Shared leadership in public sector self-managing work teams: Rhetoric or reality?

(1)

Shared leadership in public sector self-managing

work teams:

Rhetoric or reality?

A study on the relationship between self-managing work teams and shared

leadership in public sector teams

Author: Fabio Batenburg Student number: 1551515

Thesis supervisor: Prof. dr. S.M. Groeneveld Second reader: Dr. B.S Kuipers

Date: June 22, 2020 Word count: 16.863

(2)

Abstract

In this study, the relationship between self-managing work teams (SMWT’s) and the emergence of shared leadership in public sector teams has been examined. The design of the study is nested. It is a single case study design in which the different SMWT’s in the municipal organization of Krimpen aan den IJssel have been compared. By doing so, variance in leadership behaviour can be ascribed to the team design. The main goal of the study was to find patterns between the different types of SMWT’s and their leadership behaviour. I have used qualitative interviewing as the main method of data collection. Two types of SMWT’s have been examined: monodisciplinary teams and multidisciplinary teams. Evidence has been found that the type of SMWT is affecting the degree to which shared leadership is present in the public sector teams. Contrary to theoretical expectations, this effect seems to be negatively influenced by the interdependencies (task interdependency, goal interdependency and stakeholder interdependency). No big differences have been found for an effect of the external dependencies (political control and accountability) on the relationship between self-management and shared leadership.

(3)

Introduction

The concept of self-management has been given more and more attention by scholars in recent years. In the last decade, academic literature on the topic of self-management and self-managing work teams has increased exponentially. When implemented correctly, a self-management design is believed to increase both employees’ wellbeing as well as the organizational performance (Stewart & Barrick, 2000). Yet, there are also some authors that show the possible downsides of self-management designs in organizations (e.g. Barker, 1993). But why would an organization consider to change working processes to a design that is centred around the principles of self-management?

Self-management is considered to be an anti-bureaucratic form of organizing. To cite Groeneveld & Kuipers (2014, p.3): “increasing demands and competing values force public organizations to introduce new work forms that veer away from rigid bureaucratic structures and procedures while remaining in control.” Flat organizational designs (e.g. self-management) make organizations more flexible and adaptable towards their -continuously changing- environment (Burns & Stalker, 1961; Pennings, 1987; McHugh, O’Brien & Ramondt, 2001). This is especially relevant in current public administration. For example, think of globalisation and the increasing role that (inter)governmental and multi-level networks have on the daily working processes of public organizations. Therefore, it could be argued that self-management designs could nowadays be more effective and/or efficient than the ‘static’ bureaucratic and hierarchical models. This makes it a relevant topic to study (see for example Lane & Wallis, 2009, p.117; McHugh, O’Brien & Ramondt, 2001, p.35).

In this thesis, I look into management as an organizational design choice and study self-management in relation to leadership behaviour. The focus is on self-self-management as an organizational design feature and not on self-management which can be categorized as individual or team behaviour. Self-management implies that teams have a high degree of autonomy in their daily working processes. This is contrary to hierarchical bureaucratic systems where work processes are merely fixed and demarcated. Perhaps, the contradiction between self-management as an organizational design feature and bureaucracies on the other hand is a bit harsh; there could be hybrid forms. However, the ability to manage work processes is expected to vary. The argument here is that teams in a bureaucratic context are often managed by a traditional manager or supervisor, whereas team members in a post-bureaucratic context have more freedom to shape their own work and/or leadership processes. Therefore, leadership

(5)

behaviour among team members is expected to be different in both approaches. In some post-bureaucratic organizations, much of the (traditional) management tasks have been delegated towards the team members. This causes the leadership dynamics and work routines of the team to be heavily influenced. It is relevant to study how these leadership mechanisms work in a self-management context. I will use the concept of shared leadership to measure leadership behaviour. Shared leadership is about the division of leadership tasks among (multiple) team members. The concept implies that leadership is carried out by the team as a whole (or by multiple team members) rather than by a designated individual (Carson et al., 2007, p.1219). Shared leadership complements the self-management approach as “it is the team itself that is in the lead” (Kuipers & Stoker, 2009, p.399). It is expected that the (shared) leadership behaviour of the teams is affecting its self-managing work processes (e.g. decision-making processes), either in a positive or in a negative way. Therefore, it is relevant to study how shared leadership behaviour is present in the different types of self-managing work teams (SMWT’s) as it could influence the teams’ work processes and performance.

Research question

In this thesis, I am going to examine the possible relationship between organizational models of self-management (self-organizing models) and shared leadership behaviour in the context of public sector teamworking. The aim of the thesis is to discover and unravel the mechanisms of shared leadership in such a way that we are able to position the concept in relation to self-management designs in a public sector context. The case study that is designed for this thesis focusses on municipal teams in the municipality of Krimpen aan den IJssel, the Netherlands. Interesting is that most of the teams have different team compositions and functional backgrounds. This I will refer to with the type of SMWT. Also, the literature on self-management identifies a few interdependencies (goal interdependency, task interdependency and stakeholder interdependency) and external dependencies (political control and accountability) that may affect the emergence of shared leadership in SMWT’s. These concepts are highly relevant as they are expected to influence team behaviour and therefore could affect the relationship between self-management and shared leadership. The thesis is structured around the following research question:How and under what conditions does the type of self-managing work team affect the emergence of shared leadership in public sector teams?

(6)

To answer the main question, firstly, it is necessary to review the academic literature on the topics of self-management and shared leadership. Secondly, I will identify the various types of teams as well as the interdependencies and dependencies that could affect the relationship between self-management and shared leadership. The variation will be on the independent variable, namely the type of self-managing work team. I study this in combination with the emergence of shared leadership in public sector teams.

To properly structure the thesis and to provide a good answer to the main question, the following sub-question (1) need to be answered first:

Is there a relationship between the type of self-managing work team and shared leadership?

In other words, can we observe differences between the various teams and their leadership behaviour? When this question is answered, we can examine how the different

interdependencies and dependencies affect the relationship between self-management designs and the emergence of shared leadership (sub-question 2).

Relevance

The study is relevant as the literature on High Performance Teams in the public sector teaches us that both self-management and leadership could be identified as conditions for teams in the public sector to reach high standards of performance (Kuipers & Groeneveld, 2016, pp. 22-25). Therefore, research on the relationship between these two components that shape High Performance Teams is relevant for the academic literature on teamworking and organizational performance. Adding up to this, Kuipers & Stoker (2009, p. 400) argue that we need a better theoretical and empirical understanding of team development and its relation to performance within self-managing work teams. I examine this relation from the perspective of shared leadership.

In line with Groeneveld & Kuipers (2014, p.4) I argue that we need a better understanding of the public value context of SMWT. In other words, we need a better understanding about how the public sector context influences the functioning of SMWT. Studying the variation in SMWT’s as a result of organizational design will provide us with some new insights about SMWT in a public sector context as well as their leadership behaviour. This part of variation in design is mostly underexposed in academic literature. I will try to fill this gap by connecting it to leadership behaviour among team members. So, this thesis can add to the public

(7)

administration literature by showing how a certain design type of self-management affects leadership behaviour in a self-management context.

As for the social relevance, this study will provide knowledge about the relationship between self-management designs and shared leadership in such a way that it could help managers that are willing to implement self-management as an organizational form into their organizations. The question that I have tried to answer is why certain leadership behaviours emerge in self-managing teams based on the design type. For implementation, it means that knowledge of this mechanisms is desirable to successfully implement self-management and/or shared leadership behaviour into public sector teams.

Structure

After this introduction, I will provide a theoretical framework in which I review the literature of self-management and shared leadership. Thereafter, theoretical expectations about the relationship between self-management and shared leadership are formulated. The expectations will be tested later on in the analysis section of the thesis. This will be done after I have explained the research design and research methods in the methodology section.

In the case description, I will explain how self-managing work processes are implemented in the municipality of Krimpen aan den IJssel. The civil service model will be outlined and a description of the teams and their work processes will be given.

After that, I will present the empirical findings and explain the patterns that have been found in the data. In the analysis section, the theoretical expectations will be reviewed based on the data from the interviews and the document analysis. The goal is to formulate an answer to the main question of the thesis, which will be done in the conclusion and discussion chapter. I will end with a discussion in which I highlight the shortcomings of this inquiry as well as some (practical) recommendations for further research on the topics of self-management and shared leadership.

(8)

Literature Review and Theoretical Framework

In this chapter, I will review the academic literature on the topics of self-management and shared leadership in order to develop a theoretical framework that will guide the empirical research. The framework consists of several parts. First, it is important to define and position the concept of self-management in the context of (public sector) teamwork. The next part is about leadership behaviour, more specifically about shared leadership. I will elaborate on the topic of shared leadership and try to link it to the concept of self-management. After this, it is relevant to dig deeper into the (inter)dependencies that could affect self-management and shared leadership. Three interdependencies and two external dependencies have been identified. In the last section of this chapter, theoretical expectations will be formulated to test the relationships between self-management, the (inter)dependencies and shared leadership. The result is a conceptual framework for empirical research that will be reviewed later on in the analysis section of the thesis.

Teamwork

Various scholars have attempted to define the concept of teamwork. This conceptualisation is important as teamworking is the core for both SMWT’s and shared leadership. Therefore, it is necessary to give a clear definition of teamwork.

The definition I use is the one of Cohen & Bailey (1997). According to Cohen & Bailey;

A team is a collection of individuals who are interdependent in their tasks, who share responsibility for outcomes, who see themselves and who are seen by others as an intact social entity embedded in one or more larger social systems . . . and who manage their relationships across organizational boundaries. (1997, p.241)

Important elements of the definition are that team members share responsibilities and are interdependent in their tasks. Later in this chapter, these elements will be connected to the concepts of self-management and task interdependency.

Cohen & Bailey (1997) dig deeper into the subject of group-dynamics by introducing the concept of “groupness”. It is the degree to which groups are integrated and interdependent. Often, groups that develop a high level of groupness, that being groups which are highly integrated and interdependent, are labelled as teams instead of groups in the academic literature

(9)

(Cohen & Bailey, 1997, p.241). Contrary to Cohen & Bailey (1997) I will not use the terms ‘team’ and ‘group’ interchangeably. The focus of this thesis will be explicitly on teams and the concept of groupness will be used as an indicator for both integration and the interdependency of teams (see also Groeneveld & Kuipers, 2014, p.6).

However, as Kuipers & Stoker (2009, pp.400-401) point out, in HR-literature this is described in group development theory. The phase-model of Tuckman & Jensen (1977), originally developed for group-dynamics, describes this development process which consists of five stages through which a group passes:

1. Forming. The initial group phase of orientation among group members in which interpersonal and task behaviour is tested.

2. Storming. The second stage of the group process, where interpersonal conflicts and positioning are the bases of group influence and task requirements.

3. Norming. Overcoming the resistances of the second phase to achieve group cohesiveness and developing norms and roles occur in this third phase.

4. Performing. This is the fourth stage of group development and focuses on task performance. The roles and group structure developed in the norming phase form the basis for accomplishing the task.

5. Adjourning. In this final phase the group separates and, in a new form, starts again with the forming phase.

Types of teams

Four types of teams are identified in organizations: Work teams (1), parallel teams (2), project teams (3) and management teams (4) (Cohen & Bailey, 1997, p.241).

According to Cohen & Bailey (1997) a work team is the most common type of team composition. They define work teams as “continuing work units that are responsible for producing goods or providing services” (p.242). The key characteristic of work teams is a stable, usually full-time and well-defined membership. Another important aspect is that work teams are traditionally led by a manager who is in control of the working processes. However, as Cohen & Bailey (1997, p.242) already were aware of more than 20 years ago, recently there are alternative forms of work teams in organizations that are centred around the principles of

(10)

self-management and autonomous work processes (see for example Groenenveld & Kuipers, 2014, p.7; Kuipers & Stoker, 2009, p.399; Stewart & Barrick, 2000, p.135).

Parallel teams consist of people from different work units that are brought together to perform functions that the formal organizational structure is not equipped to perform well (Cohen & Bailey, 1997, p.242). This teams exist parallel to the formal organizational structure. The main characteristic is that parallel teams have limited authority and can only make recommendations to individuals that are higher up in the organizational hierarchy. Often, these teams are being used to solve problems and for improvement-oriented activities (Cohen & Bailey, 1997, p.242). There are a lot of similarities between project teams and parallel teams. Project teams also draw their people from different disciplines/teams to solve a certain problem. However, project teams generally are time-limited and produce one-time outputs whereas parallel teams are more or less vast. The work that a project team does is either an incremental improvement over an existing concept or the implementation of a new idea. The major difference with parallel teams is that the job or task of the project team is non-repetitive in nature (Cohen & Bailey, 1997, p.242-243).

Management teams are responsible for the overall performance of the organization. The task of a management team is to coordinate and provide direction to the teams and people under their jurisdiction. The authority of the management team is obtained by the (formal) hierarchical position of the team in the organization, which in rank is positioned above the other members and teams (Cohen & Bailey, 1997, p.243).

Kuipers & Groeneveld (2016) add an extra dimension to the different types of teams. They make a distinction based on homo- and heterogeneity. It is about differences in the functional background (i.e. expertise) of the team members. A monodisciplinary team consists of people with similar functions and functional backgrounds. For example, teachers in a primary school or a team of nurses on a certain department of a hospital. A multidisciplinary team consists of people with different functions who also have different expertise’s. Examples of multidisciplinary teams are ‘traditional’ management teams (where the managers from the different departments come together), most policy teams or crisis teams (Kuipers & Groeneveld, 2016, p.18).

(11)

Self-managing work teams

A self-managing work team (SMWT) is a specific form of work team. There are a few characteristics in which SMWT’s differ from traditional work teams. The definition of SMWT’s used by Kuipers & Stoker (2009) identifies a few of these characteristics. They define SMWT’s as “groups of interdependent individuals that are able to self-regulate their behaviour concerning relatively complete tasks” (p.399). Adding up to this, they mention that self-management refers to a reduced need for hierarchical command and control leadership in organizations. It is the team itself that is in the lead (Kuipers & Stoker, 2009, p.399).

The absence of strong hierarchical control is an important characteristic in an organizational design that is based on the principles of self-management. In the literature, a reduced need for hierarchical control is often complementary to increased autonomy of individual employees and/or teams. Stewart & Barrick (2000, p.139) refer to this individual autonomy with the concept of self-leadership. Self-leadership is the ability to decide how tasks should be carried out as well as what should be done and why. In a self-managing design, teams have been given more freedom, responsibility and authority. Team members rather than supervisors make decisions and are responsible for the composition of (their own/the teams) work processes (Stewart & Barrick, 2000, p.139). A big misconception on the topic of self-management is that a reduced need for hierarchical command implies that there is no supervisor or manager. In fact, most organizations that have implemented a self-management approach still have some sort of management- or direction team in their organizational design. The difference is that in self-managing teams, employees take over some of the ‘traditional’ management tasks and are therefore involved in making decisions that were formerly delegated to supervisors and managers (Cohen & Bailey, 1997, p.242). So, the autonomy and the authority of employees and teams is higher in a self-managing design when compared to traditional bureaucratic approaches.

In general, the implementation of self-managing work processes has two main purposes. Firstly, to increase the wellbeing of employees. This will result in a better quality of the work life. The argument is that employees will experience higher feelings of freedom and (personal) ownership in their daily work processes (Stewart & Barrick, 2000).

The second goal is complementary to the wellbeing of employees and is about increasing the organizational performance. If employees experience an increasing quality of their work life because of more self-management, this can result in better performance. Here, it is argued that

(12)

effect on the relationship between self-management and performance. The wellbeing of employees in this case is not a goal in itself, but a tool to increase organizational performance (Pearce & Ravlin, 1987).

Various scholars have studied the advantages and disadvantages of self-managing work processes. One of the major beliefs is that more self-managing work processes result in a decrease of shirking (Stewart & Barrick, 2000, p.139). Employees try not -or less- to shift responsibilities and tasks onto others because of high feelings of (shared) ownership and commitment towards their individual work tasks, other team members and the organization (Pearce & Ravlin, 1987). Pearce & Ravelin (1987, p.752) see self-management as a work-related form to increase self-determination. This self-determination is a specific form of intrinsic motivation. In line with self-determination theory (SDT) (see Deci & Ryan, 2004), Vandenabeele (2007) argues that in a public sector context, an increase of self-determination on the microlevel (individual employees) could have a positive effect on the organizational performance (pp.549-553). Therefore, the implementation of self-managing work processes could be advantageous for employee’s motivation.

While self-management has some clear advantages, it is relevant to also touch upon the possible disadvantages of the model. It is Barker (1993) who in his article “Tightening the iron cage” – a direct reference to Weber’s famous article about rationalization, control and bureaucracy - points at the downsides of a self-management model. His argument is that it is not for granted that a self-management design results in a better work environment. In his case study of SMWT’s, bureaucratic control performed by managers was replaced by concertive control by the team members. The team members created a system of “value-based normative rules” that controlled each other’s’ actions more than the former bureaucratic (hierarchical) system. The normative rules of the team became rationalized over time, causing the organizational change (implementation of self-management and mutual control) to not free these workers from Weber's iron cage of rational control (by the managers), but instead tighten them to the point where organizational performance went down because of concertive control; control based on the normative values of the team (Barker, 1993). Moreover, Barker (1993) teaches us that a model of self-management in itself is not better than the traditional bureaucratic organizational models. What is most important, is how the model is being carried out by employees.

(13)

Shared leadership

Shared leadership is about the distribution of leadership tasks across multiple team members. Carson, Tesluk & Marrone (2007) have outlined the definitions that are being used by other scholars to define the concept of shared leadership. A common element in the definitions is that shared leadership is about perception. It is a collective feeling of influence on each other by the team members. In addition to this, shared leadership implies that leadership is carried out by the team as a whole (or by multiple team members) rather than by a designated individual (Carson et al., 2007, p.1219). In their explanation of the concept, Carson et al. (2007) point at the fact that shared leadership contrasts the conventional paradigm of ‘vertical leadership’, “the traditional hierarchical form of management in which a manager is positioned hierarchically above and external to a team, has formal authority over the team, and is responsible for the team’s processes and outcomes” (Carson et al., 2007, p.1218). The autonomy of teams in combination with the absence of hierarchical top-down control, is in line with the philosophy behind SMWT’s. Carson et al. (2007) also touch upon the topic of self-management in their article. They emphasize that flat organizational structures (e.g. self-managing designs) have a need for leadership that originates from within the team as opposed to traditional (bureaucratic) structures where leadership is centred in a single individual (manager or supervisor) who is elevated by hierarchy (p.1217). According to this statement, self-management and shared leadership would go hand in hand for implementation into organizational design philosophies. However, Carson et al. (2007) also point at the fact that a self-management design in itself does not result in leadership behaviour that is distributed among team members, as other factors such as the internal team environment and external coaching could also influence shared leadership (p.1221). Chreim & MacNaughton (2016) found evidence that shared leadership is associated with reduced conflict, increased trust, consensus and cohesion between team members. But they also found that a lack of clarity about leadership roles, called role ambiguity, could result in team members that feel uncertain over their performance and that are less engaged in teamwork and decision making (p.202).

The concept of shared leadership is related to distributed leadership. However, there are some differences. The difference between distributed leadership and shared leadership is that distributed leadership focusses on the distribution of formal authority (for example traditional management tasks) in the organization or between team members, whereas shared leadership

(14)

could be defined along a continuum based on the number of leadership sources (i.e. team members) that have a high degree of influence on the team. (Carson et al., 2007, p.1220). In other words, shared leadership does not imply that every individual team member carries out leadership tasks. It is merely about the distribution and influence that multiple team members

can have on each other and the team tasks. Carson et al. (2007) argue that shared leadership

represents a condition of mutual influence, embedded in the interactions among team members “that can significantly improve team and organizational performance” (p.1218). However, as I have outlined in de paragraph about disadvantages of a self-management design, mutual influence could also negatively affect the organizational performance (Barker, 1993). The question that rises is if mutual control increases (or decreases) organizational performance. As for shared leadership, what we do know is that there is a positive effect of shared leadership on the organizational performance. Shared leadership may result in greater effectiveness than the leadership of a single internal team leader due to feelings of shared ownership (Carson et al., 2007, pp.1224-1228).

Interdependencies

The literature mentions various interdependencies in relation to self-management and leadership behaviour. These interdependencies are interesting, as it is expected in this thesis that interdependencies have an impact on the association between self-management and shared leadership. Stewart & Barrick (2000) provide a clear definition of interdependence. Interdependence is defined as “the extent to which team members cooperate and work interactively to complete tasks. High interdependence occurs when team members interact cooperatively and depend on each other for information, materials, and reciprocal inputs” (p.137). Unfortunately, the definition above is too limited for interdependence as referred to in the context of this study. The definition of Steven & Barrick (2000) mostly covers task interdependence. However, in this thesis I also study goal- and stakeholder interdependency as important moderators for explaining leadership behaviour in SMWT’s. Following the example of Groeneveld & Kuipers (2014), I will elaborate on the interdependencies that are validated in a public sector context: task interdependency, goal interdependency and stakeholder interdependency.

Task interdependency means that team members depend on one another to accomplish work

tasks. Team members need to interact in order to accomplish work tasks. In other words, when task interdependence is high, it is not possible for team members to function properly without

(15)

reciprocal interactions. Research has shown that task interdependence is related to increased team motivation, responsibility and commitment towards the job (Campion et al., 1993, p.827).

Goal interdependency is about the linkage between individual goals and the group goal(s).

Team members depend on one another for accomplishing the team goal. If goal interdependency is high, individual team members fulfil a piece of the puzzle, meaning that team members rely on each other for accomplishing the team goal(s). For an individual, it is not possible to reach the (team) goal without help of others. A similar, clear and defined goal for both the individual and the team is believed to be crucial for team effectiveness (Campion, Medsker & Higgs, 1993, p.827).

Stakeholder interdependency means that team members are dependent on one another because

of their contact(s) with external actors that are necessary for individual performance. In other words, if stakeholder interdependency is high, interactions between a team member and an external actor could have consequences for the work processes of another team member. It refers indirectly also to the external cooperation that team members need from other persons that are not part of the team (or the same organization). It is about reciprocal management of boundary-spanning activities that can increase team effectiveness and cohesion (Groeneveld & Kuipers, 2014, p.8; Campion et al., 1993).

External dependencies

In the literature, two more concepts have been identified that are relevant to study in a public sector context. These are political control and the accountability of the team. Political control and accountability cannot be defined as interdependencies. However, both concepts are relevant for this study because of the effect they could have on design and (leadership) behaviour. Political control and accountability influence work processes in public sector teams and could therefore affect team behaviour. Both can be labelled as external dependencies in the relation between SMWT’s and leadership behaviour.

Political control focusses on the dichotomy between politics and administration. It is about the

tools that politicians have to influence the work of civil servants. This can vary between the teams (e.g. policy teams have more direct control of politicians then executive teams) and can influence work processes. For political control, it is about top-down steering of political actors on the SMWT’s.

(16)

Accountability can also influence self-management and leadership behaviour in teams. Its

degree varies across organizations and teams. In a public sector context, accountability is about responsiveness and the justification of actions towards citizens. The question rises to what extent teams are flexible to develop their own work processes (according to the principles of self-management) when the production or output must be guaranteed. Hood (1991, p.11) refers to this with his lambda value that is about robustness and resilience of work processes. According to Hood (1991, p.4), accountability requires a clear assignment of responsibility for action and a clear statement of goals.

Theoretical expectations

Based on the literature I expect that there is a relationship between self-management as an organizational design choice and shared leadership (behaviour). Within shared leadership, leadership behaviour varies across team members rather than that it is centred in a designated individual. In other words, leadership transforms into a team process. This is in line with a self-management design approach where the team itself is in the lead when it comes to decision making processes. I expect that an organizational design that is centred around the principles of self-management complements the emergence of shared leadership in public sector teams. It results in leadership that is more distributed among team members. This results in the following expectation:

There is a relationship between the implementation of self-management as an organizational design and the emergence of shared leadership in public sector teams. When self-management is implemented, it is expected that the degree to which shared leadership behaviour could be observed, is higher.

The focus of the thesis is on SMWT’s. However, as explained in the section about types of teams, it is possible to differentiate on the homo- and heterogeneity of teams. Because of the differences between teams in for example tasks, composition and responsibilities, it is expected that the type of team has an influence on shared leadership processes. In other words, even when the organizational design is the same, it is expected that monodisciplinary teams differ in their fulfilment of self-managing work processes and leadership behaviours in comparison to multidisciplinary teams. This results in the following expectation:

(17)

There is a relationship between the type of SMWT and the degree to which shared leadership is present in public sector teams.

The next expectation is about the interdependencies in relation to self-management and shared leadership. I expect that goal interdependency, task interdependency and stakeholder interdependency could have an effect on shared leadership behaviour. When team members are (inter)dependent on each other to perform successfully, it is likely that this affects team behaviour. Team members need one another. In other words, leadership becomes a team process (shared leadership). Therefore, I expect that there is an effect of the interdependencies on the relationship between the type of SMWT and shared leadership. The higher the degree of interdependence, the more chance that shared leadership emerges in public sector teams. This results in the following expectation:

The degree to which shared leadership is present in the different types of SMWT’s is dependent on the degree of goal-, task- and stakeholder interdependence. The degree of shared leadership is expected to be higher when the degree of interdependence is also found to be high.

As for the external dependencies, I expect that there is a negative effect of political control and accountability on the relation between self-management and shared leadership. This means that when the degree of political control and/or accountability of the team is higher, the emergence of shared leadership as a result of the implementation of self-managing work processes is expected to go down. The reasoning is that political control and (high) accountability could be categorized as traditional bureaucratic characteristics where leadership processes by nature are shaped through hierarchy. Political control in itself is a control mechanism that contradicts the principles of shared leadership, it facilitates a principal-agent relationship. As for accountability, traditionally, work processes that require a high degree of responsibility are strictly managed. Therefore, it is likely that this affects the emergence of shared leadership in a negative way. This results in the following expectation:

The degree to which shared leadership is present in the different types of SMWT’s is dependent on the degree of political control and accountability. The degree of shared leadership is

(18)

expected to be higher when the degree of political control and accountability is lower in the teams.

Conceptual framework

The following conceptual model is a visual representation of the expected relationships between self-management and shared leadership. In the framework, the focus is on SMWT’s. The model starts with the type of team. Two types of SMWT’s have been identified: monodisciplinary teams and multidisciplinary teams. The variance is based on the homo- and heterogeneity of the team. The effect of the type of SMWT on shared leadership is the next step that is modelled. From the literature review, it is expected that this relationship is positively influenced by the

interdependencies (task interdependency, goal interdependency, stakeholder interdependency)

and negatively affected by the external dependencies of the team (political control and accountability). It is expected that a self-managing organizational design has a positive effect on the emergence of shared leadership in a (public sector) team.

Figure 1 gives an overview of the conceptual framework with all variables and hypothesized

relationships between them. The model will be tested later on in the analysis section.

(19)

Methodology

In this chapter, I will discuss the research design and research methods. First, the design of the research will be explained. Thereafter, the concepts of the conceptual framework will be operationalized. Next, I will touch upon the method of data collection and data analysis. An assessment of the reliability and validity of the study will be provided afterwards. Lastly, a summary of the case of Krimpen aan den IJssel will be given in the case selection.

Research design

The research was conducted among 25 self-managing work teams in the municipality of Krimpen aan den IJssel, the Netherlands. The teams are part of the administration of the municipal organization. The goal of the study is to unravel how and under what conditions the types of SMWT’s are affecting the emergence of shared leadership in public sector teams. Hence, a qualitative approach suits best if we want to understand the mechanisms of leadership behaviour in SMWT’s. A qualitative method is particularly useful for answering the how and

why question(s). It enables the researcher to study different aspects and examine them in

relation to each other while at the same time having a holistic view on the process(es) and its total environment (Meyer, 2001). The design of the study is nested. I have designed a qualitative single case study in which I compare different types of SMWT’s within the organization. The focus of the study is on the team level. I am interested in the (shared) leadership behaviour of teams and not in individual behaviour as a result of the implementation of self-management. Therefore, a single case study design in combination with a comparative approach is suitable as variation in leadership behaviour could be ascribed to the teams, ceteris paribus the organizational design. The division of a single case into a subset of smaller cases “provides the opportunity to identify both similarities and differences across the subcases. What is gained from this form of analysis often serves as the foundation for theoretical generalizations” (Paterson, 2010). The differences between the cases, in this study the observed leadership behaviours, can provide new ways of understanding how shared leadership behaviour (as a result of self-management) may unfold in each of the different types of SMWT’s.

The variation of case selection is on the independent variable, the type of SMWT. Interviews were conducted with employees of the civil service of Krimpen aan den IJssel. First, in cooperation with the HR-team of the organization, a distribution has been made based on the

type of team. All 25 teams were divided in categories based on the homo- or heterogeneity of

(20)

monodisciplinary teams and 3 multidisciplinary teams. In total, 12 interviews were conducted with employees of 6 different teams.

Type of team Team Function of respondent

Interview 1 Monodisciplinary KCC Employee of customer contact centre

Interview 2 Monodisciplinary KCC Employee of customer contact centre

Interview 3 Monodisciplinary Consulenten Sociaal Domein Social domain consultant

Interview 4 Monodisciplinary Consulenten Sociaal Domein Social domain consultant

Interview 5 Monodisciplinary Communicatie Junior communication advisor

Interview 6 Monodisciplinary Communicatie Junior communication advisor

Interview 7 Multidisciplinary Beleid sociaal domein Senior policy advisor

Interview 8 Multidisciplinary Beleid sociaal domein Senior policy advisor

Interview 9 Multidisciplinary Beleid samenleven Senior policy advisor

Interview 10 Multidisciplinary Beleid samenleven Senior policy advisor

Interview 11 Multidisciplinary Procesregie sociaal domein Administrative assistant

Interview 12 Multidisciplinary Procesregie sociaal domein Process director KrimpenWijzer

Table 1: Interview overview.

The municipal organization of Krimpen aan den IJssel has been chosen for three reasons. Firstly, due to accessibility. The municipality gave permission to access relevant documents and for the conduction of interviews with her employees. It was very helpful and a big contribution to the inquiry.

Secondly, due to knowledge of the organization. At the time of my internship, the organization initiated the changing process to develop a work structure that was based upon the principles of self-management. This means that it is easier to observe differences and to put things into perspective; it leaves room for ‘verstehen’ (Meyer, 2001, p.330). Another advantage is that knowledge of the organization makes it easier to figure out which people are relevant to conduct interviews with.

Thirdly -and most importantly-, due to the possibility to compare leadership behaviour among the different types of SMWT’s, ceteris paribus the organizational design. All teams have the same conditions of self-management as a result of organizational design. Therefore, variation in leadership behaviour can be contributed to the type of SMWT and other variables than the (self-management) design. A (single) case study design enables us to unravel these mechanisms.

(21)

Operationalization of concepts

The data is collected both on the team level (team arrangements, other documents) and on the level of individual employees (interview data). However, the concepts have been operationalized on the team level. An important side note is that most of the indicators are derived from quantitative research and are originally measured on a five-point Likert-scale (survey data). I have transformed many of the indicators into qualitative indicators that could be applied and measured in the form of interview questions.

Self-management. The first concept that needs to be operationalized is self-management. In

this thesis, self-management is examined as an organizational design choice. As elaborated upon in the theoretical section, there are a few characteristics that help to identify self-management. For the operationalization of the concept, Groeneveld & Kuipers (2014, p.16) have measured the degree to which teams are actively involved in managing their own work activities. Despite the fact that their focus is on (active) behaviour and not so much on the design, in an interview context, the indicators of Groeneveld & Kuipers (2014) could teach us something about the organizational design choices and whether they represent self-management processes in the teams. I use the following indicators to measure self-management: Teams are

involved in making decisions in the organization; The team reaches decisions together; Teams decide for themselves when external actors are needed to solve a problem; Teams make their own decisions about the division of tasks; Teams make decisions about the planning of tasks autonomously (Groeneveld & Kuipers, 2014, p.16).

Shared leadership. Shared leadership is the next concept that needs to be operationalized. In

general, “leadership can be conceptualized in relation to either the strength of influence (i.e., its quality or effectiveness) or the source of influence (i.e., a single versus multiple team members)” (Carson et al., 2007, p.1219). To measure shared leadership, I will focus on multiple sources of influence. According to Carson et al. (2007, p.1220) shared leadership can be measured along a continuum based on the number of leadership sources that have a high degree of influence on the team. Carson et al. (2007, p.1225) measured shared leadership based on

density: A measure of the total amount of leadership displayed by team members as perceived

by others on the team. Teams in which many team members are rating other team members to show leadership behaviour will score higher than teams in which only one member or a few

(22)

qualitative research context it is not possible to perform a network analysis based on density. Therefore, I have formulated the following indicators that have been used to measure (shared) leadership behaviour: initiative (Are there team members who often take the initiative?),

reliability (Does the team rely on a person in the team for leadership?), representativeness (Which person(s) represent the team towards other teams/stakeholders/the direction?) and shared ownership (Do team members feel that leadership in the team is distributed among multiple persons).

Type of team. For this study, two types of self-managing work teams have been distinguished:

monodisciplinary teams and multidisciplinary teams (Kuipers & Groeneveld, 2016). I am interested in the variation between team compositions in terms of functional backgrounds, expertise and work processes. In line with Campion et al. (1993) and Groeneveld & Kuipers (2014), I have adopted a measure for the heterogeneity of teams. The following indicators have been used to measure the type of team: Team members vary in their area of expertise; Team

members have a variety of different backgrounds and experiences; Team members have skills and abilities that complement each other (Groeneveld & Kuipers, 2014, p.14). I have added

one more indicator to this list: Team members perform similar (or different) work tasks and

activities. Also, I want to control for team size as a variable. The size of the team could possibly

affect work- and leadership processes in a team. Therefore, the number of team members is an indicator for measuring team size (Cohen & Bailey, 1997).

Task interdependency. Task interdependence is about group members that depend on one

another to accomplish the work (Campion et al., 1993, p.827). I used the measurement of task dependence from Campion et al. (1993) and Groeneveld & Kuipers (2014) and transformed them into interdependence measurements. This resulted in the following indicators: Team

members rely on information and/or materials from other members to accomplish work tasks; Jobs are related to one another; Team members cannot perform well without other team members.

Goal interdependency. Goal interdependence is about the need to common goal setting by

team members (Campion et al., 1993, p.827). For the operationalization of the concept, I choose indicators that are in line with the indicators that Groeneveld & Kuipers (2014) have used to measure goal dependency: Team members have a clear view about the objectives of the team;

(23)

mind when they work together (Groeneveld & Kuipers, 2014, p.15). However, for

interdependence I have added the following indicator: Team members rely on each other to

reach team goals;

Stakeholder interdependency. Stakeholder interdependency is about the need for external

cooperation. To operationalize stakeholder interdependence, I used existing measurements of boundary-spanning behaviour. Groeneveld & Kuipers (2014, p.15) identified indicators for boundary-spanning behaviour that can be used to measure stakeholder dependency. However, as I am interested in interdependence instead of dependence, I will only use two indicators:

Team members reach out to individuals outside the team that can provide project-related expertise or ideas; Team members proactively seek the advice and support of individuals outside the team. For interdependence, I have added the following indicator: Team members rely on the contacts of other team members with external actors for performance.

Political control. Political control in this case is about the influence of political actors on (daily)

work processes (Calvert et al., 1989). For each of the teams, I am interested in the connection between team members and the political actors in the municipality. The interaction(s) between politicians and municipal officials varies a lot among teams. Some teams are connected in such a way that their work tasks are primarily aimed at serving the politicians and the councilors of the municipality, whereas other teams barely have to deal with political influences due to fixed work processes. Therefore, I will use the degree to which political actors influence work

processes in teams as an indicator for political control. In other words, the amount of top-down

control from political actors on the teams will be measured. Teams in which many team members interact with political actors on a regular bases are expected to score higher than teams in which interactions with political actors are scarce.

Accountability. Accountability in a bureaucratic context is about the reliability of output(s).

For this study, I am interested in what Hood (1991) refers to with his lambda value. It is about robustness and resilience (Hood, 1991, p.11). For the operationalization of the concept, I have used the following indicators that are derived from Hood’s theory: There is a clear assignment

of responsibility for action among team members; There is a clear statement of goals among team members (Hood, 1991, p.4). I have added an extra dimension to measure accountability

which focusses more on reliability: Team members participate in work processes that are

(24)

Because the research is conducted during the time of the Coronavirus, it is relevant to control for accountability of teams during the crisis. This resulted in the following indicator: Work

processes are continuing in times of a crisis.

Data collection and analysis

Data collection

Data is mainly collected through interviews. The interviews were semi-structured. The questions focused on the relationship(s) between the types of SMWT’s, team processes and leadership behaviour in the teams. The questions were related to team composition, self-management, team processes, (inter)dependence of team members and leadership behaviour. The standardized format of the interviews can be found in appendix 3. Besides the standardized questions, a semi-structured method of interviewing allowed for anticipation on interesting answers to dig deeper in why something was the case. This facilitated process-tracing.

I also used data from a podcast that was recorded between A&O Fonds Gemeenten, the town clerk of Krimpen aan den IJssel, an employee of the HR-team and a member of the Employee Council. In the podcast, the town clerk explained the philosophy behind the change process of the municipal organization. He explained why principles of self-management have been implemented, what the current state of the changing process is and, in the end, what he hopes to achieve with the new work structure. The data was very helpful for understanding the change process and facilitated process-tracing.

Finally, I have studied documents of the organization regarding teamwork, self-management and the organizational change. All documents could be found at the municipality’s intranet called ‘Konnect’. On Konnect, there is an online platform which is called ‘werken in teams’. The relevant documents have been posted in this section. With the permission of the municipality, I was able to study the documents.

As for informed consent, all respondents have been asked for permission to record the interviews and for the usage of their data for research purposes. The data is anonymised in such a way that only the name of the respondents’ team and the function title is listed. For the collection of all other data sources, the same method of informed consent has been applied.

(25)

Data analysis

The research design is nested. The method of analysis that has been applied is a within-case analysis with a comparative nature. Within the case, leadership behaviour of 6 self-managing work teams has been compared. This makes a comparative within-case analysis the right tool for analysing in this thesis.

I have performed a document analysis and an analysis of the gained interview data (qualitative content analysis; Bryman, 2012). The interview data has been coded with a coding scheme and is analysed afterwards. The coding scheme is very basic. I have related all operationalized variables to specific colours. Thereafter, I have analysed both the interview data and the documents, and marked all sentences that were related to a certain variable and/or indicator in the same colour. In the end, I have bundled all marked sentences of a specific colour to see if I could identify patterns in the data. Because of the small number of respondents, this method of coding was suitable and most efficient for this study.

Furthermore, I have executed a form of process-tracing as well. It has helped by understanding the organizational changing process and its effect(s) on the team processes. I have used the A&O Gemeenten podcast and the documents on Konnect to learn more about the teams, the municipal organization, and especially about the team development phases after self-managing work processes were implemented. Process-tracing helped to put these observations into perspective.

Another data source that has been consulted for analysis is (old) survey data from 2019 on teamworking called the ‘samenwerkingsmeter’. When the organizational changing process started, the organization has measured various aspects about cooperation among team members and leadership behaviour in the teams. Every employee had to fill in the survey individually. Therefore, data is collected at the microlevel (individuals). The survey used a Likert scale in which the respondents had to score the questions between 0 (totally wrong) and 5 (totally agree). A summary with average scores per team has been made on the team level. In the survey data, for each team there are aggregated scores in the area of leadership. The aggregated scores have been used to control for the perceived leadership behaviour in the selected teams at that point in time. I am using these aggregated scores to control for selection bias. If the aggregated scores on the team level from the 2019 survey data vary a lot across the selected teams, this means that the teams’ leadership behaviour was already different before the implementation of

(26)

self-managing work processes. If not, changes in leadership behaviour can be ascribed to the organizational change.

Reliability and Validity

Reliability

Measurement reliability is about “the dependability or consistency of the measure of a variable” (Neuman, 2014, p.212). In the words of Bryman (2012), “reliability is concerned with the question of whether the results of a study are repeatable” (Bryman,2012, p.47). There are three types of reliability. Stability reliability is about reliability across time. It means that If an indicator is measured at different points in time, the result(s) should be consistent (Neuman, 2014, p.212). For this research, stability reliability is considered to be a bit tricky due to the coronavirus and its influence on daily routines and team processes. This might have influenced some of the test results. However, I tried to tackle this ‘potential problem’ by asking the respondents to compare the current team processes to the ‘normal’ situation without the coronavirus. Besides, variation in behaviour due to uncontrolled circumstances, could teach us something about the mechanisms of both self-management and shared leadership. To ensure stability reliability, there has been controlled for the aggregated results on leadership behaviour of the investigated teams, as presented in the consulted survey data from the ‘samenwerkingsmeter 2019’. No outliers have been found in the data. The semi-structured interview method also helped to ensure stability reliability. When the same questions are asked at a different point in time ceteris paribus all other circumstances, test results are expected to be alike.

Representative reliability is about reliability across groups. The results should be consistent for

various subpopulations or cases. The indicator should deliver the same results when applied to different groups (Neuman, 2014, p.212). For this research, representative reliability is considered to be high. The semi-structured interview method contributes to representative reliability as it provides a framework for observations to be consistent. I measured the indicators based on preconceived questions in teams with similar designs and found similar results. Therefore, variation in behaviour could be contributed to the research. To ensure complete representative reliability, all teams and/or employees need to be interviewed. However, this is not realistic within the available time.

(27)

Equivalence reliability is about reliability across multiple indicators. The results of the measure

should be consistent across different specific indicators that all measure the same construct. This form of reliability is more important in statistical (quantitative) research (Neuman, 2014, p.213). For each of the concepts, multiple indicators have been included. By doing so, I tried to measure all dimensions of the concepts and enlarge the equivalence reliability. the equivalence reliability is considered to be high. Indicators provided similar results for questions in the same category.

By transcribing all of the interview data, it is possible for other researchers to replicate the study. This enlarges the overall reliability of the study.

Validity

Validity is about truthfulness. It is about how well we measure social reality using our constructs about it (Neuman, 2014, p.212). It refers to how well an empirical indicator fits with actual reality. In detail, it refers to how well an empirical indicator and the conceptual definition of a construct ‘fit’ together (Neuman, 2014, p.215). I will touch upon three forms of validity that are relevant for qualitative research.

The first is the internal validity. It is about causal interference between variables. A high internal validity means that it is likely that the independent variable or intervention causes the dependent variable to change (Morgan, Gliner & Harmon, 2000, p.529). For the research, 6 self-managing teams of the same organization were selected and clustered based on the type of team. The implementation of a self-managing model caused the leadership dynamics of the teams to change heavily. Both the interview data and the document analysis confirmed this claim. An important interview question to test the internal validity was how (shared) leadership dynamics among team members were shaped in the ‘old’ model. If differences in (shared) leadership behaviour after the organizational changing process are being observed -which was indeed the case- it is likely that the implementation of self-managing work processes could have an effect on the (shared) leadership behaviour of the teams. So, internal validity is considered to be high.

Secondly, External validity is about the question of whether the results of a study can be generalized beyond the specific research context (Bryman, 2012, p.47). In general, external validity is easier to obtain in quantitative research due to the number of respondents (large-N studies) (Bryman, 2012, pp.47-48). For this research, the N is considered low (N=12). The

(28)

results can be generalized for the investigated teams due to data saturation of interview data for both team types (monodisciplinary and multidisciplinary), the survey data, and the document data (e.g. the team arrangements). Despite the fact that the study is generalizable for the investigated teams and that it is likely that similar patterns can be observed in the other teams of the organization as well, we cannot make this claim. This goes even further if we want to generalize the data for other organizations as well. Therefore, the external validity is considered to be low. A single case study design does not provide for good generalizations across organizational boundaries. To reach a higher external validity, I argue that a quantitative research approach is necessary if we want to make a representative claim about the (possible) causal effects between self-management designs and the emergence of shared leadership behaviour in public sector organizations/teams. I will explain this argument further on in the discussion chapter.

The third type of validity is called construct validity. In the literature, it is also known as measurement validity and applies primarily to quantitative research and to the search for measures of social scientific concepts. It has to do with the question of “whether a measure that is devised of a concept really does reflect the concept that it is supposed to be denoting” (Bryman, 2012, p.47). Bryman (2012, p.47) gives the example of an IQ-test: Does the IQ-test really measure variations in intelligence? If so, construct validity is considered to be high. For this research, it means that the indicators of the variables that have been used in the conceptual framework of the thesis, have to measure the concepts for the results to be valid. To obtain construct validity, I have operationalized all concepts based on indicators that have been used by other scholars to measure the same concepts as in the framework of this thesis. I studied the indicators that various scholars have used in their articles and copied the most appropriate indicators to measure the concepts in the conceptual framework.

(29)

Case description

The case is centred around the municipality of Krimpen aan den IJssel. It is a small municipality with around 28.000 inhabitants. In 2019, the municipality initiated a new policy for their civil service model that have transformed the working processes from a more hierarchical structure to a flat organizational design with self-organizing as the main characteristic of (team)working. By doing so, the teams became responsible for much of the former management tasks. In the municipal model, self-management behaviour is considered to be high. Here, self-management implies that individual employees as well as the teams have a lot of authority and autonomy to shape working processes and to make formal decisions. The role of the direction in the design is limited.

In 2018 and 2019, a lot has changed to the organizational structure of the municipality. An important change is that the former management team (MT), which existed of the heads of the departments, now is transformed into one integral direction team (DT). The ‘new’ DT has shrunk in terms of formation and responsibilities. The MT existed of all 5 heads of the departments and the town clerk. The new DT consists of only 3 members and the town clerk. This change is in line with the organizational development strategy. The philosophy behind the change was to empower the teams with more authority and autonomy in daily working routines, a so-called bottom-up approach. It implied that the teams got more responsibilities and freedom (read: less restrictions) and by doing so, the aim of the organization was to reach higher standards of performance in terms of efficiency and effectiveness. Thus, by reducing the influence of the DT, a large part of the former management tasks, powers and responsibilities were now delegated towards the teams. Examples of delegated tasks are the authority to sign certain documents, the selection and recruitment of new personnel, (dis)functioning of the team(members) and the distribution of the workload.

The design is not a model of complete self-management. The DT still creates a framework for the teams to work in and has a (political) control function. The norm is that the direction only intervenes in policies and decision-making processes that touch upon sensitive political topics. The DT has formulated in their team arrangement that they ‘form the main link with the Mayor and Executive Board’ (in Dutch: College van B&W) of the municipality of Krimpen aan den IJssel. The DT connects the teams with the Executive Board and coordinates the long-term agenda of the municipality. In their team arrangement they state that in the end, the DT is

(30)

‘integrally responsible, while taking the principals of self-organising of the teams into consideration’.

So, the model can best be categorized as a self-organizing model in which there is a lot of authority and autonomy for the teams as well as for individual team members. The organograms in appendix 1 & 2 give an overview of the old and new organizational structure in which the composition as well as the role of the teams has been changed.

In the new model, there are approximately 175 employees divided over 25 teams. The work teams are composed around the societal tasks of the municipality. Cooperation between- and in the teams is centred around the content of specified topics. Therefore, it often happens that employees participate in specific project- and parallel teams (e.g. there is a project team that is responsible for the development strategy of the organization. The project team consists of employees from 4 different teams) besides being in one of the 25 work teams, also called ‘thuisteams’ of the organization.

(31)

Empirical Findings

In this chapter I will discuss the empirical findings of the research. I will outline the different findings per variable and declare which patterns could be identified based on the interview data and additional sources. After this, a table is presented in which all research findings are summarized. Further on, in the analysis chapter, I will go in-depth about the possible relationships between certain variables and I will compare the findings with the theoretical expectations.

Interdependencies

Task interdependency

Task interdependency varied a lot between the observed teams. In three teams (team 1,2 and

3), task interdependence is considered to be low. In fact, and in line with theoretical

expectations, this was the case for all three of the monodisciplinary teams. Team members emphasized that they did not need other team members to perform well in their daily working tasks. Moreover, their daily working routines are merely fixed and clearly formulated, which causes the team members to work independently (for the most part). Therefore, it is argued that task interdependency is low for all of the monodisciplinary teams.

However, task interdependence is considered to be high for the other three teams (4,5 and 6). Again, there is a relation with the type of team here. Team 4,5 and 6 can all be labeled as multidisciplinary teams. The employees of this teams argued that they need other team members to be able to carry out their individual work tasks in a proper way. Some of the employees even stated that they are not able to function without the help of team members. Therefore, for all of the multidisciplinary teams, task interdependence is considered to be high.

Goal interdependency

Goal interdependency also varied a lot between the observed teams. In two of the teams (team

1 and 2), goal interdependence is considered to be low. Both of the teams are monodisciplinary.

Team members argued that without other team members, they would still be able to reach the team goal. For both teams, the team goal was centered around the provision of good services for the inhabitants of the municipality. An explanation for goal interdependence to be low for these teams, is because -in comparison to the other four teams (3,4,5 and 6)- the work processes

Shared leadership in public sector self-managing work teams: Rhetoric or reality?