The Successor Decision in Family Businesses

(1)

The Successor Decision in Family

Businesses

Master thesis

Martine A.M. Lohues

University of Groningen

MSc Business Administration: Small Business and

Entrepreneurship

(2)

4.5 Analysis ...74 4.6 Research quality...75 5. Results...77 5.1 Propositions ...77 5.1.1 Proposition 1...77 5.1.2 Proposition 2...77 5.1.3 Proposition 3...79 5.1.4 Proposition 4...79 5.1.5 Proposition 5...80 5.1.6 Proposition 6...81 5.1.7 Proposition 7...82 5.1.8 Proposition 8...82 5.1.9 Proposition 9...84 5.1.10 Proposition 10...85 5.2 Selection criteria ...86 5.3 Notations...88 5.3.1 Timeframe...88 5.3.2 Information sources...89

5.3.3 Successor decision and conflicts...89

(4)

5.3.6 Succession candidates...91 5.3.7 Age differences ...91 6. Conclusions ...93 6.1 Propositions 1 ...93 6.2 Proposition 2...93 6.3 Proposition 3...93 6.4 Proposition 4...94 6.5 Proposition 5...94 6.6 Proposition 6...94 6.7 Proposition 7...94 6.8 Proposition 8...95 6.9 Proposition 9...95 6.10 Proposition 10...96 6.11 Proposition 11...96 6.12 Proposition 12...96 6.13 Proposition 13...96 6.14 Proposition 14...96 6.15 Proposition 15...97 6.17 Proposition 17...97 6.18 Proposition 18...97 6.19 Proposition 19...97 6.20 Proposition 20...97 6.21 Proposition 21...97 6.22 Proposition 22...98 6.23 Proposition 23...98 6.24 Proposition 24...98 6.25 Proposition 25...98 6.26 Proposition 26...98 6.27 Proposition 27...98 6.28 Proposition 28...99

7. Research limitations and suggestions for future research ...100

(5)

7.2 Research suggestions...100

Literature ...102

Appendix 1: Successor attribute importance ratings ...114

Appendix 2: Successor attribute importance ratings ...115

Appendix 3: Most influential selection criteria and selection criteria influence ratings...116

Appendix 4: Why does a part of the family business population use ‘gender’ as a selection criterion?.117 Appendix 5: Different types of commitment...119

Appendix 6: Variables between ‘contingency variables’ and successor attribute ratings ...120

Appendix 7: Business respondent questionnaire ...121

Appendix 8: consultant respondents questionnaire...125

Appendix 9: Mr Aalberts...129 Appendix 10: Mr Baas...135 Appendix 11: Mr Cornelis ...143 Appendix 12: Ms Dalen...151 Appendix 13: Ms Evertse ...157 Appendix 14: Mr Fonse...165 Appendix 15: Mr Graaf...170 Appendix 16: Mr Hartman ...180

(6)

Summary

The literature on employee selection is dominated by the psychometric selection paradigm, which prescribes how businesses ideally select their employees. However, severe critiques have been addressed to the paradigm’s assumptions. Taylor (2008) therefore calls for research into how non-bureaucratic businesses actually select their employees. This research paper addresses his call for family businesses and provides an answer to the question: How do family businesses select their leadership successor(s)?. Twenty-one propositions have been formulated using family business literature. Using eight interviews with six members of six family businesses and two consultants, these propositions have beentested, improved and completed. Finally, twenty-eight propositions are provided which show among others that: family businesses differ in how they select their successors; family businesses do not pass the steps formulated in the

psychometric selection paradigm; who makes the final successor selection decision varies among family businesses; the majority of selectors do not formalize selection criteria; a minority uses assessments to test whether candidates are suitable successors; many selectors prefer a successor that is member of the family; perceived difficultness of the decision is determined by several variables; shared leadership is a popular succession option; the urge of incumbents to meet feelings of equality may lead to specific behaviours; the selection criteria selectors use varies between family businesses; the period within which the succession decision is made can be very long; several selectors fund their decision on tacit knowledge; the majority of successor

(7)

Preface

As part of the master Small Business and Entrepreneurship of the University of Groningen, I present you my master thesis ‘the Succession Decision in Family Businesses’. The link between family businesses and me is not difficult to make; since my early years, I am involved in the business which my parents have founded in 1985. When I was a child, the fact that my family was running a business seemed so straightforward that I could hardly imagine families who did not live ánd work together. However, when I grew up, I started to learn how other families earned their living and realised that owning and managing business as a family is not a matter of course.

For my masters, I choose to study a subject for which I have feeling for: small businesses and entrepreneurship. In the first week of my masters, a short movie was shown about a father and his two sons who once worked together in their family business. However, the father had made the unfortunate decision to split the business in two entities when succession was brought up and he was not able to select one of his sons as his successor. As a result of this decision, the profits of the business dropped and the business became near to bankruptcy. This case made me think about family business succession and triggered me to write my master thesis on family business successor decisions.

Although my active participation in the family business has diminished in the last few years, the proud Ifeel for the business is still present. I would like to take this opportunity to wish my parents and my brother, who a few years ago decided to take over the business,all the luck in the world.

Acknowledgements

First of all, I would like to thank my supervisor, Mr Zwart, who guided me during the

development of my master thesis, gave me the incentives to persevere, put me back on the right track when I strayed off, provided me with useful feedback and kept his patience with me. In addition, I would like to thank Ms Croonen for being my second supervisor.

I could not have completed this thesis without Mr Aalberts, Mr Baas, Mr Cornelis, Ms Dalen, Ms Evertse, Mr Fonse, Mr Graaf and Mr Hartman, who offered me their time, welcomed me with enthusiasm and provided me with useful information on their successor selection processes. I would like to give special thanks to Mr Baas, for his support in my search for research

(8)

I want to thank my friends, family and family-in-law for thee love and joy they have provided during my master. I would like to give special thanks to my six friends from high school, Jolanda, Jorien, Anet, Anna, Jorien and Ariska, to my sister Marlieke and my brother Mark, my cousins Marije en Annemarie and my sister-in-law Roselien and brother-in-law Mark.

I am grateful for my boyfriend Sander, who supported me in the last one and half year, provided useful comments on my work, gave me a sense of perspective, confidence and joy and did not lose his believe in me.

Last, but certainly not least, I would show my gratitude to my parents Ben and BinyLohues, who worked hard and enabled me to finish my Bachelor and Master, who made me persistent,

(9)

1. Introduction

1.1 A dominant paradigm and its critics

HRM and organizational psychologyscholars have amassed extensive knowledge on employee selection in public and private organizations (Gatewood, Field & Barrick, 2008; Heneman, Judge & Kammeyer-Mueller, 2012; Noe, Hollenbeck, Gerhart & Wright, 2006; Schmitt & Chan, 1998; Smith & Robertson, 1993; Thorndike, 1949). The vast majority of this knowledgebuildson what is referred to as the ‘psychometric selection paradigm’, an approach to personnel selection that finds it origins in the late nineteenth and beginning of the twentieth century in the bureaucratic model of organization and scientific management and that is dominated by organizational

psychologists. The paradigm has dominated selection research from its inception on andhas been preached ‘with an almost unique authority and consistency’ (McCourt, 1999, p. 1015), as the practices drawn from it are said tofoster organizational performance and organizational efficiency (e.g. Huselid, 1995; Wright & McMahan, 1992).

The psychometric selection paradigm starts with the assumption that any business in search for an employeeshould select that job candidate who is predicted to perform thetasks belonging to the vacancy the best. Stated differently,on the one hand there is a job consisting of tasks that requires a jobholder who possesses specific knowledge, skills, abilities and other characteristics (KSAOs) in order to be able to perform thetasks of the job. On the other hand, thereare job candidates who possesscertain KSAOs whichmatch or do not match the required KSAOs. It is the companies’ task to select the candidate who possesses the required KSAOs the most. To this end, an organization has two instruments at its disposal; job analysis, whichassists in the identification of job tasks and required KSAOs;and selection methods, which assists in testing how much of the required KSAOs candidates possess.

In the last two decades, however, severe critiques have been voiced to both the theoretical assumptions and the practical usefulness of the psychometric selection paradigm(Cassell, Nadin, Gray & Clegg, 2002; Derous & de Witte, 2001; Herriot, 1992; Jewson & Mason, 1986;

(10)

and jobs and persons are stable entities (McCourt, 1999; Derous & de Witte, 2001; Rees & Doran, 2001). However, opponents (McCourt, 1999; Derous & de Witte, 2001; Rees & Doran, 2001) explain that what can be considered as ‘good practices’ varies for different situations as a result of contingencies, that stakeholders insurmountable have and should have personal

influence on selection decisions, that being successful in a job requires more than performingjob taskswell (it includes for example job satisfaction and seniority), that organizational members insurmountable behave irrational and that jobs and persons change in today’s fast changing environment.

Psychometric selection advocates address organizational’ disobedience to apply the paradigmin practice by a lack of promotion of the advantages of the paradigm and the ‘fact’ that organizational members are influenced by ‘old practices’ instead of ‘science’ (Guion, 1989, as cited in McCourt, 1999). These arguments, however, can easily be refuted, as the paradigm has crystallized in its current form around 1945 alreadyandhas been promoted ever since (McCourt, 1999). For example, Thorndike (1949) devoted one full chapter of his textbook on personnel selection to ‘selling’ the model. Nevertheless, these attempts have been of no avail as‘the field’is adamant in its refusal to employ the paradigm’s practices. It is more likely that its refusal on part of organizations is the direct result of its intrinsic flaws (McCourt, 1999; Taylor, 2006). In this research paper, we react to the criticisms that have been directed to the paradigm.

1.2 Taylor’s call for elaboration

Taylor (2006) argues that employee selection is a managerial practice that has effects beyond the boundaries of the organization, with particular implications for social justice. However,

organizations find it difficult to orient their selection practices to social justice (Taylor, 2006). Hence, a certain outside interference in organizational selection processes is desirable, in order to bring about selection processes that are fairer to candidates. Unfortunately, the current

psychometric selection paradigm has not succeeded in this, as showed above. Therefore, the development of a new paradigm is required.

(11)

researchers to elaborate on actual selection practices in those businesses. He pleads for the analysis of current selection reality using empirical gathered data on ‘real’ selection practices of small and emerging companies. Based on this analysis, new best practices can be developed that are better groundedin selection reality and that more likely appeal to and will be employed by small and growth businesses.In this research paper, we answer to Taylor’s call toelaborate on selection reality and to make the first steps in developing a reality-grounded selection paradigm. 1.3 The influence of family involvement on business behaviour

The vast majority of businesses worldwide face considerable influence of a family,whether throughout ownership, management and/or governance(e.g., La Porta, Lopez-de-Silanes &Shleifer 1999).For example, between 26 and 90 per cent of Dutch businesses is considered a family business, depending on the definition at hand. Applyingone such definition (a business is a family business when more than 50 per cent of shares is held by one family, one family has decision-making influence on business strategy or succession decisions and the majority or at least two members of the business management are from one family), family businesses account for 49 per cent of Dutch employment and almost 53 per cent of Dutch GDP (Nyenrode.nl). In addition, depending on the definition of family business used, between 19 and 92 per cent of the American businesses is considered a family business, accounting for 15 till 59 per cent of workforce and 12 and 49 per cent of GDP (Shanker & Astrachan, 1996).Moreover, family control is also found in large businesses; 35 per cent of all Fortune 500 businesses are family-controlled (FFI.org/looking/facts.html, out of Dyer, 2003).

Unfortunately, the commonality of family businesses is not recognized in the general management literature (Dyer, 2003; Schulze, Lubatkin, Dino & Buchholtz, 2001). For example, in a content analysis of issues of the Academy of Management Review (vol. 25(2) to vol.

27(1))and the Academy of Management Journal (43(5) to vol. 44(6)),Dyer (2003) found that only one of the 128 articles used ‘family’ as a variable to analyse firm behaviour. Similarly, in a small review of textbooks on strategy, organization theory, organizational behaviour,

organization development and human resource management, Dyer (2003) concluded that only one textbook mentioned the phrase ‘family business’, in only five of its 545 pages.

(12)

encourages family members to be considerate of each other (Schulze et al., 2001). Often, parents see the success of children to be inextricably bounded with their own success (Dyer, 2003). As a result, who you are is more important than what you do (Dyer, 2003). Thus, businesses wherein family values impact behaviour will be different to businessesin which such impact is absent. Finally, strong feelings of altruism and antipathy characterize relationships in family businesses (Dyer, 2003). These relationships can for example result in nepotism and serious interpersonal conflicts (Dyer, 1986). As relationships in an organization greatly influence governance structure, career development and leadership succession, the unique characteristics of family businesses relationships differentiate family business behaviour from non-family business behaviour.

In view of the above-described impact that family’ involvement can have on family business behaviour, Dyer (2003) argues that researchers should question whether current

management theories and practices, which are often developed using samples consisting of large public businesses, can be applied to family businesses. In addition, Dyer calls on researchers to include the family as a variable in their research, especially in research areas where variations in goals, values and relationships will likely influence research findings. This research paper responds to Dyer’s acknowledgement of family influence in family businesses and his callto include the family as a variablein organizational research.

1.4 Call for research into family business successor selection

Dyer (2003) explains that one of the research topics in which the inclusion of family as a

variable is required is leadership succession. Indeed, variations in goals, values and relationships between family and nonfamily businesses are expected to result in variances in leadership

selection processes. For example, Laakkonen and Kansikas (2011), Santiago (2000), Sten (2004), Tatoglu, Kula and Glaister (2008) and Ward (1987) argue that family businesses prefer family members to non-family members as their successors, in spite of the chance that siblings will possess characteristics that ‘fit’ their family businesses is as small as winning a lottery (Nicholson, 2008). Rather, the selection of family members as successors is a likelyresult of familial goals. In addition, families often value the equal treatment of family members highly (Martin, 2001; Poza, 1989; Ward, 1987; Ayres, 1990). For example, Ayres (1990) argues that the tendency of business owners to provide their children with equal parts of business ownership and management is a result of family business owners’ preference to treat siblings equally.

Furthermore, researchers argue that relationships among relatives often take precedence over maximum profits (Davis, 1968) and that the relationship between a potential successor and actively and non-actively involved family members is an important successor attribute

(Chrisman, Chua & Sharma, 1998, Sharma & Rao, 2000). In conclusion, these examples show that familial goals, values and relationships will likely influence business behaviour as a result of family involvement in the business. Thus, research into family business successor selection processes is warranted.

(13)

Breton-Miller, Miller & Steier, 2004; Dyer & Handler, 1994). First, Le Breton-Miller et al. (2004) conclude that scholars should conduct empirical and theoretical research into the selection of family business successors, as selection can have a large impact on succession effectiveness. In addition, Chrisman et al. (1998) attempted to identify which successors’ attributes family

business owners consider important, as we will discuss in more detail in paragraph 2.4. Although they succeeded in this, the authors did not survey the actual use of these characteristics as

selection criteria by family selectors when selecting a successor. Therefore, the authors call for research into actual usage of these attributes as selection criteria. Third, Dyer and Handler (1994) explain that we know very little about performance difference between family businesses.

Therefore, they ask for a comparison between selection processes in different family businesses.Finally, Dyer (2003) specifically asks for modifications in selection practices developed for non-family businesses,in order to apply them in family businesses.

Surprisingly, however, very little research has been conducted into family business successor selection (Astrachan & Kolenko, 1994; Brockhaus, 2004; Chrisman et al., 1998; Dyer, 2003; Le Breton-Miller et al., 2004; Sharma & Rao, 2000). First, le Breton-Miller et al. (2004) conclude that although family business scholars have paid considerable attention to subjects such as characteristics of the incumbent, succession planning and the presence of a board of directors, they have largely neglected succession selection variables such as selection of the CEO and TMT (top management team), design of a formal and legitimate selection process, final selection criteria, selection committee (jury, rules) and person and firm fit. The authors found that these variables were mentioned the least in succession literature. The authors explain that ‘the

selection process is neveraddressed in the empirical literature, and only very rarely in theoretical work, while it isflagged as critical in fully one third of the anecdotal articles’ (p.319).

In addition, Chrisman et al. (1998) explain that desirable successor attributes and who is considered the most desirable successor has hardly been researched. In addition, Astrachan and Kolenko (1994) concluded that writers and researchers have overlooked human resource activities associated with recruitment, selection, development, compensation and performance evaluation in family businesses. They explain that the identification and validation of these practices in family business is lacking.

(14)

1.5 Research goal, research question and research method

This research paper is a response to McCourt’s (1999) and others’ critics to the psychometric selection paradigm, Taylor’s (2006) call for elaboration on selection reality and the development of a new selection paradigm that is able to guide selection practices, Dyer’s (2003)

acknowledgement of the influence of a families’ involvement in family business and other researchers’ ask for research into family business successor selection. Our research goal is to gain understanding of how leadership successor(s) are selected in family businesses. The research paper will provide an answer to the question:

How do family businesses select their leadership successor(s)?

Dictated by the severe criticism on the psychometric selection paradigm, we will not use theframework as a point of reference. Instead, we respond to Taylor’s (2006) call for a more realistic approach to selection. Therefore, we will elaborate on how family businesses actually select their leadership successors. Consequently, our research paper is descriptive in nature.

Instead of focusing on one part of the decision-makingprocess, like several other researchers didbefore (e.g. Ayres, 1990; Chrisman et al., 1998; “Family Firms Lax,” 2008; Handler &Kram, 1988; Ibrahim, Soufani & Lam, 2001; Kets de Vries, 1993; Laakkonen & Kansikas, 2011; Lansberg, 1988; Levinson, 1971; Martin, 2001; Sharma and Rao, 2000; Sharma, Chua & Chrisman, 2000; Sten, 2004; Tatoglu et al., 2008; Ward, 1987; Williams, 1984 in

Lansberg, 1988),we zoom out and picture the complete selection process. As far as we know, such attempts have not been made or have not been successful before, as our literature research did not yield one article that presents research into the entire process. Rather, research efforts have been directed atsmall parts ofthe process, often only touching upon the topic when addressing other research subjects. For example, Chrisman et al. (1998) and Sharma and Rao (2000) researched successor attributes that family business owners consider important, Le Breton-Miller et al. (2004) discussed the topics that need attention in successor selection,

Tatoglu et al. (2008) addressed the identity of the person who selects the successor, Sharma et al. (2000) discussedthe development of explicit succession criteria and Dumas (1990) shortly points to an earlier unpublished research in which it was indicated that family-owned businesses do not use evaluation tools to test successors’ abilities.Nevertheless, we will use these bits and piecesin our exploration of successor selection reality.

We expect differences between family businesses selection processes, as family

(15)

We start our examination of successor selection reality with an extensive literature review in order to make use of earlier research contributions and to focus our attention to several main topics. Although no efforts have been madeto record the entire successor selection process, earlier researchers have left bits and pieces on the successor decision process. These pieces will be used toformulate propositions that describe the successor selection process. Then, we will preliminary test the accuracy of the propositions using six interviews with family members of six different family businesses and two interviews withtwo family business consultants. Making using of the interview data, we will adapt and fill the developed propositions were needed. Finally, an answer to our research question is provided in the form of several final

propositions.Future researcherscanthen test these propositions using larger sample sets, deduce good practices from it and build a new paradigm. This paradigm is grounded in reality and thus will be better able to guide organizational selection processes.

1.6 Definitions

1.6.1. Family business defined

Many family business researchers (e.g. Chrisman et al., 2005; Dyer, 2003; Westhead & Cowling, 1998) have emphasized the need for family business researchers to provide a clear definition of family business in order to facilitate the comparison of research results (Westhead & Cowling, 1998). Indeed, such a definitional pause is required as a universal definition of family business does not exist (Chua, Chrisman & Sharma, 1999; Littunen & Hyrsky, 2000; Shanker &

Astrachan, 1996; Westhead & Cowling, 1998). For example, Westhead and Cowling (1998) showed the large variety of family business definitionsthat researchers have adopted. Deploying these various definitions, the authorsfound that the percentages of businesses that was considered to be a ‘family businesses’ in their sample varied from 15 to 81 percent. Therefore, they

recommend researchers to explain and justify their definitions. In this view, we shortly discuss the definition of family business employed in this research paper.

(16)

Such a differentiation between family and non-family businesses based on behaviour is emphasized by Chua et al. (1999), who define a family business by its behaviour:

A family business is a business 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 in a manner that is potentially sustainable across generations of the family or families. (p. 25)

Thisbehaviour-based definition is favourable because it focuses on the ‘essence’ of the family business (Chrisman et al., 2005): that what truly separates family businesses from non-family businesses and justifies investigation of family businesses. Different from other definitions, which ‘use’ components of involvement to define family businesses (through management, ownership and/or governance), this description does not prescribe how a family should be involved in a business to qualify as a family business. Indeed, although family involvement in a business is required for a business to be qualified as a family business, the components by themselves do not necessarily influence business behaviour. Stated differently, family involvement in a business is a necessary but not a sufficient requirement for a business to be considered a family business. It is the behaviour of a business that needs to be influencedby family involvement in order for a business to be qualified as a family business (Chua et al., 1999). Again, the above provided definition centres this behaviour.

A second advantage of the definition is that it allows for variations in types and amounts of family involvement and thus for family business heterogeneity. Indeed, a family can be involved in a business through ownership, management, governance control and/or trans-generational succession in various extents (Chua et al. 1999; Klein, Astrachan & Smyrnios, 2005; Westhead & Cowling, 1998), while still behave as a family business.In light of these advantages, the above-stated definition will be used in this research paper.

A drawback of the provided definition, however, is its subjective nature. Indeed, the classification of businesses in family and non-family businesses would likely rely on the

subjective judgments of researchers. The measurement of for example ‘intentions’, ‘vision’ and ‘potential sustainability’ is ambiguous and multiple interpretable. Nevertheless, because of the exploratory nature of this paper and the use of a small sample size to test our propositions, we think some subjectivity is ‘allowed’. Indeed, given our small sample size, we do by no means have the pretention to provide statistically proven hypotheses. However, subsequent researchers testing our propositions using a large sample set and aimed at obtaining statistically significant results may need to operationalize the definition.

(17)

the business to achieve that desired future state. In addition, the ‘dominant coalition’ refers to the organization’s powerful persons that control the organizational agenda. Furthermore, the

definition holds the phrase ‘intention’ because a business that tries but fails to pursue and shape the vision of a dominant coalition still qualifies as a family business. Moreover, the definition does not specify the ‘identity’ of the dominant coalition. Therefore, as discussed earlier, the dominant coalition can be either owning or managing persons, as long as they control the organizational agenda and are controlled by members of the same family or a small number of families (Chua et al., 1999). Finally, to clarify the definition, we add the following description of Chua et al. (1999):

Simply stated, if (1) a family-controlled dominant coalition has shaped the vision that is being pursued or intends to shape and pursue a new vision and (2) the intention

orbehaviour is potentially sustainable across generations, the business is a family business. (p. 26)

1.6.2 Family business succession defined

As opposed to family business researchers’ discord on a definition of family business,

researchers are rather unitedon what succession of such a business constitutes. Many definitions consist of some combination of the components, ‘actions and events’, ‘transition’, ‘leadership and/or ownership’ and ‘family members’ (e.g. Hutson, 1987; Le Breton-Miller et al., 2004; Morris, Williams, Allen & Avila, 1997; Schwendinger, 2011; Sharma, Chrisman, Pablo & Chua, 2001). We concur in this structure and refer to family business succession as:

The actions and events by which leadership atthe top of a firm is transferred from one family member(s) to (an)other family member(s).

(18)

Finally, family business succession success is determined bysatisfaction of family members with the succession process and firm performance after succession (see Morris et al., 1997 for the operationalization of these concepts).

1.6.3 Leadership defined

For a useful description of ‘leadership’, we turn to the general management literature, as family business researchers do not provide a useful description of who or what position is actually succeeded (see for example Le Breton-Miller et al., 2004; Dyer, 2003; Handler, 1990) when succession takes place. To this end, we will use Jones’ (2013) textbook;‘Organizational Theory, Design and Change’.

We have visualized Jones’ (2013) understanding of organizational top-management hierarchy in figure two. As we can see, the shareholders are in top of the hierarchy, followed by the (chair of the) board of directors, corporate managers, divisional managers and functional managers. The board of directors is the representative of a company’s shareholders and act as trustees to protect the interests of shareholders. In this role, it has the legal authority to hire, fire and discipline corporate managers and to monitor the way corporate managers use organizational resources. The chair of the board of directors is the ‘manager’ of the board of directors and, according to Jones (2013), the person with the most authority in the organization.

Family business leadership succession Outside family Within family Generation-Neutral Between generations Extended family Nuclear family Extended family Nuclear family

(19)

After the (chair of the) board of directors, Jones (2013) argues that it is the Chief Executive Officer (CEO) who has the most authority (Jones, 2013). The CEO is the head of corporate management and the person that is ultimately responsible for setting organizational strategy and policy and the allocation of resources (Jones, 2013). It has received this authority from the board of directors. A company’s CEO can exert influence on organizational effectiveness and decision making in five principal ways (Jones, 2013):

1. She is responsible for setting the organization’s goals and designing its structure. 2. She selects key executives to occupy the topmost levels of the managerial hierarchy. 3. She determines top management’s rewards and incentives. 4. She controls the allocation of scarce resources such as money and decision-making power among the organization’s functional areas or business divisions. 5. Her actions and reputation have a major impact on inside and outside stakeholders’ views of the organization and affect the

organization’s ability to attract resources from its environment. (Jones, 1993, p. 61-62) Shareholders

Board of Directors

Senior Vice Presidents and Vice Presidents

General Managers or Divisional Managers Chair of board of directors

Chief Executive Officer (CEO) Chief Operating Officer (COO) Executive Vice Presidents

Functional Managers Ownership Trusteeship Corporate management Divisional management Functional Management

(20)

According to Jones (2013) the CEO and the chair of the board of directors is often the same person (CEO duality). As opposed to Jones (2013), several researchers have argued that it is the CEO who has the ultimate legal authority and responsibility (see for example Vancil, 1987) and who is the most powerful of the power centres (Brady & Helmich, 1984).

Whoever the person with the most authority might be, in this paperwe refer to the leader at top of the firm as a person or persons who influence(s) organizational effectiveness and decision making in the five above mentioned ways. It is possible that this person (or persons) also perform the tasks of the chair of the board of directors, although not necessarily. We expect it to be possible to identify such a person in every family business. Family business members may refer to this person as ‘CEO’, although not necessarily. Indeed, different organizations have provided different titles for this function (Oliver, 2002). Nevertheless, in order to prevent

confusion we focus on the ability of a person to influence organizational effectiveness and decision making in the above mentioned ways rather than the title attached to such a person. In the remainder of this research paper, we will name the person that have influenced organizational effectiveness and decision making up to the succession moment and for whom a successor needs to be selected the ‘incumbent’, while we will name the to-be-selected person, who in the future will influence organizational effectiveness and decision making and who succeeds the incumbent the ‘successor’ (thereby conforming to the family business succession literature). Although we will sometimes refer to this person in singular, it is also possible that there is more than one incumbent for which more than one successor will be selected.

1.7 Research structure

We will continue our research paperwith a more thorough discussion of the psychometric selection paradigm and the criticism addressed to it in chapter two, in order to develop a better understanding for our decision to departure from the model. This chapter should be seen as a side-road, after which the main part of the paper starts. Paragraph 2.1 includes a more detailed discussion of the psychometric selection paradigm, including elements of job analysis, selection and employment, KSAOs and selection methods. In chapter 2.2, the criticism to the paradigm will be discussed in depth, thereby using McCourt’s (1999) article as a guide. In chapter 2.3, we address Taylor’s (2006) request for elaboration and his theoretical foundation for it, for which he used Archer’s (1988, 1995, 2000) and Reed’s (2005) critical realist framework.

We then start the ‘main’ part of our research paper in chapter three. This chapter discusses theextensive literature review we conducted on successor selection practice. The chapter includes propositions that will be tested further on. In paragraph 3.1, we discuss how we conducted the extensive literature review.Then, in paragraph 3.2,the selection process will be ‘sketched’. The process will be pictured using propositions on among others selectors, use of selection criteria and selection methods, selection difficulties andfeelings of equality. Each subparagraph includes one or two propositions on a subtopic. Paragraph 3.3 then zooms into the selection criteria, as the literature is veryextensive on this subject. First, to facilitate the

(21)

paragraph 3.3.1. Then, in paragraphs 3.3.2, each selection criterion will be discussed separately. In paragraph 3.3.3, we discuss why usage of selection criteria varies between family businesses, introducing the so-called ‘contingency variables’.

Chapter four includes a discussion on how the proposition formulated in chapter three will be tested. In paragraph 4.1, we position our research ina research typology. In paragraph 4.2, we discuss how we collect research data. Then, in paragraph 4.3, we introduce our respondents. In addition, we present the concepts, measures and research questions in paragraph 4.4, after which we will discuss how we will analyse our research data (paragraph 4.5). Finally, in paragraph 4.6, we discuss how we controlled the quality of our research results.

Chapter five presents the results of our research. In paragraph 5.1 and 5.2, we present the results to the propositions we have formulated in chapter three. In paragraph 5.3, we present the results that could not have been coupled to a proposition.

(22)

2.A paradigm and its decline

2.1. Psychometric selection paradigm explained

In paragraph 1.1, we shortly introduced the psychometric selection paradigm and the criticism directed to it. However, to really understand the paradigm, its criticism and our motives to deviate from it, a more extensive explanation of the model is required. Therefore, this paragraph provides a more detailed explanation of the model, using three normative textbooks on employee selection: ‘staffing organizations’ (Heneman et al., 2012), ‘human resource selection’ (Gatewood et al., 2008) and ‘human resource management: gaining a competitive advantage’ (Noe et al., 2006). The last textbook has been used in a human resource management coursethat was part of the bachelorBusiness Administration of the University of Groningen (2008). Although the three books differ somewhat in focus anddepth, theircontent is rather similar. The following

description is a combination and brief summary of the three books.Remember that these books are prescriptive in nature and present the psychometric selection paradigm. As such, they include descriptions of how selection should be employed in organizations, rather than how selection actually isemployed.

2.1.1 Selection and HRM

As discussed in the introduction, selection is ‘the process by which an organization attempts to identify applicants with the necessary knowledge, skills, abilities and other characteristics (KSAOs) that will help it achieve its goals’ (Noe et al., 2006, p. 725). Selection is one of the subsystems of human resource management (HRM) and interacts with other HRM systems such as job analysis, training and compensation (Heneman et al., 2012; Gatewood et al., 2008; Noe et al., 2006). As with other HRM systems, selection strategy is shaped byan organization’s strategy, its human resource strategy and the interaction of these two, which in their turn are shaped by an organization’s mission, goals and objectives. Stated differently, the strategy of a well-developed staffing system within an organization is tuned to the strategies of the other HRM systems, the overall HRM strategy and the strategy, mission, goals and objectives of the organization as a whole. As a result, the policies and programs of the selection system result from the interaction of these strategies, mission, goals and objectives. The interdependencies of the above-mentioned entities are visualized in figure three (Heneman et al., 2012).

(23)

2.1.2 Job analysis

As described in the introduction, businesses that are in searchfor a person to fill a job vacancy should start with developing a ‘picture’of how the perfect jobholder (the person who holds the job)looks like. A job analysisfacilitatesin the development of such a picture (Gatewood et al., 2008; Heneman et al., 2012).Job analysesstart with the collection of information about the open vacancy, making use of job analysis methods such as job analysis interviews and job analysis questionnaires (Gatewood et al., 2008; Heneman et al., 2012). These activities produce a list of job tasks, duties, responsibilitiesand contexts laid down in a job description. Subsequently, the job descriptioncan betranslated into job specifications1; a list of specific knowledge, skills, abilities and other characteristics (KSAOs) that predict performance of job tasks (Gatewood et al., 2008; Heneman et al., 2012). Knowledge refers to a body of information, usually of a procedural or factual nature, that is necessary for successfully performing a task (Gatewood et al., 2008; Heneman et al., 2012; Noe et al., 2006). In addition, a skill refers to an individual’s level of proficiency in performing a specific task (Gatewood et al. 2008; Noe et al., 2006). A skill is not an enduring characteristic of a person, but depends on experience and practice (Heneman et al., 2012). Conversely, ability is a more general and enduring trait or capability of an individual, useful for performing a range of tasks (Gatewood et al., 2008; Heneman et al., 2012; Noe et al., 2006). An ability, as opposed to a skill, is less likely to change over time and is useful for a wide set of tasks in different jobs (Heneman et al., 2012). ‘Other characteristics’ is a

Figure 3: Staffing & Strategy (adapted model from Heneman et al., 2012, p. 22) Organizational

mission, goals and objectives Organizational

strategy HR strategy

(24)

category for leftover factors that do not fit into one of the other categories. An example of such other characteristics is ‘possession of a driving license’ (Heneman et al., 2012).

Finally, job specifications are translated into selection methods, which measure how much of the required KSAOs a candidate possesses. These selection methods can be developed in-house or bought ready-made from anoutside supplier. Paragraph 2.1.4 includes a short discussion of different selection methods. Picture four visualizes the job analysis process.

2.1.3 Recruitment

Recruitment usually follows job analysis and precedesselection. Recruitment is occupied with the identification and attraction of applicants for potential employment. It comprises those activities that influence the number and type of individuals who apply for positions within organizations. A well-developed recruitment system ensures the attraction of a sufficient amount of qualified applicants and increases the probability that the selected applicants will accept a job offer (Heneman et al., 2012; Gatewood et al., 2008). As discussed in the introduction,this research paper focuses on the selection of family business successors. Therefore, we will not explain this system in more detail and close this paragraph with the remark that recruitment activities affect selection activities and vice versa.

2.1.4 Selection

After a job is analysed, described in terms of its tasks, duties, responsibilities and contexts, specified in terms of required KSAOs and candidates are recruited,selectors need to assess which candidate is best qualified for the job. To this end,information oncandidate’s KSAOs need to be collectedusing one or a few selection methods (Gatewood et al., 2008; Heneman et al.,

2012).These selection methods should measure those KSAOs that have been identified in job analysis and that are necessary for performance on the job. Although this condition seems straightforward, many selection devices still measure too broad KSAOs instead of the particular ones required for job performance (Gatewood et al, 2008). In addition, a method should be able to differentiate among applicants. Since differences between candidates are sometimes subtle, selection methods should be sensible to small differences (Gatewood et al., 2008).

Translated into

Produces Translated into

Job Analysis Method E.g. job analysis

interviews Selection methods E.g. Job Interviews and Ability Tests Job Specifications Knowledge, Skills, Abilities and other Characteristics (KSAOs) Job Description Job Tasks, Duties, Responsibilities and Job Contexts

(25)

Examples of selection methods are collection of biographical data, reference checks, structured and unstructured job interviews, personality tests, cognitive ability tests, work-sample tests and drug tests (Gatewood et al., 2008; Heneman et al., 2012; Noe et al., 2006).For example, a

structured interview is a standardized method of assessment that needs careful preparation and is based on job specifications, whereas the unstructured interview is an often unplanned,

unstandardized method, resting on the interviewer’s intuition. Since the first one scores higher on validity, it is the favoured selection method. In addition, cognitive ability tests assess abilities involved in thinking, including, for example, memory, reasoning and verbal and mathematical abilities. These tests can assess both these specific abilities and general mental ability and have shown to be among the most (if not the most) valid methods of selection.

Because of time and price limits, a business can usually employ one or a few selection methods (although usage of more than one is preferred). Therefore, the selector should make a choice which selection methods to use. This choice should be dependent onfor example a method’s quality, which is measured by reliability, validity, generalizability, utility and legality. Obviously, the higher a selection scores on these measures, the higher the quality and the more its usage is preferred. Reliability refers to the consistency of a performance measure. Reliable selection methods are free from error and are consistent over time and in different contexts. Second, validity can be defined as the extent to which performance on a measure is related to performance on a job. If the validity of a selection method is high, a high score on a method will translate in high performance on the job. Generalizabilityis ‘the degree to which the

validity of a selection method established in one context extends to other contexts’ (Noe et al., 2006, p.721). Utility is ‘the degree to which the information provided by selection methods enhances the bottom-line effectiveness of the organization’ (Noe et al., 2006, p. 229). Factors such as costs of the method are included in this last standard. Finally, legality refers to the conformance of selection methods with existing laws and legal precedents. A selection method should reduce the occurrence of discrimination, in order to minimize the change of legal prosecution (Gatewood et al., 2008; Heneman et al., 2012; Noe et al., 2006).

Table one provides an overview of several selection methods and their typical scores on standards of reliability, validity, generalizability, utility and legality.

Method Reliability Validity Generalizability Utility Legality

Biographical data High High/Low S.D. High S.D.

Reference checks Low Low Low Low S.D.

(Initial) Interviews S.D. S.D. Low Low Low

Personality tests High Low/n/a Low Low Low

Cognitive ability tests High Moderate/n/a High High S.D.

Work-sample tests High High S.D. High High

Drug testing High High High S.D. S.D.

(26)

After applicants are assessed, a company’s selector should decide which applicant to extend an offer of employment. This decision does not automatically result from the deployment of

selection methods. For example, when using multiple selection methods, candidates’ scores need to be combined. One such combination strategy is called ‘multiple regression’, in which the different predictor scores (selection method scores) are given a weight according to their importance to job performance (e.g. the result on a job interview may be judged less important for job performance than the result on a work-sample test) and a total score will be calculated using a special formula. This enables compensation when an applicant scored negative on one predictor and positive on another predictor (Gatewood et al., 1998; Heneman et al., 2012). In addition, the candidates’ scores have to beranked in some way. One of the possibilities to do this is the top-down selection approach, wherein applicants’ scores are rank-ordered from high to low. The person on top of the list is extended a job offer first. Then, moving to the bottom, other applicants are extended job offers until all positions are filled (Gatewood et al., 2008).

2.2 The criticism to the psychometric selection paradigm

In the last two or three decades, the psychometric selection paradigm has received severe criticism, primarily directed at its low take-up on part of private and public organizations andits assumptions(Cassell et al., 2002; Derous & de Witte, 2001; Herriot, 1992; Jewson & Mason, 1986; McCourt, 1999; Rees & Doran, 2001; Taylor, 2006). Although best practices that are preached by the paradigm might be useful for large bureaucratic organizations, they have shown to be less applicable to other sorts of companies (Taylor, 2006). This paragraph discusses the criticism addressed against the paradigm, its development alongside the development of the bureaucratic model of organisations and scientific management and how this development hasresulted in flawed assumptions, especially when applied in non-bureaucratic organizations. 2.2.1 Disobedience

As discussed in the introduction, one of the most important requirements of a paradigm developed for appliance in a field is its acceptance and usage among organizations (McCourt, 1999; Tsang, 1997). Therefore, the normative selection paradigm can only be considered ‘adequate’ when a majority of organizations has implemented it. However, many authors havereported that organizations refuse to employ the model in its absolute form (Cassell et al., 2002; Herriot, 1992; Jenkins, 1986; Jewson & Mason, 1986; McCourt, 1999; Rynes & Boudreau, 1986; Shackleton & Newell, 1991; Smith & Abrahamsen, 1992; Taylor, 2006) and that not only do small organizations disobey, large organizations also largely diverge from its ‘good practices’ (McCourt, 1999; Taylor, 2006). For example, Rynes and Boudreau (1986) surveyed 145 of the largest Fortune 1000 companies and concluded that most of these companies diverged radically from the paradigm’s good practices. Likewise, Shackleton and Newell (1991) found in a

(27)

correlation of-0.25 between a methods’ usage and its validity, meaning that if the validity of a selection method increases, its usage declines.

Advocates of the psychometric selection paradigm ascribe selectors’refusal to apply the paradigm to a lack of professional competence (McCourt, 1999), knowledge (Rowe, Williams & Day, 1994; McCourt, 1999; Rynes & Boudreau, 1986; Williams, 1993) or management

education (Amba-Rao & Pendse, 1985; Deshpande & Golhar, 1994; IPD, 1994). In addition, they argue that the paradigm’s advantages have not been promoted enough yet and that

organizational members are influenced by ‘old practices’ instead of ‘science’ (Guion, 1989, as cited in McCourt, 1999). These arguments can however simple be refuted, as the paradigm crystallized in its current form around 1945 already and has been promoted and taught ever since (McCourt, 1999). For example, Thorndikedevoted in 1949 already one full chapter of his

personnel selection-book to ‘selling’ the model. Nonetheless, the attempts have been of no avail, as the fieldrefuses to employ the paradigm in practice, as showed above. Rather, its refusal on part of organizational selectors is the direct result of intrinsic flaws of the paradigm(McCourt, 1999; Taylor, 2006). McCourt (1999) argues that organizational members are aware of these flaws and reject the model as a result.

2.2.2 Flawed assumptions

The psychometric selection paradigmhas developed alongside the development of the bureaucratic model of organizations and scientific management in the late nineteenth and beginning of the twentieth century. In the public sector, the late nineteenth movement, that stressed the importance of applying the bureaucratic model to governmental organizations, emphasized the need to select civil servants based on merits and the need to separate jobholders from jobs. Slightly later, the scientific management movement recommended applying its concepts of economy and efficiency in selection procedures in private organizations. In order to process the large numbers of applicants that would apply as a result, selection technology was needed that was widely applicable, impersonal and capable of distinguishing applicants based on physical and mental characteristics. Upcoming psychologists that had just developed tests to distinguish among physical and mental personal characteristics offered this technology

(McCourt, 1999). For example, around the turn of the 19thto the 20thcentury, Hugo Munsterberg developed work simulation tests for railway and telecom companies (Hothersall, 1984). From that time up, psychologists developed methods that enabled private and public organizations to select employees in a ‘scientific’ manner. In 1945, with the completion of the Assessment Centre method, the psychometric selection model crystallized in its current form and the paradigm did not change fundamentally since (see McCourt (1999) for an extensive overview of the history of the psychometric selection paradigm).

(28)

(flawed)assumptions are: (1) good practice is universal; (2) good practice is impersonal; (3) organizational performance is simply the sum of individual performances (McCourt, 1999) and (4) jobs and persons are stable entities (Derous & de Witte, 2001; Rees & Doran, 2001). These assumptions will be discussed separately in the subsequent paragraphs.

2.2.2.1 Good practice is universal

As opposed to many other management areas, the normative-prescriptive selection framework is almost completely insensible to contingency factors such as company size or technology,

whether companies have a personnel department or not and national culture (McCourt, 1999). Rather, the paradigm enforces companies of all shapes and sizes to follow the prescribed selection procedure, in order to maximize individual performance. However, this focus runs counter to what is seen as a truism in many other organizational research areas (such as company structure and leadership); that good practice is a contingent matter (McCourt, 1999). Several empirical studies have shown thatcontingencies influence selection procedures and ‘best practices’ (e.g. Pfeffer &Cohen, 1984; Roe, 1989). For example, Roe (1989) found that the societal context of selection determined the design of a selection procedure. Furthermore, Pfeffer and Cohen (1984) found that usage of internal labour markets is dependent on for example unionization. In addition, a business of a large size is more likely to have larger amounts of similar jobs, for which more jobholders are required. As a result, the costs of an extensivejob analysiscan be effective, as its costs can be spread among a large amount of jobholders. In contrast,a business consistent offive employees is probably unable to spread those costs. In such companies, less extensive job analyses might be more appropriate.In conclusion, the assumption that good practice is universal seems flawed.

2.2.2.2 Good practice is impersonal

In order to prevent organisational stakeholders from influencing selection decision in an attempt to defend their own interest, the psychometric paradigm dictates impartially selection of

(29)

(1987) argues that involving the line manager into the selection procedurefosters the success of an appointment. Therefore, we argue that the assumption that selection should be impersonal is flawed.

2.2.2.3 Organizational performance is simply the sum of individual task performances

As discussed in one of the preceding paragraphs, the psychometric selection paradigm is directed at predicting candidate’s performance on individual job tasks. Unfortunately, such emphasison job tasks neglects the impact that ‘non-functional’ factors can have on organizational success. For example, in order to preserve harmonious industrial relations, it may be desirable to select on basis of seniority (Macdonald, 1989). However, seniority might not directly predict individual job performance andtherefore be neglected in candidate assessments. Similar, family harmony is often one of the measures of family business performance (Churchill & Hatten, 1997). Therefore, selecting the ‘family-favourite’ or the firstborn son (when primogeniture is a family selection rule) for the family business successor position may be advisable, when it fosters family

harmony. However, family favourability or being first-borne does not directly lead to individual task performance andis therefore not assessed in psychometric selection procedures.

In conclusion, McCourt (1999) explains that the psychometric selection paradigm wrongfully assumes that organizational performance is simply the sum of individual

performance. Indeed, aspects such as job satisfaction, seniority, loyalty and person-organisation fit likely influence firm performance but are often neglected in the normative selection

framework. Thereby, the ignorance is another flaw of the psychometric selection paradigm. 2.2.2.4 Jobs and persons are stable entities

Finally, the psychometric selection paradigm assumes that jobs and persons are stable entities, which enables prediction of task performance. However, jobs and persons adapt to today’s fast changing environment (Derous, & de Witte, 2001). As a result, today’s job descriptions may not reflect tomorrow’s job tasks. Similarly, a candidate’scurrent KSAOs may not reflect future job performance. Therefore, the identification of specific job specifications and measurements of specific KSAOs is flawed in many organizations were flexibility is a key business objective (Derous & de Witte, 2001; Rees & Doran, 2001).

2.2.3 Why the psychometric selection paradigm is still advocated in selection literature

Considering the many flaws of the psychometric selection paradigm, an appropriate question is why the model is still advocated. According to McCourt (1999), the paradigm persists because (1) it is still appropriate for ‘pure’ bureaucratic organizations that process large amounts of applicants, (2) of the existence of a self-sustaining research agenda, (3) the paradigm has established a professional jurisdiction that dominates the selection ‘field’ (for example, the journals in which most selection research is published are controlled by advocates of the paradigm;see McCourt, 1999, p. 1014 for an overview how professional jurisdiction stimulates the perseverance of a paradigm) and (4) of the advantages that accompany paradigm

(30)

and economies of scales in the development of new instruments (McCourt, 1999). In short, one could argue that the psychometric paradigm has developed into a dogma.

2.3 Taylor’s alternative perspective: elaboration of selection practices

As discussed in the introduction of this research paper, selection practices in organizations need to be controlled to restrict the occurrence ofdiscrimination as much as possible. However, as a result of itsflaws, the current ‘dominant’ psychometric selection model has been unable to

control selection practices, as the above paragraphs have made clear. As a result, one could argue that selection practice is currently left uncontrolled, an undesirable state of affairs.

Fortunately, Taylor (2006) offers an alternative perspective to selection that

enablescontrol of organizational selection practices. To this end, Taylor (2006) proposes the study of employee selection by acknowledging organizational agents’ freedom to act within a framework of imposed conditions. To this end, heuses Archer’s (1988, 1995, 2000) framework, viewed through Reed’s (2005) interpretation. We first introduce Taylor’s perspective of this framework and provide a picture that visualizes the framework. We then discuss how Taylor adapts this framework to employee selection and how it functions as the starting point of our research paper.

Taylorexplains that previous ‘actors’ have developed cultural and structural conditions, within which ‘current’ individuals and groups behave freely, accepting and challenging the conditions (T1-T3, figure five). After a certain period of time, academic analyses, policy review and ‘agentic’ reflection take place (T3-T4). In this phase, current action within the framework of imposed conditions is mapped andreflected on. This reflection either results in a changed

(31)

Taylor applies Archer’s perspective to selection theory and practice. He explains that selectors today are influenced by selection structures that have been developed in previous times. Indeed, the psychometric selection paradigm is developed by previous actors in order to regulate

selection in bureaucraticorganizations (from the end of the nineteenth century till around 1945, T1-T2).Theparadigm, which is manifested through legislation, HRM techniques and social norms, functions as a structure that conditions selection practice today. However, within that structure, selectors behave freely, adapting psychometric selection procedures in a way that suits them (the period from around the first half of the twentieth century till now, T2-T3). Indeed, managers are no defenceless marionettes that thoughtlessly follow prescriptions dictated by the psychometric selection paradigm, nor are they controlling puppet-masters who behave

(32)

selection procedures are hardly ever implemented in practice in its exact prescribed form. However,organizations are not ‘anarchistic’ selectors as well. Instead of randomly picking their employees from the streets, organizations generally use some sort of selection method that is funded in the psychometric selection paradigm (Herriot, 1992). For example, more than half of the 300 British organizations in Mabey’s (1989) sample used aptitude tests for selection reasons, and all of them used interviews. Similarly, the majority of the organizations surveyed by Cassell et al. (2002) addressed equal opportunity concerns, yet in a less formalized and procedural manner than prescribed by the psychometric selection paradigm.

Finally, in the last three or four decades, researchers such as Collinson et al. (1990), Derous and de Witte (2001), Lupton (2000), McCourt (1999), Silverman and Jones (1976) and Taylor (2006) have started to elaborate on succession practice (T3-T4). For example, Collinson et al. (1990) and Silverman and Jones(1976), found that selection tools such as interviews are blurred through managerial politics, individual preferences and prejudices and organizational or local cultures. Unfortunately, however, these elaboration activities have focused on large organizations, as discussed in the introduction. Elaboration of selection practice in small and emerging companies has been practically absent(Taylor, 2006). Similar, reflection of selection practice in family businesses have been scarce and fragmented (Le Breton-Miller et al., 2004; Dyer, 2003; Astrachan & Kolenko, 1994). Furthermore, like small and emerging businesses, it is unlikely that analysis of selection practices in bureaucratic organizationscan be translated to family businesses, as family involvement influences selection behaviour substantially, as explained in chapter one. Therefore, Taylor (2008) calls for elaboration of selection practice in non-bureaucratic organizations. To this end, he instructs researchers to gather data by conducting qualitative empirical research. Using these data as input, researchers can analyse desirability of current selection action and test accuracy of current structures. If action is judged undesirable and structures inadequate, new best practices (i.e. structures) can be developed that guide future action in a favourable direction. Such a new structure is more likely to be taken up by non-bureaucratic organizations, as they are imbedded in their selection reality and thus provide those businesses with more realistic selection procedures.

In this research paper, we answer to Taylor’s call by making the first steps in the elaboration of actual successor selection practices in family businesses. To this end, we gather empirical data on selection theory.We provide future researchers a picture of how family

(33)

3.Successor selection

The empirical literature on family businesses successor selection is brief and fragmented, as already discussed in the introduction.Nevertheless, the bits and pieces in the literature are valuable input for the development of propositions that describe the successor selection ‘process’2. In this paragraph, we discuss the empirical literature on successor selection and present the propositions that together provide a picture of the successor selection process.

In paragraph 3.1, we shortly discuss the literature review that has beenconducted to identifythe ‘bits and pieces’on successor selection. In paragraphs 3.2,we present and discuss relevant literature and formulatetenpropositions that describe how family businesses likely select their leadership successors. Subsequently, in paragraph 3.3, we zoom into the selection criteria that are identified by researchers and whichwe expect family businesses to use when selecting a successor. This paragraph will be concluded with a discussion oncontingencies that possibly determine which selection criteria are used by which family businesses.

3.1 Literature review

In search of articles on family business successor selection, we made use of Business Source Premier, a popular business research database including more than 2200 full-text business journals. We used among others ‘family business/family firm/family-owned business’,

‘selection’, ‘successor’, ‘(selection) criteria/criterion’, ‘succession’, ‘successor’, ‘choice/choose’, ‘decision’, ‘sibling(s)’, ‘relationship(s)’, ‘family members’, ‘interest’, ‘candidate(s)’ and

combinations of these words as search terms. We checked off the ‘scholarly (peer reviewed) journals’ option, to restrict or search to peer reviewed articles. In addition, we used the search engine ‘Google’ to search in the scholarly journal ‘Family Business Review’, a journal oriented to family businesses, as Business Source Premier does not presents most of the articles published in this journal.

We have read and skimmedmore thantwo hundred articles in our literature review. Most of the articles that were judged ‘relevant’ include brief discussions of successor selection, while addressing other parts of the succession process. Second, as opposed to a discussion of the full ‘process’, all the reviewed articles discussparts of successor selection, such as use of explicit selection criteria, the person responsible for selection and selection planning. Third, the reviewed articles present all sorts of papers (e.g. research, advises, stories). However, for the development of the propositions, we haveonly used empirical contributions. Thereby, we develop an

understanding of how selection actuallyis performed as opposed to how selection should be performedandwe thusscreen our propositionsfrom the psychometric selection paradigm. Nevertheless, as a result of theparadigm’s dominance in selection literature, we do expect ‘our’reviewed articles to be impacted by the norms dictated in the paradigm in some way. For example, we expect the subject choice of family business scholars researching successor selection to be affected bythe subjects emphasizedinthe psychometric selection paradigm.

(34)

Unfortunately, thisinfluenceis the insuperable result of our choice to useliterature to formulate propositions (instead of for example reasoning). Nevertheless, we (partly) control for the paradigm’s influence by presenting our findings in the form of propositions, which are tested in the second part of our research and in subsequent research.

Finally, although our propositions are funded on empirical research,we sometimes discuss the non-empirical literature as well, for example to show the literature’s preoccupation with a certain topic. However, to repeat, this literature will not be used as an input to our propositions.

3.2The selection‘process’ 3.2.1 Selection variances

In a survey into 1060 small firms, Chrisman, Chua, Pearson and Barnett (2012) concluded thatfamily businesses are a heterogeneous business population.Consistent to this heterogenic view of family businesses, we found the ‘process’ by which family businesses select their successor to vary for different family businesses (see the subsequent paragraphs). For example, in some family businesses the incumbent makes the selection decision, while in other family businesses the board of directors make this decision (Laakkonen & Kansikas, 2011; Tatoglu et al., 2008; Williams, 1984, in Lansberg, 1988). Moreover, some family businesses select one successor, whereas others select more than one successor (Kets de Vries, 1993; Laakkonen and Kansikas, 2011; Levinson, 1971; Ward, 1987). In addition, some successor selectors consult family members before making a selection decision, while others do not (Tatoglu et al., 2008; Williams, 1984, in Lansberg, 1988).

Therefore, in contrast to what is dictated by the psychometric selection paradigm, we suggest that:

1. The ‘process’ by which family businesses select their successor(s) varies between different family businesses.

Note that the word ‘process’ is a term we simply picked to refer to (some of) the elements of the successor decision that will be discussed in the following propositions. We do not use this word to imply that family businesses consider the decision as a process consisting of different steps that need to be passed through. Such a notion is a typical ‘normative’ idea (see paragraph 2.1). More than that, we do not assume that family businesses consider the ‘decision’ an actual decision. Rather, we expect a part of the following propositions to be situation dependent.The subsequent paragraphs will indicate when this is the case.

The Successor Decision in Family Businesses