Total diversity within the TMT. In order to measure the degree of diversity within the TMT the
following variable is created (Diversity_Score). The variable is a composite summed variable of 4 independent diversity measures with a score range of 0 to 4, it involves: gender, nationality, appointment time and previous function. Where the 4 independent variables receive a score of 1 when there is some degree of diversity (mixed teams) and 0 when there is not (e.g. only men or women). Firms with a score of 0 have no diversity at all, where firms with a score of 4 has maximal diversity.
In order to measure the ‘overall Blau index diversity’ a variable is created which is a summed variable of the means of the Blau index variables above (Blau_Index_Overall).
3.2.4 Moderator
Family firm. The survey of 2018 only contains the question whether the firm is a family firm or not,
this question is also asked in the 2020 survey (0: no family firm, 1: family firm). Additionally, in the 2020 survey the following question was added: ‘To what extent does the family affect business operations?’. This question was measured on a 5 point Likert scales (1: very small, 5: very strong). The question was added in order to increase the reliability of the interaction effect, so distinguishing the ‘effect’ from just the label family firm. Firms that answered 1 (very small family effect) and 2 (small family effect) were indicated as non-family firms, the remaining firms are family firms.
3.2.5 Control variables
In order to increase the accuracy, robustness and reliability, this research includes three control variables: sector, firm size, TMT size and supervisory board. These control variables may neutralize disturbing factors. The disturbing factor may have an effect on the quality and strength of the relation between TMT diversity and firm performance.
Sector. According to Grosvold et al. (2007), industry sector plays an important role in the
decision-making process regarding the composition of the TMT. The survey included a question about the main activity of the (0: services, 1: production).
Firm size. Given the numerous prior studies which add firm size according to Ruiz-Jiménez & Mar
Fuentes-Fuentes (2016), the survey included a question about the number of employees. The natural logarithm of the number of employees is applied to correct for outliers.
TMT size. In the field of upper echelons theory, most of the studies focuses on TMT size as a control
variable (i.e. Marimuthu & Kolandaisamy, 2009). TMT size is measured as the number of members within a team.
Supervisory board. Because there are several Dutch SME’s with supervisory boards even though it is
not imposed by legislation (van Ees and Postma, 2004), we measure supervisory board as control variable.
3.3. Analysis plan
To test the hypotheses stated in this study, i.e. the influence of TMT diversity on firm growth, correlation and regression analyses were performed. Because the dependent variable firm growth is dichotomous (0 = decreased/stable, 1 = increased), binary logistic regression was used. In addition, descriptive statistics were analyzed for all variables collected in this study.
First, normal distribution of all continuous variables was assessed visually using standard histograms and Q-Q plots and by using Shapiro-Wilk tests for normality. Data are reported as mean ± standard deviation for continuous, normally distributed data and median with interquartile range for continuous non-normally distributed data. Categorical data are presented as number (percentage).
Correlations between continuous variables were determined using Pearson correlation coefficient. To measure the strength and direction of the associations between continuous variables and dichotomous variables point-biserial correlations were used.
All hypotheses stated in this study were tested using logistic regression models with “firm growth” as dependent variable. This approach consisted of four consecutive steps. First, a multivariable logistic model with all control variables was performed. Second, a model with all control variables and independent variables was constructed. The third step consisted of the same model in step two, with the addition of the moderator variable “family firm”. Fourth and finally, a model with all control variables, independent variables, moderator variable and the interaction terms of each independent variable with the moderator was constructed. This final step was performed to test the moderation hypothesis 5 and its sub-hypotheses. In case of significant moderation, marginal effects were calculated and visualized using a marginal effects plot. Model fits of different multivariable logistic models were compared based on log likelihood, Chi-square values and Nagelkerke’s R2 value. The effect of independent variables on
This study uses different measures of diversity of categorical variables within the TMT, for example the dummy variable “mixed team gender” as well as Blau index gender. Because these variables are different ways to represent diversity of the same factor, they are highly correlated. In order to prevent multicollinearity within the multivariable models, the primary diversity measure of independent variables gender, nationality, previous function and tenure was Blau index. The same regression models with mixed team variables were added to the appendices.
In order to verify the validity of the logistic models, model assumptions were checked. One of the assumptions of logistic regression is that the relationship between the logit of the dependent variable and continuous independent variables should be linear. This was checked using Box-Tidwell tests, in which the interaction term of each continuous variable with its logarithmic transformation was added to the model. In case this was non-significant, the linearity assumption was not violated. In addition, there should not be collinearity between independent variables within the model, as multicollinearity can decrease the validity of the model. Multicollinearity was checked using variance inflated factor (VIF), using an iterative process, in which every independent variable in the final model was checked for collinearity with other variables in the model. Finally, regression requires that all observations should be independent from each other, i.e. data collected from the same firm at multiple time points should not be included in the model. Thus, data of 2020 was used of firms with data both in 2018 and 2020.
4. Results
This chapter includes the findings of this research and consists of two sections. Section 1, includes the descriptive statistics and correlations. In the second section, the results of testing the hypotheses is discussed.
4.1 Descriptives
Of the 145 firms, a number of firms were included both in 2018 and 2020 (n = 19), in which case the data of 2020 was used in order to meet the assumption of independent observations for logistics regression. In addition, twenty firms are excluded from analysis because the number of employees exceeded 250 and therefore did not count as small and medium-sized enterprises. Finally, 22 firms are excluded because the team could not operate independently or due to missing data about team members (n = 4). In total, the sample consists of 90 firms of which 49 participated in 2020 (net response rate 25.5%).
The number of employees of these 90 firms ranged from 7 to 240 employees, with an average of 119 employees. Most firms (94%) belonged to the category of medium enterprises. Also, most firms (57%) were in the services sector as opposed to the production sector. Furthermore, the majority of firms (62%) were family firms. If we then look at the characteristics of the different TMT’s, we found that the average TMT consists of three members. The mean age of the TMT members was 51 years (S.D. 6 years) and in general, the variation of age within the TMT is low, with a mean coefficient of variation of 12% (S.D. 8%). The mean percentage of women in the TMT was 16% (S.D. 22%) and on average, most TMT members were Dutch (86% [S.D. 31%]). Roughly, on average about half of TMT members had a previous function within the same firm (54%), and most team members have been part of the TMT for two years or more (62%).
Characteristics of firms with growth vs. no growth
Because we are interested in the dependent variable ‘firm growth’, characteristics of firms with increased growth and decreased or no growth (see table 14, Appendix 3.0) were compared. Firms with increased growth are larger (median firm size 117 vs. 93; p < .01) and more likely to operate in the services sector (64% vs. 41%; p = .05). In addition, the median age in the team of firms with increased performance is higher (53 vs. 48 years; p = .01) and the median proportion of directors with tenure over two years is higher (1.00 vs. 0.42; p < .01). Other variables are not significantly different, i.e. gender diversity (Blau index, proportion female), nationality diversity and overall diversity are similar. The vast majority of firms with increased growth has 100% Dutch members.
