(I ): 16,366 W - : . M W : . M B C 2019 S E B J U G F S3539636 J -H B F G ?” D P S - M - F “ T O G P : W E A O G ’ I P M ’ T

E

A

O

G

’

I

P

“T

O

G

P

:

W

D

P

S

-

M

-

F

F

G

?”

BY

J

-H

B

S3539636

U

G

F

E

B

J

2019

S

:

.

M

B

C

-

:

.

M

W

A

I would first like to thank my thesis advisor Dr. Maryse Brand of the Faculty of Economics and Business at the Rijksuniversiteit Groningen. The door to Prof. Brand’s office was always open whenever I ran into difficulties or had a question about my research or writing. She consistently steered me in the right direction whenever she thought I needed it.

I would also like to acknowledge Dr. Michael Wyrwich as the second reader of this thesis as well as my fellow students for their valuable hints and comments. I am gratefully indebted for their time and effort that they committed to continuously improve this thesis.

Finally, last but by no means least, I express my very profound gratitude to my family Anke, Johann-Heinrich, and Jonas as well as to my fiancée Lea for providing me with unfailing support and continuous encouragement throughout my study and through the process of researching and writing this thesis. This accomplishment would not have been possible without them. Thank you.

Jan-Hendrik Buhrfeind

A

Based on company growth theory and the literature about firm professionalization, this study investigates the outcome of growth – professionalization relationship in small- and medium-sized family firms in Germany. Motivated by the important but often neglected impact of outcome of growth and the still unclear professionalization construct, this research contributes to the existing literature by identifying a driver and proposing a new dimension of firm professionalization. The regression analysis of 58 survey responses from CEOs in Germany suggests a positive and significant influence of environmental complexity as an outcome of growth on firm professionalization whereas organizational complexity does not exert a significant effect. A larger customer base, more regulation, and more suppliers – summarized as environmental complexity - positively influence firm professionalization. Moreover, CEO’s need for achievement and unwillingness to change did not exert a significant influence on firm professionalization. Furthermore, the survey answers provided a better understanding of firm professionalization leading to a potential new dimension called customer-orientation. Firms are advised to take advantage of the arising environmental complexity and emphasize the professionalization process in order to cope with an advanced firm environment.

L

CEO Chief Executive Officer DoA Decentralization of authority EC (b) Environmental complexity (broad) EC (n) Environmental complexity (narrow) EU European Union

FCS Financial control systems FSM Formal strategy-making GS Governance systems

HRCS Human resource control systems

IHK Industrie- und Handelskammer (German Chambers of Commerce and Industry) KMO Kaiser-Meyer-Olkin

MIT Mittelstands- und Wirtschaftsvereinigung der CDU/CSU (SME and Economic

Association of CDU and CSU (MIT)

NfA Need for achievement OC Organizational complexity P-P plot Probability plot

PCA Principal component analysis R&D Research & Development RBV Resource-based view RQ Research question SD Standard deviation

L

&

LIST OF TABLES

Table I. Definition of small- and medium-sized firms 11

Table II. Comparison of selected characteristics between population, sample, and respondents 26

Table III. Descriptive statistics 29

Table IV. Industry according to the family firm’s main activity 29

Table V. Distribution of the family firms across Germany 30

Table VI. Results of the varimax-rotated factor model 32

Table VII. Regression analysis on professionalization 34

Table VIII. The different perceptions of professionalization 36

Table IX. Correlations 62

LIST OF FIGURES

Figure I. Conceptual model of the proposed relationships between the outcome of growth, CEO’s

T

C

1 Introduction ... 6

2 Literature Review ... 9

2.1 The Definition of Small- and Medium-sized Family Firms ... 9

2.2 The Multidimensional Approach Towards Firm Professionalization ... 11

2.2.1 Financial and human resource control systems. ... 13

2.2.2 Monitoring and resource providing governance systems. ... 13

2.2.3 Decentralization of authority. ... 14

2.2.4 Formal strategy-making. ... 15

2.3 The Outcome of Growth ... 16

2.3.1 Organizational complexity. ... 16

2.3.2 Environmental complexity. ... 17

2.4 The CEO’s Personality Traits: Unwillingness & Reluctance and Need for Achievement .... 18

2.4.1 Unwillingness & reluctance. ... 19

2.4.2 Need for achievement. ... 20

3 Methodology ... 22

3.1 Data Collection and Samples ... 22

3.2 Measurements ... 23

3.2.1 Organizational complexity. ... 23

3.2.2 Environmental complexity. ... 24

3.2.3 Need for achievement and unwillingness & reluctance. ... 24

3.2.4 Professionalization. ... 25

3.2.5 Control variables. ... 25

3.3 Validity and Reliability ... 26

3.3.1 Representativeness. ... 26

3.3.2 Key informant approach and retrospective bias. ... 27

3.3.3 Non-response bias. ... 27

3.3.4 Common method bias. ... 28

4 Analysis ... 28

4.1 Descriptive Variables ... 28

4.2 Factor Analysis and Construct Composition ... 30

4.3 Regression Analysis ... 33

4.4 Regression Results ... 34

4.5 Analysis of the Meaning of Professionalization ... 35

5 Discussion ... 37

5.1 Limitations and Future Research ... 39

5.2 Managerial Implications ... 40

5.3 Conclusion ... 41

References ... 41

Appendices ... 52

Appendix I: Questionnaire Overview ... 52

Appendix II: Variable Overview ... 57

Appendix III: Correlation Matrix ... 62

1

I

The economic importance of family firms has been widely acknowledged, but especially in Germany (Gottschalk et al., 2014; Porta et al., 1999). According to Klein (2010), family firms account for about 60 % of all firms and provide about 58 % of private employment in Germany. Mostly, family firms can be assessed as small- or medium-sized (Donckels & Fröhlich, 1991; Corbetta & Montemerlo, 1999). Family firms distinguish themselves through the intertwined relationship with the founding family (Gersick et al., 1997). Considering growth, Wiklund et al. (2003) point out that family firms often differ from non-family firms in terms of the willingness to grow and the firm’s ability to control growth. They suggest that the expected outcomes are often noneconomic (e.g. maintaining control over the firm) and thus vary strongly to maximization of financial gains, as suggested by basic economic theory. Therefore, one can suggest that the outcomes of growth affect family firms differently compared to their non-family counterparts.

General organizational development models (e.g. life cycle model) describe in predetermined stages the development of firms in general and thereby also of family firms (Gabrielsson, 2007; Gedajlovic et al., 2004). Analytical models often consider growth as an outcome, it therefore being the dependent variable, with a variety of independent influence factors, ranging from geographical location to entrepreneurial characteristics (Gilbert et al., 2006). On the other hand, growth leads to organizational issues within the firm (Fombrun & Wally, 1989). Penrose (1959) states that the main constraint on a firm’s growth is a managerial one. In her point of view, increasing size forces the firm to undergo fundamental changes with respect to managerial function and basic administrative structure, which subsequently fundamentally shape the very nature of the organization. McKelvie & Wiklund (2010) refer to this as the outcome of growth, which is particularly related to potential managerial challenges that arise when a firm increases in size. It is not sufficient (or even not possible) to merely add resources to deal with its growth (Flamholtz & Randle, 2012); rather the firm must professionalize in order to cope with the consequences of growth. Lewis & Churchill (1983) refer to this step as a transformation the firm should undergo. Flamholtz & Randle (2012) describe the recognition of this need as the experience of “growing pains” (p. 48), due to misbalanced organizational development and corresponding internal complexity. Researchers have investigated such managerial issues in differently sized and growing firms (Flamholtz & Randle, 2012; Fombrun & Wally, 1989; Hanks et al., 1994; Kazanjian, 1988; Kazanjian & Drazin, 1990). Still, the literature misses an empirical link that the outcomes of growth drive family firms towards the next development stage – to professionalize.

contribute (Gedajlovic & Carney, 2010; Schulze et al., 2001). Several authors criticized this unidimensional approach and proposed a multidimensional perspective on family firm professionalization (Dekker, 2012; Dekker et al., 2013; Dekker et al., 2015). The authors consider firm professionalization as the implementation of formalized systems and structures. This view offers a broader and more inclusive perspective on firm professionalization and its various facets. Nevertheless, the explanatory power of it is rather weak (Dekker et al., 2015). Thus, our understanding of firm professionalization and the firm’s transformation process requires further research.

By linking the professionalization literature to the growth literature, this study aims at identify and empirically test drivers of firm professionalization and enhance our understanding of the meaning of firm professionalization. So far, professionalization is often treated as an independent variable and its influence on performance is investigated (cf. Dekker et al., 2015). This study takes firm professionalization as the dependent variable and investigates its drivers. Furthermore, this empirical study acknowledges the important role CEOs have in firms in general, and in small- and medium-sized family firms in particular, by incorporating distinctive characteristics of the CEO that may influence the willingness and intensity of professionalization within the firm.

The motivation for this research is threefold. First, small- and medium-sized family firms account for a majority of firms in a country’s business landscape. This is especially true in Germany where family firms are deeply rooted in the overall economic environment. By considering the major firm type, it seems like a promising avenue to broaden the knowledge about family firm professionalization, and also to analyze what drives these firms to professionalization and what firm professionalization is in this context. Second, growth literature mostly focuses on growth as an outcome and often neglects the outcome of growth (McKelvie & Wiklund, 2010). To my knowledge, no existing study links the outcomes of growth with family firm professionalization. Therefore, addressing this gap and identifying what drives firms towards professionalization is a major motivation for this study. Third, CEOs of family firms often have different priorities in comparison to their counterparts in non-family firms. The different priorities can be traced back to certain characteristics that, again, might influence the professionalization process in firms. The distinct behavior shown by family firms and the specific characteristics of small- and medium-sized firms emphasize the importance of investigating these firm types further. This study tries to shed a little light on the complex relationship between individual characteristics and firm behavior, in order to explain firm-level decisions and incite researchers to focus more on this relationship in the future.

efficiently organized structure with an effective use of resources. Additionally, awareness about the influence of the owner’s characteristics on firm behavior cannot be stressed enough, especially in the family firm context. For instance, the unwillingness to share decision-making power may lead to organizational issues in a larger firm. Awareness may attenuate the influence of this personality attitude.

Based on the linkage of two streams of literature, I have formulated four research questions: One main question and three subquestions. Firstly, I did not find an empirical investigation of the outcomes of growth as drivers of family firm professionalization in the literature. The company growth theory provides an idea of the relationship but the empirical proof of firm professionalization’s drivers is still vague. Following this theory, to manage an increasingly complex firm calls for better resource allocation, a more effective use of organizational structures, and service-providing/firm-monitoring institutions to cope with this complexity – in essence the need to professionalize. This leads to my main research question:

RQ1: What are drivers of professionalization for small- and medium-sized family firms?

In order to answer the rather generic main research question, I formulated three subquestions on a lower abstraction level. Relying on literature research, I chose to investigate the outcomes of growth and individual CEO characteristics as potential drivers of firm professionalization. Furthermore, I want to validate and perhaps enhance the conceptualization of the professionalization construct by asking practitioners about their understanding of firm professionalization.

The outcome of growth is often associated with complexity, but it is not clear in what firm areas complexity starts and where it ends. In order to answer RQ1, we first need to determine if complexity indeed arises as an outcome of growth. I formulate my second research question as a subquestion of RQ1:

RQ1A: What is the “outcome of growth” in small- and medium-sized family firms?

In the same vein, it must be clear what professionalization means. The professionalization construct is ambiguously discussed in the literature (see section 2.2) and no clear idea of what professionalization means for firms has yet emerged. Building on the work of Dekker (2012), this study aims to develop the dimensions that might constitute professionalization according to existing theories. Again, I formulate my third research question as a subquestion of RQ1:

Fourthly, referring to the CEO’s personality, their influence on professionalization and the CEO’s particular role in small- and medium-sized family firms – which are subject of investigation in this study – my last subquestion is:

RQ1C: Do the CEO’s personality traits influence the professionalization on firm-level?

The paper continues as follows: First, the literature review will clarify important terms, definitions, and concepts used in this study. Moreover, hypotheses are derived building on existing theory using two different literature streams. Second, the methodology section describes the data collection process, sample creation, and the operationalization of the variables. Furthermore, a discussion about the validity and reliability is added. Third, I present the analysis of the data and the results. It includes relevant statistical parameters such as correlations, regression weights and test values. Next, I discuss important findings and interpret them by using the theory. Eventually, the discussion section offers limitations and future research as well as managerial implications derived from this study. The overall conclusion completes this paper.

2

L

R

This section first elaborates on what constitutes a family firm and the two approaches to identify a small- and medium-sized firm. Second, I discuss the literature on the unidimensional and the multidimensional perspective of firm professionalization to clarify the dependent variable of my research. Finally, I describe the outcome of growth, chosen CEO’s characteristics, and develop my hypotheses.

2.1 The Definition of Small- and Medium-sized Family Firms

The research history of family firms goes back over 30 years and yet no accepted generic definition of what constitutes a family firm has emerged, and there are many questions that remain unanswered (Littunen & Hyrsky, 2000; Wagner et al., 2015). From early on, the general agreement was that the ownership and involvement of family members in management/governance are sufficient to constitute a family firm (Handler, 1989; Miller & Rice, 1967). Arguably, this insinuates that families provide unique resources and capabilities to the firm that non-family firms do not have (Pearson et al., 2008). However, this view has softened over time and additional understandings of what may constitute a family firm have emerged (Chua et al., 1999). Nevertheless, does that mean that a firm is different simply due to family ownership and involvement?

are generally accepted across literature: Family-owned and -managed, family-owned but not -managed, and family-managed but not -owned. However, they emphasized that this alone does not define a family firm. Firms with the same level of ownership and management can still act very differently. Some do consider themselves as family firms, whereas others do not. One must identify that it is uniqueness that results from family involvement in patterns of ownership, governance, management, and succession, since these influence the firm’s goals and strategies. The authors suggest operationalizing the firm’s vision (using the firm to achieve a better future for the family) for this matter due to the influence on a firm’s behavior in order to distinguish family firms from non-family firms. Put differently, Chua et al. (1999) suggest a theoretical definition based on behavior where a “family firm is a firm governed and/or managed with the intention to shape and pursue the vision of the business held by a dominant coalition controlled by members of the same family or a small number of families […]” (p. 25). A dominant coalition thereby describes the actors within a firm equipped with power to control the firm’s orientation (Hambrick & Mason, 1984). It is not necessarily that one single family hold the control over the firm, but a dominant coalition of families. In order to apply this theoretical definition to practical use, they suggest asking firms for their vision to assess if the corresponding firm is a family firm.

This definition offers several benefits. First, with the attempt to pursue the family’s vision, the firm may act differently compared to non-family firms. Thus, firms who consider themselves as a family firm, and therefore act like a family firm, are taken into account. Second, it incorporates the general accepted combinations of ownership and management and, as a result, has strong support in the literature (Handler, 1989; Miller & Rice, 1967). A firm that is family-owned and -managed but does not consider itself to be a family firm is, following the definition, not per se a family firm. This means that the general accepted conditions (managed and/or owned) that constitute family firms are no longer the sole determining conditions. Instead of ownership and management structures, this definition additionally emphasizes the behavior of firms as a distinguishing factor. This study will build on the general idea of Chua et al.’s (1999) definition. That means a firm is seen as a family firm if families hold its shares’ majority and/or at least one family member is involved in the management and/or the firm considers itself as a family firm.

The emphasis of this study is not just to investigate family firms, but small-and medium-sized family firms. In general, small- and medium-sized firms can be classified qualitatively and quantitatively. Nooteboom (1994) points out three distinct characteristics that define such firms. First, they have a

small scale, which means they do not enjoy economics of scale. These firms usually face resource

medium-sized firms. Their definition is based on staff headcount, turnover, and balance sheet (European Union Commission, 2003). Table I shows the numerical range of the EU’s definition.

Table I.

Definition of small- and medium-sized firms (European Union Commission, 2003)

Firm classification Employees Turnover Balance sheet

Small-sized firm < 50 ≤ 10 Million Euro ≤ 10 Million Euro

Medium-sized firm < 250 ≤ 59 Million Euro ≤ 43 Million Euro

The quantitative approach suggests clear benchmarks that are easy to implement in the research compared to the suggested by Nooteboom (1994) which are rather difficult to investigate. Therefore, this research utilizes the quantitative definition in the method section to identify small- and medium-sized firms, but both definitions are relevant in order to understand the specific role that small- and medium-sized firms have in the economy.

2.2 The Multidimensional Approach Towards Firm Professionalization

Firm professionalization is an ambiguous construct. Family firm literature often refers to professionalization as the entrance of a non-family member in the management (Gedajlovic et al., 2004; Lin & Hu, 2007; Zhang & Ma, 2009). It is assumed that non-family managers contribute knowledge and capabilities that cannot be contributed by family members because they are possibly less qualified (Gedajlovic & Carney, 2010; Schulze et al., 2001). Subsequently, the firm makes use of the provided knowledge and capabilities – they professionalize (Stewart & Hitt, 2012; Tsui-Auch, 2004).

1985). Recapitulating, solely relying on the employment of non-family managers in order to describe firm professionalization seems to be inadequate since it is not clear what benefits a non-family manager can contribute to the firm. This unidimensional approach may have especially negative consequences in small- and medium-sized family firms, which often face resource constraints and thus they might be less willing to invest in costly monitoring.

Next to the agency theory argumentation, the expectations of family members to achieve formal education and outside work experience has risen in recent years (Aronoff, 1998; Venter et al., 2005). This challenges the view that family members cannot contribute the required capabilities and knowledge (Gedajlovic & Carney, 2010; Handler, 1990; Schulze et al., 2001). Combining education and work experience gained outside the family firm as the general human capital of successors, Sardeshmukh & Corbett (2011) find a positive link to opportunity perception. They suggest that education refines generic and analytical skills and improves the ability to transfer management concepts to their own family firm. It also helps gaining legitimacy in the firm and mitigates non-family employees’ thoughts of nepotism. Moreover, outside work experience shows that the successor is able to work under different bosses and circumstances, which may benefit the development of managerial techniques (McCall et al., 1988). Overall, this suggests that family members are as likely as non-family members to have obtained managerial education and experience (Dyer, 1989). The unidimensional approach seems to be too limited to capture all aspects of firm professionalization, which calls for a more inclusive, comprehensive, multidimensional approach.

but only explained a small part of the variance (adjusted R2_{=0.06). Her research nevertheless provides}

a useful avenue for my research; I can build on her work and use her multidimensional approach as a foundation.

In what follows, I review and, if necessary, adjust the proposed five dimensions according to insights presented by the literature in order to enhance the firm professionalization construct. In the analysis section, these dimensions assemble the dependent variable – the firm professionalization construct.

2.2.1 Financial and human resource control systems. Competencies in the fields of human resource management and financial management are found to be crucial for small- and medium-sized firms, since these firms are relatively labor-intensive and have limited access to finance (Brock & Evans, 1989; Storey & Greene, 2010). This supports the idea that the dimensions (1) Financial Control Systems and (3) Human Resource Control Systems can be a valuable part of explaining firm professionalization since this becomes even more important with increasing complexity. The firm’s tasks to recruit the right people and to keep hold of the firm’s money are supported by the implementation of formal recruitment systems (Flamholtz & Randle, 2012), formal training systems (de Kok et al., 2006), budgets, and formalized financial goals (Daily & Dollinger, 1992; Pérez de Lema & Duréndez, 2007; Songini, 2006). This implies changing the informal setting of a family firm to a more formal one by introducing formalized systems (Chua et al., 2009; Dyer, 2006; Gedajlovic et al., 2004). Thus, the proposed dimensions by Dekker (2012) (1) Financial Control Systems and (3) Human Resource Control Systems are accepted and incorporated in the professionalization construct for this research.

2.2.2 Monitoring and resource providing governance systems. Moreover, Dekker (2012) proposes “Non-Family Involvement in Governance Systems” as a dimension of family firm professionalization. She includes non-family involvement in their professionalization construct due to its deep connection with family firm professionalization in the literature. As argued during the beginning of this section, non-family involvement does not seem to be a good predictor of firm professionalization in general, and for small- and medium-sized family firms in particular. Non-family involvement can be one component of it, but surely not a whole dimension of professionalization. Supporting the notion that non-family involvement is not crucial, Dekker et al. (2013) find in their exploratory study regarding the dimensions of professionalization four different clusters of family firms. One of them, domestic configuration, describes a firm type where management is still largely in the hands of the family. This firm type focuses on family managers and formalized structures, e.g. formal control mechanisms, which means that they professionalize without the involvement of non-family managers.

firms since utilizing a supervisory board is costly (accounting for expense allowance for board members) and transfers formal power to outsiders. Therefore, this dimension should rather describe a well-developed governance system rather than simply the involvement of non-family members.

Hillman & Dalziel (2003) elaborate on the two distinct functions of boards, those being the monitoring of managers and the provision of resources. Using agency theory, they explain that governance systems attenuate agency costs since they monitor the agent’s behavior. Moreover, the authors apply resource dependence theory to suggest that governance systems shall provide resources to the firm in form of expertise and advice. Based on the different roles that governance systems should fulfill within a firm context, one can conclude that it is not the non-kinship of board members that is important, but rather the added value. Therefore, the firm’s monitoring ability as well as its service/advice needs should drive the governance involvement (Grundei & Talaulicar, 2002; Voordeckers et al., 2007). Van den Heuvel et al. (2006) point out that CEOs of small- and medium-sized family firms perceive the role of service as more important. The provided expertise is especially valuable for smaller firms that might suffer from managerial deficiencies. Considering the previous discussion, I refine the proposed dimension by Dekker (2012) to (2) Monitoring and Resource Providing Governance Systems.

Centralization of authority seems to be a characteristic of family firms (Chrisman et al., 2013; De Massis et al., 2014), and decentralization of authority is necessary in order to deal with increasing organizational complexity (Chittoor & Das, 2007). Considering the foregoing discussion, I accept (4) Decentralization of Authority as one dimension of the professionalization construct and will use it in this research.

2.2.4 Formal strategy-making. Dekker (2012) furthermore elaborates on (5) Top Level Activeness, which essentially refers to the number of board meetings as a determinant of the quality of decision-making in family firms. Admittedly, improving decision-decision-making quality must be a goal for professionalizing family firms. Quick decision-making is often considered one of the major advantages of small- and medium-sized family firms in comparison to larger firms (Ward, 1997). Decision quality, however, refers to the decision’s contribution to the achievement of firm goals (Mustakallio et al., 2002), whereas decision commitment describes the degree to which affected parties accept the reached decision (Korsgaard et al., 1995). Taking these three dimensions (speed, quality, commitment) together, the concept of good decision-making becomes clearer. But how can this be achieved? Formal strategic planning might be the key to success. Family firms are prone to “undershoot” (Ward, 1988, p. 115) their strategic potential by neglecting planning. In order to achieve good and comprehensive strategy-making, and, as a result, good decision-making, planning is vital. This becomes especially true for growing firms, which exhibit increasingly complex structures (internal and external). Grant (2003) emphasizes that formal planning systems are instrumental in coordinating strategy-making in complex environments. However, small- and medium-sized firms often reject planning, arguing that more disadvantages than advantages are involved, such as concentrating limited resources to unknown means (cf. Ward, 1997). Essentially, these CEOs consider planning to be an impediment to flexibility. This means more bureaucracy (Andersen, 2004) due to different forecasting methods, and such, that must be discussed and evaluated (cf. Wolf & Floyd, 2017), and also because it is more difficult to deviate once a plan is made. It seems that it strongly depends on the actors involved (and particularly the CEO) whether and, if yes, to which degree they make use of strategy-making.

I argue that family firms should adopt formal strategy-making mechanisms to cope with the increasing complexity in order to ensure good decision-making. The balance between flexibility and formal strategy-making is crucial to avoid being overcome in bureaucracy. Nevertheless, strategy-making supports the firm by coping with the outcome of growth (increasing complexity). As a result, the dimension “Top Level Activeness” of family firm’s professionalization is adjusted to (5) Formal Strategy-making.

Control Systems, (4) Decentralization of Authority, and (5) Formal Strategy-making. I use these dimensions to operationalize family firm professionalization in my research (see section 3.2.4).

2.3 The Outcome of Growth

Company growth literature mainly describes growth as an outcome (McKelvie & Wiklund, 2010). Researchers aim to find predictors of growth and explain variations in growth rates (Baum et al., 2001; Gilbert et al., 2006). A significantly smaller stream of literature examines the aftereffect of growth with a precise focus on the issues associated with managing a grown firm (McKelvie & Wiklund, 2010; Phelps et al., 2007) and operating in a more complex environment (Stewart & Hitt, 2012; Weidenbaum, 1996; Yildirim-Öktem & Üsdiken, 2010; Zeng & Williamson, 2007; Zhang & Ma, 2009). This section includes an elaboration about organizational and environmental complexity as an outcome of growth as well as the development of the first two hypotheses.

2.3.1 Organizational complexity. Company growth theory suggests a managerial approach (firm-level) in strategy formulation, planning, formal governance, control systems and decentralized structures, mainly due to increasing organizational complexity after firm growth (Deakins et al., 2002; Irvin, 2000; Kroeger, 1974; Penrose, 1959; Songini, 2006). Firms shall implement appropriate processes and structures to manage their growth. However, the biggest concern for many firms is that they are not able to cover their expenses and they, therefore, continue to seek growth. Thereby, they forget the threat of being overwhelmed by rapid firm growth and are overcome by it. Why is that so?

Flamholtz & Randle (2012) identify the underlying issue as being the organization’s failure to match internal infrastructure with complexity; they summarize it by using the term “growing pains” (p. 48). They observed this in a broad variety of firms within different industries and sizes. Growing pains are issues that occur as a result of inappropriate internal organizational development in relation to size and complexity. Something has gone wrong in the process of organizational development and a further step is required to deal with it. Flamholtz & Hua (2002) find a significant link between growing pains and financial performance, which suggests that growing pains are a signal of success (because the firm has grown), but provide bleak outlooks if the firm does not adjust their internal infrastructure in time.

employees. The examples show the potential harm a mismatch between internal infrastructure and size can cause.

According to Dooley (2002, p. 5014), organizational complexity can be defined as “the amount of differentiation that exists within different elements constituting the organization”. This definition is in line with the aforementioned growing pains. For instance, a firm where people are not aware of what others do may indicate high differentiation between two departments. This is then called organizational complexity. I argue that if firms are exposed to such complexity, they are driven to implement formal systems (Davila, 2005), decentralize authority (Chittoor & Das, 2007), and seek advice from governance councils (Van den Heuvel et al., 2006) in order to cope with it. The implementation of formal systems is necessary because the informal approach requires being in contact with everyone. Due to increasing complexity and corresponding rise in formal communication needs, it becomes beneficial to coordinate communication through formalized systems. The firm benefits from adopting a more decentralized decision-making structure in order to ensure appropriate responses to new (and probably more complex) tasks. Increasing organizational complexity as an outcome of growth increases the need for firms to professionalize and adopt formal control mechanisms, strategic planning, and decentralization of authority. As a result, I hypothesize:

H1: Organizational complexity positively drives small- and medium-sized family firms towards

professionalization.

2.3.2 Environmental complexity. A second outcome of growth is external complexity (environmental-level). Finkelstein & Hambrick (1996) link environmental complexity with diversification. Gabrielsson (2007) and Sanders & Carpenter (1998) describe internationalization as an important form of diversification. Exporting to different markets outside of the home country and/or running a subsidiary beyond the home market inevitably lead to new issues, and this subsequently requires professionalization of the firm. Zhang & Ma (2009) identify new market necessities as a stimulant for professionalization. For instance, new and perhaps international competitors push the firm to improve its underlying processes and structures in order to stay profitable and competitive. In fact, the intense market competition as an outcome of growth forces firms to adapt to the new environment (Stewart & Hitt, 2012; Weidenbaum, 1996; Yildirim-Öktem & Üsdiken, 2010; Zeng & Williamson, 2007; Zhang & Ma, 2009). Increasingly competitive and complex markets raise the complexity firms deal with. If firms increase in size, revenue or profits, they usually find themselves in a new environment with an increased number of external influence factors (e.g. new customers, new competitors, governmental control).

found five different components. First, the customer component which incorporates distributors and actual users of the firm’s product/service.Second, the supplier component that describes labor and material suppliers. Third, he describes competitors for suppliers and customers as the competitor component. Fourth, he identifies a socio-political and a technological component as the last components of the external environment. Subsequently, he orders the responses in a simplex-complex continuum in order to predict the environments of perceived uncertainty. This is related to our perception of environmental complexity as an outcome of growth because it reflects the increasing complexity that results from growth (more external factors mean more complexity). In his proposed continuum, environmental complexity incorporates the number of external influence factors on a perceptual-level. As a result, the environmental complexity is dependent on the observing organization (Schneider et al., 2017). For instance, after a period of growth a firm my find itself in an environment with increased customer demand, perhaps more suppliers of raw material or labor, increased governmental control due to the economic relevance, and new technology approaches. It seems that all these factors define a firm’s environmental complexity.

Child & Rodrigues (2011) use a similar view of complexity. They relate key organizational actors (e.g. competitors, suppliers) with key factors that influence the environmental conditions (e.g. government, technology change). So, again, environmental complexity describes an increasing number of factors influencing the behavior of the firm. Additionally, they include the perspective of “cognitive complexity” proposed by Boisot & Child (1999) in their assessment of environmental complexity. Cognitive complexity describes the challenge of knowing and understanding the firm’s environment. This seems to be particularly challenging for small- and medium-sized firms; their management teams often consist of a few individuals whose skills are pivotal in the firm’s managerial capabilities (Brunninge et al., 2007).

It seems that the perceived number of external influence factors, and the ability to grasp it, build an adequate way to assess a firm’s environmental complexity. The five dimensions suppliers, competitors, customers, socio-political & technological influence, and cognitive complexity then build a comprehensive picture of the firm’s environment. I argue that in order to cope with the increasing number of external influence factors, the firm is driven towards the next step in the organizational development, namely professionalization. Thus, I hypothesize:

H2: Environmental complexity positively drives small- and medium-sized family firms towards

professionalization.

2007; Man et al., 2002) where the CEO has even more substantial influence over the firm. The CEO’s individual beliefs attitude towards professionalization might contribute significantly to explaining variations in professionalization intentions among family firms. In this section, I describe two well-acknowledged personality traits in the literature and develop two hypotheses based on “upper echelon theory” (Hambrick & Mason, 1984).

2.4.1 Unwillingness & reluctance. As mentioned in section 2.2, family firms generally have lower agency costs than their non-family counterparts (Fama & Jensen 1983; Jensen & Meckling 1976; Pollak, 1985). Several authors even argued that family firms are free of agency costs (Becker, 1974; Eisenhardt, 1989) but face challenges widely unknown to non-family firms (Litz, 1997; Schulze et al., 2001). For instance, Hendry (2002) points out that they are exposed to a higher risk of attracting less qualified employees, and/or tying them to the firm (e.g. through shares), due to limited financial resources and the family’s unwillingness to share control. In the same vein, Brunninge & Nordqvist (2004) suggest that family firms are less likely to tie external directors to the firm and endow them with formal power because they are reluctant and unwilling to share control. This indicates that family firms face problems that non-family firms usually do not have. The underlying issue seems to be the unwillingness to share control and thus increase agency costs.

The unwilling and reluctant attitude often shown by family-owned and -managed firms also affects the installation of advanced managerial control systems and governance systems (Gedajlovic et al., 2004). Tabone & Baldacchino (2003) indicate that family firms have difficulties in controlling management decisions because they lack the necessary knowledge and tools. In the same vein, Jorissen et al. (2001) confirm that family firms make lesser use of financial controls than their non-family counterparts. In an empirical study, García Pérez de Lema & Duréndez (2007) test the managerial differences between family firms and non-family firms, and confirm the above findings for a sample of 639 private small- and medium-sized Spanish industrial firms. That suggests that family firms in Germany may also make lesser use of managerial control systems and governance systems.

However, I think that if a person is generally uncomfortable with changing plans, this will reflect in firm decisions such as the introduction of formalization or delegation of control.

Summarizing the above paragraphs, I argue that family firms, usually exposed to a lower degree of agency costs due to the overlap of management and control, are reluctant to increase such agency cost during the process of professionalization through delegation of control and formalization. Non-family firms, on the other hand, who are exposed to higher agency costs from the start due to the separation of management and control, rather try to increase their level of professionalization in order to lower agency costs. Using “upper echelon theory” (Hambrick & Mason, 1984), which suggests that strategic choices (e.g. growth/professionalization) reflect norms and values of powerful actors within the firm, a connection between the decisions on the individual-level and firm-level outcome is drawn. Put differently, the CEO must make the decision to delegate decision-making power, introduce costly control systems, and implement an appropriate monitoring structure, in order to professionalize the firm. If the CEO is unwilling or reluctant to do so, it negatively influences the professionalization process within the firm because the firm’s behavior reflects the CEO’s characteristics. As a result, I hypothesize:

H3: CEO’s unwillingness and reluctance to lose control have a negative effect for small- and

medium-sized family firms on firm professionalization.

creation process of firm professionalization is then out of their hands and the achievements are no longer related to them. As a result, I expect that the positive influence of need for achievement on professionalization will turn negative when the need for achievement is too high and the CEO becomes reluctant to share power any further and decentralize decision-making. This attenuates the professionalization process within the firm, suggesting that there might be an optimal level of need for achievement regarding firm professionalization inherited by the CEO. As a result, I hypothesize:

H4: CEO’s need for achievement has a curvilinear effect for small- and medium-sized family firms

on firm professionalization (inverted-U shape).

I have summarized the proposed relationships in a conceptual model shown in figure I. The dependent variable is the multidimensional construct of firm-level professionalization from section 2.2. The model shows the expected positive influence of organizational (firm-level) complexity and environmental (environmental-level) complexity as an outcome of growth on firm-level professionalization. I developed the hypotheses in section 2.3. I also added the expected impacts of CEO’s characteristics on firm-level professionalization from section 2.4 to the model. I expect the individual-level variable “unwillingness and reluctance” to have a negative influence on professionalization. For “need for achievement”, I propose a curvilinear relationship.

Figure I.

3

M

First, I present the data collection and sample creation process; second, I present the measures and the development of the questionnaire; finally, the validity and reliability of the data set is discussed.

3.1 Data Collection and Samples

I made use of two samples to test the hypotheses. First, I created a convenient sample through my personal and family network, which resulted in a sample size of 85 CEOs of German small- and medium-sized family firms. Second, I identified relevant firms from the orbis database. Only firms located in Germany meeting the criteria of the European Union for small- and medium sized firms in terms of number of employees were chosen. The required legal forms are limited-liability, partnership, and sole proprietorship to exclude cooperatives, corporations, and parties. Furthermore, only firms with an available e-mail address and a minimum turnover of 10,000 Euro were included to ensure that the selected firms can eventually receive the survey via e-mail, and to ensure that I contact only active firms. These are firms with at least a little turnover. Moreover, the population contains all industries except banking and insurance to exclude investment firms, which are not relevant for this study because they often operate as holdings. From 8,938 firms obtained through this search strategy, a sample of 7,200 firms was randomly created in two steps by the qualtrics sample-creation software using the setting “most random”. I distributed an individual link via e-mail to make sure that I could send reminders to unfinished respondents. This was done for both samples. After accounting for bounced back emails and invalid addresses, I ended up with a total audience size of 6,628. For further support, I additionally contacted the German Chambers of Commerce and Industry (IHK) in Cologne, Hamburg, and Stade as well as the SME and Economic Association of CDU and CSU (MIT) - with more than 25,000 members, this is the strongest and most influential party-political business association in Germany. However, neither the IHK nor the MIT were willing to support my endeavor. Two reminder messages to unfinished respondents were sent in order to boost the response rate. As well as this, I provided a QR-code with an introduction text to a cooperating business-to-business firm in the case that their customers are interested in this study. Three firms used this option and participated in my survey.

ensure that the meanings of the items/questions are clear to CEOs, and to avoid typical biases in surveys e.g. through double negative items. Again, they did not find inconsistencies but all parties emphasized that the questions regarding “Decentralization of Authority” were difficult to understand and not clear. I adjusted the questions concerned to make them clearer. After the second round of reviewing, the survey was found viable, well-structured, and clear. At the end of the survey, the participants could enter their e-mail address to receive a summary of the results as a reward for participating in order to increase the response rate. Additionally, the participants had the opportunity to share their thoughts to capture all potential insights from the CEO arising during the survey. Appendix I provides a detailed overview of the questionnaire.

Within three weeks, I collected 97 responses (1.5 % response rate) with 77 finished surveys (79.4 % completion rate). I added three surveys because they are sufficiently complete and include the main variables. From these 80 sufficiently completed surveys, I selected the firms that meet the utilized definition of a family firm as well as the size requirements, resulting in 58 firms (72.5 % of the surveys). I decided to include the five firms that were above the threshold of 250 employees but met the family firm definition to make the best use of my available data. I checked the five firm’s websites to assess whether the indicated number of employees is plausible. No incorrect indications have been found. As a result, my final data set includes 63 family firms with at least one employee.

At first sight, the collected number of responses seems rather low. However, they are comparable to other studies if taken into account that survey response rates in family firm research are generally low (Winter et al., 1998), in particular the rates for German firms (Klein et al., 2005; Nagl, 2005). Research on top management teams also often suffers from low response rates (Baruch & Holtom, 2008; Hambrick et al., 1993; Minichilli et al., 2010). Bearing in mind this challenging initial situation, the number of responses is neither large nor unusually small.

3.2 Measurements

Where possible, measures are adapted from or related to existing validated scales from literature. Appendix II provides an overview of the variables, items, translation, and sources. I assess the construct’s composition and validity during the factor analysis in section 4.2.

explicitly ask the respondents to think about the firm in recent years in order to mitigate potential causality issues.

3.2.2 Environmental complexity. Earlier, I explained that environmental complexity consists of five dimensions. The first four dimensions are Duncan’s (1972) perceptual measures of environmental complexity: customers, suppliers, competitors, socio-political/technological components. Duncan showed the respondents eight statements about whether the number of customers/suppliers/competitors has increased, whether the government regulatory control over the industry has increased, and whether it has become more difficult to meet technological requirements of the industry. The fifth dimension is Child & Rodrigues’s (2011) cognitive complexity. To measure, I use two statements “It has become harder to understand what is going on in the firm's environment” and “It is easy for me to proceed the gained information about the firm's environment (market data, competitor's behavior etc.)”. The latter one is a reversed item. Then the respondents are asked to what extent they agree with the statements on a 1 to 5 Likert-scale, with 1 as strongly disagree and 5 as strongly agree. I explicitly asked the participants to think about the firm's environment in the recent years to attenuate potential causality issues.

3.2.3 Need for achievement and unwillingness & reluctance. A well-known tool used to measure “needs” is the Manifest Needs Questionnaire by Steers & Braunstein (1976). It consists of 20 items measuring an individual’s needs for achievement, affiliation, autonomy, and dominance. This measure, however, is very extensive and more compact measures are available. Wu & Dagher (2007) use a rather simple, but still related to the “achievement scale” of Steers & Braunstein, four-item scale to measure “need for achievement”. They asked the respondents to what degree they would agree with the statements, “I need to meet the challenge, I need to continue learning, I need personal growth, and I need to prove that I can succeed”. Robichaud et al. (2001) provide evidence that the items are reliable, valid, and load the correct construct. The items are measured on a 1 to 5 Likert-scale with 1 as strongly disagree and 5 as strongly agree.

3.2.4 Professionalization. The dependent variable consists of five dimensions. Firstly, the dimension “Formal financial control systems” consists of the three items utilized in the study by Dekker et al. (2015). The respondents are asked to what extent the firm makes use of budgets, budget evaluation systems, and formalized financial goals and objectives. The items are measured on a 1 to 5 Likert-scale with 1 as “to a very slight extent” and 5 as “to a very strong extent”. Secondly, the dimension “Monitoring and resource providing governance systems” is made up of four items taken from Van den Heuvel et al. (2006). For this research, I utilized the two "most important" rated tasks by CEOs for service role and monitoring role, respectively. The respondents are asked to indicate on a five-point Likert-scale to what extent the governance system of their firm fulfils the tasks “Building organizational reputation”, “Formulate/ratify organizational strategy”, “Direct succession problems”, and “Evaluate/control management performance”. Thirdly, the dimension “Human resource control systems” consists of five different items. The respondents are asked to indicate on a five-point Likert-scale to what extent they make use of the items. These range from “Formal recruitment system” to “Formal scheduled staff meetings”. Dekker et al. (2015) have successfully used this scale in their research. Fourthly, the dimension “Decentralization of authority” consists of ten items suggested by Martin et al. (2016). On a five-point Likert-scale, the respondents are asked to what extent they delegate the tasks/decisions “Decision about the number of workers required”, “Whether to employ a worker” etc. A high extent suggests a strong decentralization of authority whereas a slight extent suggest a strong centralization of authority. Finally, the dimension “Formal strategy making” relies on seven items suggested by Andersen (2004), Hart & Banbury (1994), and Ward (1997). The first four items can be summarized as a rational component (e.g. “We have a written mission statement that is communicated to employees”), whereas the last three rather assemble an involving component (e.g. “Business planning in our firm is ongoing, involving everyone in the process to some degree”). Again, the items are measured on a 1 to 5 Likert-scale with 1 as strongly disagree and 5 as strongly agree.

3.2.5 Control variables. Several control variables are included in this study. The firm professionalization literature commonly refers to firm size, more specifically, the lognormal transformed number of employees (see e.g., Delmar & Wiklund, 2008). Relatively larger firms have advantages (e.g., access to more resources) as well as disadvantages (e.g., less flexibility) with regard to professionalization that must be accounted for. Furthermore, Davidsson (1991) suggests using the

owner’s age in order to take into account potential age-related influences at the individual-level. Owners

serves as a reference point. I consider the CEO’s gender because literature mentions differences in firm behavior related to this (e.g. Faccio et al., 2016). Moreover, I control for industry to account for any environmental effects specific to certain industries. The participants are asked in which industry (construction, manufacturing, service, others) their firm operates, considering the firm’s main activity. As suggested by Lubatkin et al. (2006), I dummy coded firms as “construction”, “manufacturing”, and “service”. The reference point in my model is the “other businesses”. Additionally, I aim to consider

firm’s region to control for potential effects on professionalization aroused by differences in access to

resources. For instance, firms operating in regions with a higher density of universities might have better access to human capital (Schubert & Kroll, 2016) which subsequently helps them to professionalize.

3.3 Validity and Reliability

In this section, I check the collected data’s validity and reliability. First, I evaluate the representativeness of the data set. Therefore, I compare several descriptive data with the population and the convenience sample. Subsequently, I conduct analysis to check for potential biases often in common research.

3.3.1 Representativeness. I compare the net-sample’s characteristics (respondents) to the population, the convenient sample, and the family firms in terms of region, number of employees, and firm age. Table II provides an overview. For the population, the relevant characteristics were available. In order to compare the convenience sample as well, I conducted additional internet research and sometimes the firms were directly called to get additional information.

Table II.

Comparison of selected characteristics between population, sample, and respondents

Population, n=8,938 Convenience Sample, n=85 Respondents, n=97 Family Firms, n=63

Most frequent regions North

Rhine-Westphalia (24 %)

Lower Saxony (90 %) Lower Saxony

(34.1 %); North Rhine-Westphalia (23.9 %) Lower Saxony (44.4 %); North Rhine-Westphalia (23.8 %) Average number of employees 105 62 103 90

Average firm age in years

36.4 25.1 44.8 41.8

personal contacts to firms mainly located in Lower Saxony and, thus, a higher percentage is plausible. In terms of number of employees, the firms from the population employ on average 105 people, the responding firms 103 people, whereas the selected family firms employ on average only 90 people. The responding firms were on average 8.4 years older and the selected family firms 5.4 years older than the population firms were. This favors our study since the goal is to investigate family firms with established routines who have long been in business and have experienced consequences of growth.

3.3.2 Key informant approach and retrospective bias. This study relies on survey answers provided by CEOs. This is an often-used approach in small- and medium-sized firm research (Dehlen et al., 2014; Kammerlander et al., 2015) and is based on the assumption that CEOs have the best knowledge about the firm’s behavior, characteristics, and activities. The e-mail recipient is asked to forward the survey to the CEO or, if not available, a member of the top management team. From the final data set of 63 family firms, eight respondents indicated that they are not the CEO. The survey design ensured that the respondents subsequently must indicate their position in the firm if they are not the CEO. The indicated positions are, for instance: Head of R&D, authorized representative, and division manager. I consulted the firm’s website to find additional information regarding the organizational structure and to assess whether the respondent could have sufficient insights into the firm. All respondents are in a position to adequately answer the survey.

Moreover, the specific research design avoids retrospective bias. The retrospective bias can seriously skew the survey answers and therefore a good research design is crucial in order to obtain appropriate data (Huber & Power, 1985). The survey questions regarding the independent variable – the outcome of growth – have a strong emphasis on the past situation in the firm. On the other hand, the survey questions regarding the dependent variable – professionalization – emphasize the status in the firm. This specific design strengthens the causality of the “outcome of growth – professionalization” relationship by attenuating the retrospective bias.

Furthermore, I considered the F-value of Leven’s test for equality of variances and, again, for the first three variables the null hypothesis cannot be rejected assuming equal variances. Only firm size shows significant results and makes us reject the null hypothesis (p=0.006, F=9.251). The t-test for equality of means thus indicates a significant difference between the means on a 10-% alpha-level (p=0.095). This suggests that, if there are non-response bias, they are rather small since the results are insignificant on the 5-% alpha-level.

3.3.4 Common method bias. In order to avoid the common method bias, several actions have been made. First, the participants are ensured anonymity and confidentiality, which decreases the respondents’ motivation to give socially desirable answers (Podsakoff et al., 2003). Second, to reduce potential common method bias, the data for this study should have been collected from multiple sources and with temporal differences. However, this approach seemed to be unsuitable for my research since another concern is the response rate and the limited temporal scope of this study. Therefore, I used statistical remedies to control for common method bias (Podsakoff & Organ, 1986; Podsakoff et al., 2003). Podsakoff & Organ (1986) propose using Harman’s (1976) single factor test which relies on the assumption that if a critical amount of common method bias is present, either the factor analysis including all construct items will excel one single factor or one general factor will explain most of the variance in an unrotated solution. To substantially address the relationships, the common method variance would need to account for 70 % (Fuller et al., 2016). In my study, 10 factors are exceled based on eigenvalue greater than 1. Furthermore, the one factor solution accounts for 22.46 % of the variance. As a result, common method bias does not appear to be a major issue in the current study.

4

A

Firstly, the data set was cleaned and prepared for further analysis with the statistical program SPSS. The descriptive variables are presented in section 4.1. Section 4.2 shows the analysis of the constructs. The necessary assumptions for linear regression are checked in section 4.3 and the results can be found in section 4.4. Finally, I conducted additional analysis of the meaning of professionalization by assigning firms to one of three groups, according to their answer on the “meaning of professionalization” question from my survey. Section 4.5 provides the corresponding analysis.

4.1 Descriptive Variables

my data set are larger (90 employees compared to 29 employees) and older (42 years compared to 27 years). That the firms are larger may be because I included five firms with more than 250 employees in my analysis. The high standard deviation (141.038) indicates this. Having older firms in my data set favors our attempt to investigate the consequences of growth on professionalization, since older firms are more likely to have suffered from such consequences. Table III summarizes the descriptive statistics of my data set. Furthermore, 95 % of the firms responded that they consider professionalization to be important for their firm, which indicates the managerial relevance for investigating this topic from an academic perspective.

Table III.

Descriptive statistics

N Min Max Mean SD

CEO age 63 21 67 48.13 10.949 CEO tenure* 55 0.167 35 15.45 9.993 Number of Employees 63 1 850 89.76 141.038 Firm age 63 0.167 200 41.76 39.750

* 8 respondents are not CEOs and thus are not asked for their tenure.

Almost all respondents (98.4 %) obtained at least vocational training (a very common form of education in Germany) and 58.7 % even graduated from a university or a university of applied sciences. The high level of education may not be representative for the population of Germany and, at a first glance, might sound disproportionate, but it does underpin the fact that the firms in my data set are long-established firms where the well-educated successor took over from the founder. Moreover, the firms in my data set are mainly located in the service industry (47.6 %), followed by the manufacturing industry (20.6 %) and construction (9.5 %). 22.2 % indicated that they operate in a different industry than the aforementioned. Table IV provides an overview about the industries.

Table IV.

Industry according to the family firm’s main activity

Frequency, n=63 Percent

Service 30 47.6

Manufacturing 13 20.6

Construction 6 9.5

Other 14 22.2

in alphabetical order. Unfortunately, my data set does not provide an adequate number of firms from different regions. In particular, the regions located in the former territory of East Germany are often subject to investigation but are not sufficiently present in my data set. That means that additional value from controlling for region and/or conducting further analysis with the regional data may not be expected. Consequently, I decided to exclude the different regions as control variables in my study. However, I belief that this may be an interesting and promising approach, particularly in Germany (see section 5.1).

Table V.

Distribution of the family firms across Germany

Frequency, n=63 Percent Baden-Württemberg 4 6.3 Berlin 1 1.6 Bavaria 8 12.7 Bremen 1 1.6 Hamburg 1 1.6 Hesse 1 1.6 Lower Saxony 28 44.4 North Rhine-Westphalia 15 23.8 Rhineland-Palatinate 2 3.2 Saxony 1 1.6 Saxony-Anhalt 1 1.6

Regarding firm’s growth history, 60.4 % indicated that their firm has grown in number of employees over the last five years and 66.6 % indicated that for the firm’s turnover respectively. 92.1 % (87.3 %) of the firms anticipate their future growth intention to be either stable or increasing when it comes to number of employees (turnover). The high number of firms that have grown during the last five years favors our study since the likelihood for experienced consequences of growth is higher.

4.2 Factor Analysis and Construct Composition

Before starting with the factor analysis, it is important to assess whether the applied constructs are reflective or formative to appropriately investigate the relationships and the validity. The constructs

Organizational Complexity, Need for Achievement, and Unwillingness & Reluctance are reflective

aspects of the environment. Furthermore, the flow of causality is not fully clear. A logical view can also be that the various complex aspects of the environment form the construct. Thus, the flow of causality is from the various aspects of the firm environment through the items chosen to what is defined as environmental complexity, rather than latent complexity being reflected in the chosen items. To account for this ambivalent construct in my analysis, first environmental complexity is suggested as a reflective construct, thus included in the factor analysis, and in a second step, taken as a formative construct that keeps all items. This construct is from here onwards called Environmental Complexity (broad). The dependent construct Professionalization is clearly a formative construct because the items are not interchangeable and the flow of causality is from the items to the construct; the items thus form the latent construct.

I checked the construct validity in two steps. I conducted an exploratory factor analysis including all 31 items of the reflective constructs used as the independent variables to investigate whether the measurement items reflect the constructs appropriately. The results of the closed factor analysis (predetermined number of factors: 4), more precisely a principal component analysis (PCA), will be assessed according to the following criteria: (1) each measure must have a loading > 0.45; (2) each measure must not have a loading > 0.45 to more than one factor; (3) loading into the correct factor. Items that do not meet the criteria will be deleted (cf. Song et al., 2011, slightly lower (1) and higher (2) thresholds). I analyzed the covariance matrix as the item scales are commensurable on a 5-Point Likert level (Field, 2013). I decided on the varimax rotation because it simplifies the interpretation through strong item loadings in one factor (Kaiser, 1958). The Bartlett test of sphericity (χ2 _{= 896.972;}

Table VI.

Results of the varimax-rotated factor model

Factors Environmental Complexity Organizational Complexity Need for Achievment Unwillingness & Reluctance EC_01 .782 EC_02 .745 EC_03 .448 EC_06 .493 OC_01 .533 OC_02 .692 OC_03R .686 OC_04 .782 OC_05R .696 OC_07R .686 OC_08 .618 OC_09 .594 NfA_01 .889 NfA_02 .525 NfA_03 .720 UR_01 .863 UR_02 .830 UR_03 .637 UR_05 .539 UR_07 .583 Cronbach’s alpha .552 .833 .648 .791 Accumulated percentage of variance explained 14.740 18.767 9.197 9.967

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. a. Rotation converged in 5 iterations.

b. Items with a loading of less than 0.4 are suppressed.