Linkages between management control systems at different organizational levels in a Dutch municipality

(1)

MSc Business Administration

Specialization Organizational & Management Control

Linkages between management control

systems at different organizational levels

in a Dutch municipality

By

TOM H. SCHOKKER

(2)

1

Linkages between management control systems at different

organizational levels in a Dutch municipality

Abstract:

The objective of this paper is to analyze how management control systems are related at different hierarchical levels in a Dutch municipality. The case study confirms the management control system as a package theory (Otley, 1980), which is found to be relevant for control systems at different hierarchical levels. The object-of-control framework (Merchant & van der Stede, 2012) is used to map out and improve understanding of the different management control systems in the organization. The exact couplings between those control systems are analyzed and described by using the coupling theory of Orton & Weick (1990). The analysis used data from thirteen interviews, eighteen formal documents and the notes from three attended meetings. The analysis ends with two elements that seem to have influence in the determination of hierarchical coupled management control systems. The first is the extent in which a control system is build up out of different hierarchical control systems or elements. The second is the extent in which managers or subordinates are accountable for their actions. The discussion contains a reflection on the findings and the theories used.

Keywords: Management control systems (package), coupling theory, object-of-control framework, municipal organization, goal congruence, hierarchical relatedness in management control.

Information:

Name: Tom H. Schokker

Email: Tomschokker@gmail.com Student number: 2394243

Country: The Netherlands

University: Rijksuniversiteit Groningen Faculty: Economics and business Master: Business Administration

Specialization: Organizational & Management Control Supervisor: B. van der Kolk

(3)

2

Preface

Dear reader,

In the last six months I conducted a qualitative research for two main reasons. The first reason is to finish my time as a student of Business Administration. After the bachelor Business Economics at Saxion in Enschede (2008-2012), I decided to study Business Administration at the Rijksuniversiteit Groningen (2012-2014). My study period is about to end now. The second reason is to experience what it is like to contribute on academic level. The experience of conducting research was a great pleasure and the result is right in front of you. There are many people who indirectly contributed to this thesis and who made it possible for me to conduct the research. First of all I would like to thank the Municipality of Emmen. The organization and specifically the employees at the department Policy and Direction were willing to support me with the data collection and took time to explain the functioning of and their role in the organization. In the second place I would like to thank my supervisor, Berend van der Kolk. His feedback and support really improved the quality of the thesis. And to conclude, I would like to thank my family and friends who supported me during the years of my study period.

I hope you enjoy reading this thesis.

Your sincerely,

(5)

4

1 Introduction

Management control is a critical function of an organization (Merchant & van der Stede, 2012), still there is no clear explanation of the mutual functioning of and between Management Control Systems (MCS) (Malmi & Brown, 2008). Although previous research has provided great and interesting insights in the functioning of organizations and how to control an organization, there is still a lot of research to do, as stated by Malmi & Brown (2008) for example. The aim of management control is to direct employees’ behavior and decisions in a consistent way with the organizational goals. The systems that ensure this aim are called Management Control Systems (Merchant & van der Stede, 2012) or Management Control Elements (MCE) (Chenhall, 2003). A management control systems package is a collection or set of control elements and control systems (Malmi & Brown, 2008).

In 2008 a special issue of Management Accounting Research was dedicated to the topic of Management Control Systems as a Package. In this special issue, Malmi & Brown (2008) identified research directions regarding this topic. Among other directions, they indicated the research of how the elements and systems of a management control package interrelate to each other. A possible explanation for this gap is that different authors suggest that management control systems or elements do not operate in isolation (Chenhall, 2003; Malmi & Brown, 2008; Otley, 1980). Although a part of the MCS research considers single themes or practices that are seemingly unconnected from each other and the context in which they operate, these single elements are part of a broader control system (Chenhall,2003; Malmi & Brown, 2008).

Malmi & Brown (2008) and Järvenpää & Länsiluoto (2011) both suggest an opportunity to research the linkages that exist between different management control systems within a package. Moreover, Malmi & Brown (2008) specifically indicate that it is a long standing question how the management control systems relate to each other (also addressed in Abernethy & Chua, 1996).

(6)

5 archetypes, Sandelin’s version of the control package (2008) which was also recognized by Merchant & van der Stede (2012), and Malmi & Brown’s version of the control package (2008). The object-of-control framework (Merchant & van der Stede, 2012) provides a stable basis to analyze the case study and categorize the control types used in the organization.

In the eighties, the Dutch government took the initiative to stimulate municipalities to apply business-like management control elements in order to make local authorities familiar with new ways of governance and management. These elements were for example focused on customer orientation, effectiveness and efficiency. This phenomenon is called New Public Management (Hood, 1991; Hood, 1995, van Helden, 1998; ter Bogt & van Helden, 2000; ter Bogt, 2001; ter Bogt, 2003). The design of New Public Management origins from the new institutional economics, which is rooted in the choice theory, transaction costs theory and the agency theory (Hood, 1991).

New Public Management, based on effectiveness and efficiency, was a new style of working and had effect on the internal organization. The municipalities had an inward-looking approach and relied on formal procedures. Customers (i.e. the citizens) had to adjust to these procedures in order to make use of the offered services. Municipalities had a wait-and-see attitude and paid little attention to

performance (ter Bogt & van Helden, 2000). In addition, municipalities focused their budgets mainly on the inputs of the transformation process and paid relatively low attention to the output of the processes (ter Bogt, 2001). By introducing the ideas of New Public Management, municipalities and other

governmental organizations were criticized for their inefficient operations (Hood, 1991). New Public Management forced municipalities to have a more citizen-focused approach (ter Bogt & van Helden, 2000). The impact of shifting from an emphasis on inputs and observing rules towards outputs and quantitative output measures can have consequences for the types of control to be used (ter Bogt, 2001). Although prior research contributed to knowledge about control systems in governmental organizations (e.g. ter Bogt & van Helden, 2000; ter Bogt, 2001; ter Bogt, 2003), there still remain gaps in the literature about how managers and organizations optimize the design of their control systems and what relations exist between the control systems at different levels within the organization (Malmi & Brown, 2008; Järvenpää & Länsiluoto, 2011).

(7)

6 since the introduction of the New Public Management, made the possibility to conduct this research in a governmental, not-for-profit organization interesting. Moreover, it is very interesting to follow the developments in practice about this area. As academia argue, there is no clear explanation for the functioning of the management control system (Malmi & Brown, 2008). Therefore it is interesting to see how for example managers make decisions about the design of the control system and what relations exist between those systems and elements. In what way is the management control package designed and how do managers decide or update their control package? A part of this literature gap (Malmi & Brown, 2008) will be treated in this thesis.

In order to analyze the relations between control systems the coupling theory of Orton & Weick (1990) will be used. This theory can generally describe any relation between organizational phenomena

(Beekun & Glick, 2001). Although it has been used for lots of other subjects in academic literature, it still seems relevant as it is recently suggested by Malmi & Brown (2008). In their suggestion, they indicate the interesting question about the hierarchical relations between management control systems in a package approach. Moreover, the exact theory that is suggested concerning this topic is the coupling theory by Orton & Weick (1990). The theory uses a framework in which a certain element is ranked to the extent in which it is responsive and distinctive upon another element. A detailed description of these coupling will help understand how management control systems are coupled and therefore contribute to the knowledge about the management control theory in a package approach (Otley, 1980; Malmi & Brown, 2008; Grabner & Moers, 2013).

The structure of the thesis will be as following. The next section contains the research questions and its theoretical and practical relevance. The second section is the literature review, which contains the relevant and prior literature in the field. Section three elaborates upon the research methodology. Section four will contain an analysis of the case study concerning the research question. The fifth section contains a discussion about the literature and theories used in this thesis. This thesis will end with a conclusion, the limitations and suggestions for further research.

1.1 Research question

(8)

7 organizational levels are related to each other. This contribution refines the theory of Management Control Systems and Management Control Systems as a Package. The main research question in this thesis will be:

“How do management control systems within the management control package at different organizational levels relate to each other in a Dutch municipality?”

In 2008, Malmi & Brown confirmed that the interrelations between management controls is still a gap in the literature, especially research that investigates how those controls relate to each other on different organizational levels is scarce. Result controls at top organizational level might be tightly coupled with action control at departmental level. Tightly coupled systems have a high responsiveness and a low distinctiveness. This means that the action controls are very sensitive for the changes in the result controls at organizational top level and are almost similar to those result controls. Recognizing patterns in those design choices will help contribute to the literature gap about the linkages between

management control systems at different organizational levels, using a package approach.

For theoretical purposes, the outcome of this thesis will help uncover what is currently a black box; how the different types of controls relate to each other through the different organizational levels. From a more practical perspective, managers are in the need of a good management control system, which can address different types of problems, e.g. thefts, frauds and unintentional errors (Merchant & van der Stede, 2012). However, the design of the system is dynamic and needs to respond to the context, as Chenhall discussed in his paper (2003). It is critical to know as a manager what combination of controls (e.g. results, action, personnel, and cultural controls1) is sufficient in a specific organizational context in order to manage in congruence with the organizational interests (Malmi & Brown, 2008).

The approach that will be taken to address the research question is qualitative (Yin, 2014). The case study will take place at the municipality of Emmen, where several techniques to gather data will be used, namely interviews, analyzing documents and observations. In order to get an overview of what control systems are used in the organization, these data sources provide the basis for the analysis. After that, the coupling theory (Orton & Weick, 1990) will help to analyze the control elements at different organizational levels and determine to what extent those systems are related to each other.

1

(9)

8

2 Literature review

The literature review contains an overview of the relevant literature and theories that are related to the main research question. First, the definition and function of control will be explained. After that, the key literature on management control systems and its typologies will be reviewed. This section will discuss different typologies of management control systems. The fourth subsection of the literature review is about how control systems can relate to each other, where the coupling theory will be introduced. The coupling theory will be used as an instrument to make sense of the relations between various control systems. In the last subsection the context of public organizations will shortly be highlighted in order to be conscious of possible effects on this research.

2.1 Control in organizations

What is control exactly? In many published papers about Management Control, authors start with the notion that the definition of the concept of Management Control Systems is a challenge (Fisher, 1998), because it is widely used (Abernethy & Chua, 1996; Chenhall, 2003; Merchant & van der Stede, 2012; Malmi & Brown, 2008). Some of these definitions have an overlap, while others are quite different from each other (Malmi & Brown, 2008). According to the Oxford dictionary the definition of control is: “the power to influence or direct people’s behavior or the course of events”

Comparing definitions from the literature field of management control systems, it seems that the essence of the definitions is still closely related. For instance, Abernethy & Chua (1996) defined management control systems as : “a system that comprises a combination of control mechanisms designed and implemented by management to increase the probability that organizational actors will behave in ways consistent with the objectives of the dominant organizational coalition” (p. 573).

Compared to the definition of the Oxford dictionary, the power is described by Abernethy & Chua as the management of an organization. In addition, the control mechanisms are designed to increase the probability (Oxford’s translation “influence”) to let actors behave in a consistent way.

Merchant & van der Stede (2012, p.3) stated that “management control is a critical function in

organizations”. In order to explain why management control is that important, it might be useful to have an understanding of the underlying theories of control and organizations. In 1937 Ronald Coase

published his ideas about the theory of the firm. He argued that people begin to organize their

(10)

9 is greater than the transaction cost of coordinating production within a firm. This idea is the basis for the transaction cost theory (Coase, 1937). In small businesses, it is common that one person is the decision-maker. This is one way to overcome the control problem. However, as Merchant & van der Stede (2012) explain, in most organizations – including the small ones – it is not possible to centralize all critical decisions. Therefore, other control solutions are necessary. Decentralization is the process where responsibilities are assigned to subordinates. Agency theory describes the principle-agent-problem. The interests of agents (or subordinates) are not per definition equal to the interest of the organization, i.e. the principles (Saam, 2007), for example due to the assumption that individuals are self-interested (Merchant & van der Stede, 2012). Hence, management controls are the instruments that guard against the possibilities that subordinates will perform a certain activity which is not in desirable to the

organization, or fail to do something they should do (Merchant & van der Stede, 2012). In brief, this explains what Merchant & van der Stede meant by the critical function of management control in organizations.

2.2 Management control systems as a package

In 1980 David Otley published his idea that individual control instruments should be seen as a part of a wider organization control package (Otley, 1980) and that control instruments do not operate in isolation (Chenhall, 2003). From a contingency approach, Otley (1980) recognized that the “mix” of those instruments is not determined, but several different combinations may give equally good results. In the next decades, several authors supported his idea about the control instruments as a package and started exploring the holistic functioning of the management control package (Abernethy & Chua, 1996; Sandelin, 2008; Malmi & Brown, 2008; Sandt, 2009; Ferreira &Otley, 2009; Järvenpää& Länsiluoto, 2011; Grabner & Moers, 2013). Although there is knowledge about various characteristics of a management control package, there are still research opportunities. Studying management control systems in isolation can lead to difficulties (Chenhall, 2003; Malmi & Brown, 2008; Otley, 1980). In addition, Chenhall (2003) suggests that studying a management control system in isolation is a potential threat for model under specification, because it may lead to spurious results which do not reflect reality.

Recently, Grabner & Moers (2013) discussed the conceptual and empirical issues of “management control as a system or as a package”. This paper starts with rephrasing what is actually meant by the system- and package-approach. Management control practices form a system when multiple

(11)

10 into account. A management control package can also be a system, but the difference is that it does not have to contain necessarily interdependent control practices. Management control packages can also contain standalone systems that are implemented, for instance at different points in time or by different departments for different purposes. In order to clarify in more detail what Grabner & Moers (2013) elaborated upon, the next figure visualizes the difference between those approaches in figure 1.

Figure 1. Differences in MC approaches, as mentioned by Grabner & Moers (2013)

As visualized above, the package exists of systems and elements that can be related or unrelated. Systems contain related elements. Elements are management control practices, for example budgets or work procedures (Chenhall, 2003; Malmi & Brown, 2008; Grabner & Moers, 2013). Management control package-theory suggests that the various management control systems within the package are related (e.g. Otley, 1980; Malmi & Brown, 2008). This thesis is responding to a call of Malmi & Brown (2008), which suggest that research about the linkages between and in between control elements and control systems in a package at different organizational levels is a gap in the literature.

2.3 Typologies of management control systems

To analyze the linkages between the management control systems, there are several typologies to help identify the content of management control systems. After discussing various MC frameworks and looking at their characteristics, the choice for the framework to help gather and analyze data will be made.

Simons’ levers of control (1995)

(12)

11 organizations that demand flexibility, creativity and innovation. Simons’ wrote a highly cited paper and created a framework in his paper. This framework reveals that, derived from an organizational business strategy, there are four control levers that can be applied and all four have its own characteristics, which are belief systems, boundary systems, diagnostic control systems and interactive control systems. This framework can be very useful in addressing or analyzing the way in which control systems are used by managers, rather than how an MCS is designed (Chenhall, 2008).

Speklé’s control archetypes (2001)

In his paper “Explaining Management Control Structure Variety: A Transaction Cost Economics Perspective ”Speklé introduces a new framework that discusses the variety in MC structure in

organizations. Speklé argues that control archetypes are different configurations of control devices that can be separated. The idea of those archetypes should be especially useful in addressing the entire control function, which is in line with this research. In the conclusion of his paper, Speklé recognizes that the scales on which the variables score are rough and their boundaries are left implicit. The variables are described at a conceptual level and additionally “their substance is suggested rather than defined” (p. 22).

Malmi & Brown’s management control systems package (2008)

In 2008, Malmi & Brown introduced their perception of how a control package in organizations are structured. Their control package contained five elements: planning-, cybernetic-, reward and compensation-, administrative- and cultural controls. This framework recognizes different forms of control and should be applicable in this research where multiple organizational levels will be addressed. The relatively new existence of the framework can be assumed as a bad aspect of this framework, due to lack of recognition within the prior literature. Moreover, Malmi & Brown do not clearly argue where this typology comes from, i.e. where it is built upon.

Ferreira and Otley’s (2009) PMSs framework.

(13)

12 measurement and does not distinguishes various types of control mechanisms. The cultural aspects of control mechanism are for instance not explained, although it is an important aspect in influencing employee’s behavior (Merchant & van der Stede, 2012).

Merchant & van der Stede’s object-of-control framework (2007, 2012)

Merchant & van der Stede (2007 & 2012) created a framework which is also used in Sandelin’s article (2008) and emerged from the original ideas of Ouchi (1977). The typologies of management control systems that Merchant & van der Stede (2012) distinguish are known as the framework of object-of-control. This framework was introduced by Merchant (1998). It focuses on the spectrum of control practices, but still provides sufficient rigidity with the specific control objects of action, results, culture and personnel (Sandelin, 2008).Although the object-of-control framework is widely used and

recognized, and is applicable at multiple organizational levels; there has been some critique to this framework. Sandelin (2008) argued that the framework lacks specificity with regard to the coupling between control elements because it focuses on the whole spectrum of the control practices. Since this thesis focusses on the coupling between management control elements and systems, it will not only contribute to the theory about management control systems as a package, but also describe the couplings between the control elements of the object-of-control framework. The object-of-control framework will provide a stable basis for analysis, which fits in my research setting. The four typologies are action-, result-, personnel-, and cultural controls.

Because of the downsides of the other four frameworks and because the framework of Merchant & van der Stede has a good fit with the research setting, the choice for the object-of-control framework has been made. The four typologies will be described extensively and examples will be provided.

According to Merchant & van der Stede (2012) action controls are described as the most direct form of control in their typology of controls. Actions controls can prevent organizational actors from doing what should not be done and it controls these actors to act in consistent ways with organizational goals and interests. One of the forms of action controls that Merchant & van der Stede (2012) recognize are behavioral constraints. The word suggests a “negative” and “constraining” form of action control. It ensures that employees are not able to perform actions that should not be done. To illustrate what is meant by these action controls, some examples refer to physical constraints and others to

(14)

13 or code locks on safes. By administrative constraints is meant action controls such as passwords on systems, digital authorization or restriction to approve expenditures above a certain amount of money. Another form of action control suggested by Sandelin (2008) is personal supervision. Usually with this type of action control, a principle supervises an agent by direct observation and will try to coach the agent by guiding the agent in activities and behavior. When an organization is working with action controls it is important to recognize that effective action controls are only available when (1) an organization can determine what actions are (un)desirable; and (2) they are able to ensure that the (un)desirable actions (do not) occur, i.e. are (un)controllable. Aside from extreme situations, action controls can be effective if they are applied in the right context.

Result controls is a reactive type of control. By an ex-ante approach it rewards employees when “good” actions or behavior, which lead to results, are shown. Mostly these actions are in line with

organizational goals. Therefore, a critical question regarding the use of result controls is, what are good actions or what is good behavior? Result controls influence actions and behavior because of the

consequences that are set by the firm. Those consequences are set in such a way that an employee will get rewarded for “good” actions or behavior, i.e. performance in line with the company’s goals. The organization does not dictate the actions that an employee should take, but instead it empowers the employee to decide for its own how to behave. The reactive characteristics of result controls are in that sense present when the employee gets evaluated over the previous evaluation period. In this

evaluation, the principle will review and judge the employee’s performance and reward for “good” performance. These types of control are used at many organizational levels and therefore extremely relevant for the specific research question. A well-known example of a result control is

pay-for-performance. With this type of result control the employees with best performance of the performance indicators, will be rewarded the most. Merchant & van der Stede (2012) give a label to organizations where this occurs; meritocracies. “In meritocracies, the rewards are given to the most talented and hardest working employees, rather than those with the longest tenure or the right social connections” (p.30). The potential downside of result controls are the creation of “bonus cultures” of greed and short-terminism. This implies that employees will do anything to achieve their individual performance level of being rewarded and forget the wellbeing of for instance team performance or related business units. An important conclusion about the use of result controls is that it can work as a good motivator for

(15)

14 Cultural controls can be regarded as a softer form of control. It does not imply restrictions of actions or rewarding after an evaluation period. Cultural controls exist to shape organizational behavioral norms and to encourage employees to monitor and influence each other’s behaviors. These controls are influenced, for example, by formal documents as code of ethics, statements of mission, vision or management philosophy. It starts with the philosophy or vision of the firm; what type of organization does the firm want to be? A “hardcore” profit-making organization?, a firm that contributes to society?, a non-profit organization or a mixture of some of these? The vision statement of an organization determines the world view of a firm and their contribution in their environment. From this starting point, a firm can create a culture within the organization. Merchant & van der Stede (2012) refer to this phenomenom as tone at the top. Other ways they suggest to influence a culture are codes of conduct, for example the code of ethics. This document describes in general the organizational values,

commitments to stakeholders and the ways in which management would like the organization to function. The purpose of these codes is to help employees understand what behaviors are expected even in the absence of a specific rule. Some subjects that are probably taken into consideration in codes of ethics are fair treatment, innovation, risk taking and willingness to change. By formal statements about topics like these, top management wants to create an atmosphere of shared values within the organization.

“Personnel controls build on employee’s natural tendencies to control or motivate themselves” (Merchant & van der Stede, 2012, p. 88). Personnel controls are focused on the employees of an organization and have three main purposes. First, these controls help employees to understand what is expected from them. For example by providing a job design and the necessary resources to perform the job, an organization can create expectations and an employee in response to that, knows how to perform his/her job. Secondly, some personnel controls help ensure to select the right employees for the right jobs. In the selection and hiring process a firm can look at previous experiences of an applicant or to study degrees that are achieved. By those kind of selection criteria an organization can increase the likelihood to attract the employees that fit into the job design and have the required skills. Even if an applicant does not have all the required competences, a firm can choose to make use of another

(16)

15 employees will engage in self-monitoring programs. As Merchant & van der Stede (2012, p. 88) state: “self-monitoring is an innate force that pushes most employees to want to do a good job, to be naturally committed. Self-monitoring is effective both because most people have the conscience that leads them to what is right and are able to derive positive feelings of self-respect and satisfaction when they perform a good job and see their organization succeeding.” These controls can be implemented by selection and placement, training and job design and resourcing.

This typology of management controls will help to shape this paper in recognizing various forms of controlling the behavior of the subordinates in practice.

2.4 Linkages between management control systems

The object-of-control framework by Merchant & van der Stede (2012), that just has been described, will be used as a typology to categorize the different types of controls that are observed in the organization. In order to do research at the linkages between those types of controls, the coupling theory of Orton & Weick (1990) is applied. This theory is suggested by Malmi & Brown (2008) in their call for further research regarding this subject. Coupling theory is a tool that focuses on relationships between coupling elements (Beekun & Glick, 2001). The elements that are described can refer to any organizational phenomena, e.g. processes, incentives, subordinates or information streams, that are related to each other in some way (Weick, 1976; Beekun & Glick, 2001). Orton & Weick (1990) have two variables in their loosely coupling theory. Those two variables determine to what extent (in this case) management control systems are coupled. The two variables are responsiveness and distinctiveness. The presence of either variables are either “high” or “low”. The measurement regarding the presence or absence of responsiveness or distinctiveness seems to have a subjective influence. After explaining the theory where Orton & Weick (1990) tried to convince us with, the prior literature about how to rank in using the coupling theory will be discussed.

Although Orton & Weick (1990) did not visualize their theory, it might be helpful to have an insight about the outcome of the variables and what this means for the level of coupling. Therefore, based on their explanation, a two-by-two matrix is developed (Figure 2). It separates responsiveness from

(17)

16 there is distinctiveness without responsiveness, the system is decoupled. If there is both distinctiveness and responsiveness, the system is loosely coupled.” (Orton & Weick, 1990, p. 205). An example of low distinctive control system may be cultural controls. If cultural controls are designed to be applied organizational wide, it is hard to distinguish those controls at different organizational levels. But due to this design, a change in cultural values at the top level will be directly translated to the other

departments. This implicates that there is a high level of responsiveness. This example of cultural controls is according to Orton & Weick (1990) a tightly coupled system.

Responsiveness High Low D istin cti vene ss H ig h Loosely coupled systems Decoupled systems Lo w Tightly coupled systems Noncoupled systems

Figure 2.Coupling theory (Derived from Orton & Weick, 1990).

In a working paper from 2009 Sandt explored on the approach to research management control systems as a package. In this paper he also used the coupling theory that originates from Orton & Weick (1990). As mentioned above, by using the framework of Orton & Weick it is not always possible to determine the extent to which phenomena are coupled (Sandt, 2009). Especially in case of this research where the relatedness of certain control elements and systems are explained. In practice there won’t be such extreme situations as described by Orton & Weick, i.e. either high or low distinctiveness or

(18)

17 organizations. Their analysis shows a distinction between the vertical and horizontal coupling of result controls. Although their contributions about loosely coupled systems are of great relevance to the field, there is no emphasis on how certain systems are coupled. Since this might contribute to create a better understanding of the mutual functioning of management control systems, the focus of this thesis will be on describing the coupling in detail. The analysis concerns a transparent view on the coupling between two management controls on different hierarchical levels. It uses a detailed description of the (sub-) dimensions that are known in prior literature (Orton & Weick, 1990; Beekun & Glick, 2001; Brown et al. 2007). Beekun & Glick were the first to critically analyze the applicability of the loosely coupling

framework. They identified certain sub-dimensions but did not couple them to the concepts of Orton & Weick. As a result, there were still no substantial triggers for future research to address this

dysfunctionality. Brown et al.’s (2007) where Sandt (2009) refers to, built upon the critics of Beekun & Glick (2001) and found sub-dimensions that are applicable to distinctiveness and responsiveness. This specification of the framework of Orton & Weick can help practitioners to apply the framework better and recognize in practice what type of coupling between phenomena actually exists. Next to describing those sub-dimensions, the practical handles will be visualized and thereby clarify the complete structure of the adjusted framework. Cited from Sandt (2009, p. 9) who refers to Brown et al. (2007): “For

characterizing distinctiveness between elements of an MCS, they differentiate three further dimensions: focus, use and components, for characterizing responsiveness they distinguish the dimensions

dependence, directness and strength.”

Dimensions of distinctiveness

The three dimensions that characterize distinctiveness according to Brown et al. (2007) are focus, use and components. By the focus of an MCS is meant where it is primarily looking at. The distinction can for instance be made between a focus on processes and a focus on people. Differences in use relates to the timing where the MCS is relevant. For example budgets are control instruments that can be used for ex ante purposes, where profit measures are used to review a period of performance, i.e. ex post purposes. Differences in components among MCS can also be an indicator for the extent of distinctiveness. To illustrate what is meant by differences in components, consider an MCS with more than one measure. Those different measures can be considered as components. To separate one MCS from another MCS is harder when the two MCSs overlap and therefore have similar components. Having common

components leads to a low distinctiveness as will be indicated hereafter.

(19)

18 of distinctiveness is given; if two dimensions are given, a medium level of distinctiveness is present” (Sandt, 2009, p. 10).

Dimensions of responsiveness

Three dimensions that characterize responsiveness are dependence, directness and strength. The dependence of one MCS on another MCS reflects the level of exchange between two or more MCSs. In case of reward systems, if performance is evaluated by only one measure, the reward is highly

dependent to that one measure. But if, instead, the performance evaluation depends on several measures, there is a lower extent of dependence for the reward. The second characteristic of

responsiveness is directness, which refers to the urgency of the relationship between MCSs. Beekun & Glick (2001, p. 232) give an example of participants being loosely coupled if they communicate with each other through different levels of hierarchy rather than directly. The strength of the coupling between MCS is determined by the frequency, intensity and probability of change caused by one element to the other. Beekun & Glick (2001) also mention by this definition that certain MCS can be loosely and tightly coupled in one time, e.g. if infrequent conversations have an intense character. When MCSs have a high level of dependence, directness and strength, they are considered to be highly responsive to each other.

The text of the previous two subparagraphs has been summarized and visualized in figure 3, where the (sub-)dimensions indicate the extent of coupling between two systems.

Responsiveness:

Level of dependence, directness and strength High Low Di st in ct iv e n e ss : D iff erenc e in f o cus, use an d c o m p o n en ts H ig h Loosely coupled systems Decoupled systems Lo w Tightly coupled systems Noncoupled systems

(20)

19

2.5 The context of a municipal organization

This section will shortly highlight the specific context of a municipality. Differences between the public and the private organizations are not new to academic literature. In recent papers, the work of Rainey, Backoff & Levine (1976) is often recognized to define the differences between the public and the private organizational context (Baarspul & Wilderom, 2012; Hasty et al. 2012). A basic difference compared to private organizations is that municipalities are an example of organizations that are owned by the state and its citizens. “Citizens therefore have the rights and expectations they do not have in relation to private organizations” (Rainley et al. 1976, p. 239). In private organizations the directors often represent the shareholders of the company. Public interest may play a role in this difference, which is

characterized by a wider scope of concern. Aldermen represent the public interest and are politically responsible for the day-to-day activities of the departments (Budding, 2004). Cristofoli et al. (2010) complement this statement by referring to the crucial relationship that exists between citizens and the city (or the city’s representatives). Politicians are sensitive for the general public opinion and try to act in favor of that opinion, since their goal is to serve the public interest. “Citizens play an important role in influencing service provision and the related control mechanisms through their relationship with the city” (Cristofoli et al., 2010, p. 367). An additional difference is that administrators in municipalities find it harder to measure results, and therefore also harder to attain these results and become more effective (Rainley et al., 1976). The introduction of NPM may play a role in the way that organizations like municipalities are managed and controlled. As discussed earlier, the movement of NPM introduced business-like instruments that shifted the focus of control towards outputs. Because of the differences between the public organizations, and the private organizations generalization can be problematic. Therefore it might be interesting to do research in the public organizations separately.

2.5.1 The hierarchical structure in a Dutch municipality

Because the determination of relationships between management control systems at different

(21)

20 Figure 4. Hengel et al.(2014) about the hierarchical structure in Dutch Municipalities

The political body consists of the municipal council and the board of mayor and alderman. By voting every four years, the citizens have a great influence on the composition of the political bodies. In the bureaucracy, or the municipal organization, there are three levels of hierarchy, these include top management, middle management and the civil servants.

3 Research methodology

The section of research methodology will clarify the type of research, the methods of data collection and the method of data analysis. Tables regarding the data sources can be found in the appendix.

3.1 Type of research

According to Yin (2014) case-studies are most of the time suitable when the research is willing to explain some present circumstance, which is characterized by research questions which addresses “why” or “how” questions. This thesis wants to explain how management control systems at different

(22)

21 possible. Moreover, Malmi & Brown (2008) referred to this research opportunity using the case study approach where the theory will be refined. This research extends the theory about the relatedness of management control systems, but contributes specifically with new arguments about the relatedness of management control systems at different hierarchical levels.

3.2 Data collection

The data collection consists of multiple sources of data, which is called in the literature: triangulation (Yin, 2014). It starts with getting familiar within the organization. Study their structure, their formal information systems and get familiar with their culture, by taking notes on basis of observations and reading formal documents. These data sources can partly be seen as a preparation for the interviews, in order to anticipate better at the context of the organization. Thirteen interviews were conducted, which are recorded and the transcripts of those interviews are available. A conducted interview covered 50 to 80 minutes and the total transcript contains 184 pages, single spaced, Cambria 12. The interviewees are located at different organizational levels and have different internal functions, e.g. managers and subordinates. All interviews have been recorded and the transcripts are available. The aim of those interviews is first of all to verify the functioning of the management control systems on the basis of formal documents. The second aim is to map controls used on top level and department level. Finally, to understand how the control systems relate to each other using the coupling theory. For example, are the controls at department level responsive and distinctive to the controls at the top level? Which control types do relate and how do they relate to each other? The focus on these aims will help me to find a pattern in the linkages between management control systems at different organizational levels. The interviews contained of non-directive questions (Duck & McMahan, 2012).This means that the question will not directly relate to control packages or the coupling theory. The aim in the interviews was more about the story of the interviewee and the questions directed the subject of the interviews.

3.3 Data analysis

(23)

22 An overview of the management control systems was made by using the above described data sources. The first step will be to categorize all controls used at the different organizational levels. The transcripts are marked at phrases where the controls out of the object of control framework were described. Out of these markings an overview is created of the management control package. At this point, the data sources are prepared for the analysis to find any coupling between management control systems at different hierarchical levels. For example, is the target in the pay-for-performance system (Result controls) directly related to the mission and vision statement (cultural controls), to the training-targets (personnel controls), or is it a standalone system? And if, to what extent, according the coupling theory (Orton & Weick, 1990) are the management controls related? On the basis of the coded transcripts, coupling patterns between two control systems are found. In addition, the notes from informal and formal documents are used to verify conclusions from the transcripts.

The analysis concludes with the control systems (Merchant & van der Stede, 2012) that are related to each other in the organization, and thereby refine the theory about management control systems as a package.

4 Analysis

The analysis section will depict the case study that is conducted at the department Policy and Direction (P&D) (in Dutch: Beleid en Regie (BNR)) at the municipality of Emmen. Section 4.1 will shortly introduce the organization that is the object of research. In section 4.2, the characteristics of the department will be described and the activities will be listed. In this section, the management controls will shortly be named. Section 4.3 contains the relationship between those management controls that will be

described on the basis of the coupling theory by Orton & Weick. Arguments will be based upon quotes and detailed descriptions of the contextual working of management controls. The last part of this section is about the relationship between various management control systems. In this section the coupling theory will be used to build an argument about the way in which systems are coupled.

4.1 The municipality of Emmen

(24)

23 city, such as the famous zoo of Emmen, it also offers enormous space and rest by its large surface. The organization of the municipality of Emmen is characterized by the willingness to involve the citizens into the decision making process. The statement “more from the society, a different government” is

widespread across the employees of the municipality. It illustrates how the municipality wants to serve its citizens. In an interview with an employee at the department P&D this vision was confirmed by the following quote:

“I can easily write a report behind my desk, but if we want to profile ourselves in terms of “more from the society, a different government”, then the report won’t come any further than my desk drawer and it will become worthless for the citizens”

This quote indicates that the employees of the municipality of Emmen involve the citizens within the transformation process in order to increase the value that citizens attach to the services that they offer. This ideology or idea is shared among the interviewees. The involvement of the citizens in writing and executing the policy is definitely part of the norms and values of the department.

The services that are offered vary widely: economy, employment, education, traffic, constructions, sports, leisure and healthcare are examples of services where the municipality is involved in. The amount of involved employees in the organization and operationalization of these services is 1230 people. Those 1230 employees are spread over 15 departments next to some additional director functions.

4.1.1 Hierarchical structure at the municipality of Emmen

In comparison to the ideas of Hengel et al. (2014) that are depicted in figure 4, the municipality of Emmen has the same hierarchical structure. The political bodies are exactly structured the same as in figure 4. In the bureaucracy, the top management consists of the municipal manager, the directors and some staff functions that facilitate management. The middle management consists of all the 15

department managers. The hierarchical layer of civil servants is led by team managers, where the subordinates are the direct responsibility of the team manager.

4.2 Department P&D

The department P&D is a department of the municipality of Emmen where the policy about

(25)

24 degree. They are expected to follow the social developments very closely. Those developments are complex, dynamic and are changing rapidly. To answer those developments, the department formulates policy frameworks and execution programs. Their operations require a wide orientation on the society and work from different angles. In their daily activities the employees manage their own time and predominantly do not have to stick to procedures in writing policy. In the next quote the vision of the subordinates about the role of the team manager and their own position is indicated. This quote came from a subordinate from team social:

“we trust on our team manager that he will treat us in the same way in the future more or less like professionals and that he gives us space for our own responsibilities”

This vision is shared among the employees. None of the interviewees ignored this vision. The team managers encourage their employees in being professionals. Freedom and responsibility is given as long as employees meet the expected results. Outputs from the departments always take into account the interests of internal and external partners. The next quote from a subordinate of team Economy visualizes that:

“.. if a plan of an entrepreneur is economically perfect but is does not fit in whatever regulation or policy, than you need to start the discussion internally and externally. That is also a form of facilitating the society.”

The operationalization of this output is monitored by the responsible employees. The results and feedback out of the monitoring process is used to learn from and to formulate adjusted policy and develop new policy.

(26)

25

4.3 Coupling between management control systems

The management controls that are described above function as the basis for this paragraph. The

coupling of management controls at different hierarchical levels will be described in this paragraph. This description uses the dimensions and sub-dimensions that are known in the academic literature (Orton & Weick, 1990; Beekun & Glick, 2001; Brown et al., 2007). Unfortunately not all couplings between the different management controls can be described, due to time and space restrictions regarding the thesis period that are taken into account.

“Management controls include all the devices and systems managers use to ensure that the behaviours and decisions of their employees are consistent with the organization’s objectives and strategies” (Malmi & Brown, 2008, p. 290).

Merchant & van der Stede (2012) define, in line with Malmi & Brown (2008), the function of

management controls. These definitions were helpful in the decision of which couplings to describe. The management controls that seem most relevant in influencing behaviors and decisions in

subordinates’ daily jobs are included in the description of the following three couplings. The relevance is determined by the extent in which management controls are mentioned by the employees in interviews, observations and their relative importance in formal documents. The structure of

describing the coupling is as following: the first paragraph describes two management control systems, where each system is located at a different hierarchical level. The second and third paragraphs describe the distinctiveness and responsiveness between the two management controls and the last paragraph concludes on the type of coupling between the management controls. The described couplings are; between the budgeting program and the individual work plans, between the A3 annual scheme on concern level and the A3 annual scheme on department level, and finally between the strategic HR planning and the conversation cycle.

4.3.1 Coupling between the budgeting program and individual work plans

The first coupling is the political influence on the day-to-day activities at the department of Policy and Direction. Every four years, the citizens of the municipality vote for the composition of the political council. The coalition within the council generates an executive agreement for the coming four years, where in abstract terms the course of action of the municipality is described. In the time period of four years, this executive program gives direction to objectives for the individual work plans at the

(27)

26 called the budgeting program and can be seen as a quarter of the executive agreement, which is

dedicated to a specific calendar year. The relevant subjects in this budgeting program are directly translated and distributed to the individual work plans of the subordinates at the department of Policy and Direction. The individual subordinates are given an amount of hours on annual basis in order to indicate the amount of time to spend on certain policy programs. The translation of the budgeting program to the individual work plans makes it more concrete for the subordinates to know what is expected from them. Next to the focus on what should be done by whom, the individual work plan also controls the way how it should be done by emphasizing to the core values and the identity of the municipality. Because of having multiple sources of control elements, the individual work plans can be seen as a management control system (Grabner & Moers, 2013).

Distinctiveness

First of all, the budgeting program is located at the political level and the individual work plans are dedicated to the civil servants. In terms of distinctiveness, the focus of the budgeting program is looking at processes (what should be done by the end of the year?). The input for the individual work plans originates from the budgeting program. This input is being translated to assign specific task to certain subordinates. This translation makes the individual work plans much more people oriented than the budgeting program. Both management controls are practically used equally. Both management controls are made for ex-ante purposes, where monitoring during the year is available in order to control

behavior and direction. But in the same time, both management controls are used ex-post, where results are evaluated after a year. The components of the budgeting program and the individual work plan are different. The following quote from a team manager is indicating the relation between the two management controls:

“The budgeting program is clearly indicating what needs to be done for the coming year. Program line two is the program line of the Economy.. In program line two we observe the actions and activities that need to be done, has to be worked on and finally needs to be ready at the end of the year. These actions and activities are translated in to the individual work plans.”

(28)

27 Responsiveness

The political influence is visible in the individual work plans. In order to create a good view about the responsiveness, interviews took place with subordinates, team managers, an HR-advisor and the department manager. An illustrative quote about the responsiveness came from a team manager at the department P&D:

“Whatever is decided on the political level, has directly effect on my team. If the politicians decide for the execution of an action plan in order to create jobs, than my team has to create an action plan that creates jobs. It has a direct impact”.

All interviewees recognize the high influence of the political body in their daily activities. The quote illustrates the power that the politicians have regarding controlling the behavior and actions of civil servants. The budgeting program determines the direction of actions. Looking at the dependence of the individual work plans on the budgeting program, it seems rather high. The individual work plans depend upon the budget program in terms of subjects that are relevant to the policy writers. For example, the budget program’s states “Create more jobs for citizens older than 55 year”, than the first matter is the subject: employment. The second matter is the direction of employment: increasing the number of jobs on the market, especially for citizens older than 55 years. The job for the subordinates at the

department of Policy and Direction will be to pick up this signal and design policy and action plans in order to stimulate this intention. The relation in terms of directness between the budget program and the individual work plan is also rather high. The budget program serves as a direct input for the

individual work plans. It determines the allocation of priorities in terms of subject and course of action. The strength between the two management controls also seems to be high. Referring to Beekun & Glick (2001), the strength is determined by the frequency, intensity and probability of change caused by one element to the other. Since the budget program determines the input of the individual work plans, the probability of change is high because work plans are adjusted to the budget program on an intense level, which is also illustrated by the quote above on this page. Another quote that indicates the influence of the political bodies is given by a subordinate from team Economy:

(29)

28 by changes in the budgeting program. Therefore a part of the individual work plan is not responsive, but the focus in the description of this coupling is to the responsive part of the individual work plan.

Type of coupling

The coupling between the budget program and the individual work plan tends to be loosely coupled. It indicates that the two management controls are medium to high distinctive and have a high level of responsiveness. It seems that the budget program separates itself from the individual work plan by having a wider scope than the individual work plan. In addition, it focuses on processes, where the individual work plan is composed for a specific human being and thereby focusing on people. The responsiveness can be explained by the political power that is present. A nice illustration about this power is also displayed in the quote by the team manager of Economy on page 23. The individual work plan is a management control system where the budgeting program is one of the sources of input. Another source of input is, for example the A3 annual scheme, which is part of analysis after this paragraph. The different sources of input and the difference in focus (, process focused versus people focused,) causes the distinctiveness between the management controls.

(30)

29 Distinctiveness

The focus and components of both A3 annual scheme are equal. It is structured in the same way and has the same main parts. The concern scheme is written in abstract terms which are suitable for all 15 departments in the municipality. The department scheme is translated to the local context and adjusted to more concrete and relevant goals for the specific department. An interesting quote concerning this translation came from a department manager, who said:

“The EFQM model is a model that we used the last ten years in order to continue with speaking the same language in the organization. It is also an model where you emphasize on all elements of leadership, processes and employees. The EFQM model is the basis for the A3 model. Thus, in that way cohesiveness is emerging”

The quote illustrates the application of the A3 model and the ambition for goal congruence by using the A3 model. In terms of use there are no differences. In principal both A3 annual scheme are evaluated at the end of every year, but monitoring during the year is also possible. The feedback from both schemes is used in order to make a new A3 scheme. A Plan-Do-Check-Act cycle is being used in the annual cycle of the A3 schemes. The differences in focus, use and components are low, which indicates a low level of distinctiveness.

Responsiveness

The abstract goals and ambitions that are stated in the concern scheme function as the basis for the A3 scheme on department level. There are some goals that are literally copied from the concern scheme to the department scheme. For example “at the end of 2014 the planned cut offs are 100% realized”. Other goals are translated to the contextual situation at P&D. For instance the ambition on concern level that says: “A municipal organization that is visibly involved with the society”. On department level this ambition is translated into:

“- Offer internships (ambition 1)

- offer experience placements (ambition 1)

- Minimal 6 inspiration lectures that are attended better than last year”

(31)

30 in hierarchy, the urgency is high. The level of responsiveness between the concern scheme and the department scheme seems to be high, indicated by the level of dependence, strength and directness. Type of coupling

The department scheme seems to be highly responsive to and has a low distinctiveness on the concern scheme. Therefore, according to Orton & Weick (1990), those two management controls are tightly coupled. The abstract ideas that are created in the top of the bureaucracy about the identity of the organization are spread out to all departments by using the A3 annual schemes. This translation to departmental context develops the A3 scheme to become more concrete and relevant to managers and subordinates at department level. This tight coupling creates goal congruence. In addition, as indicated by a department manager in the quote above, it ensures that the employees in the organization speak the same language. It remains the vision of how the municipality is willing to take further steps, which also contains cultural controls. Using these tightly coupled management control systems creates cohesiveness through the organizational layers. It also forms the way in which the municipality profiles itself to outsiders and how subordinates represent the organization; a cultural mechanism. This is in line with Malmi & Brown’s (2008) definition of the functioning of management controls.

4.3.3 Coupling between strategic HR direction and conversation cycle

The top of the bureaucratic organization has developed a long term policy for use of human resources in the municipal organization. This policy is named the strategic Human Resources direction and contains subjects like the size, quality and composition of the workforce. The document starts with an analysis of the current situation and concludes by formulating core challenges and a strategic focus regarding the HR policy. In this manner, the progress is monitored. One instrument to translate these ambitions to the operational work floor is the conversation cycle. A team manager mentions the functioning and the origin of this conversation cycle in the following quote:

“..in the context of the new HR policy, we are now going in to a cycle. Every subordinate is speaking with his team manager at the beginning of the year about a personal development- and work plan. During the year there is a second meeting about “Where are we now? What do you need? Do we need to adjust or do you need some help of a college?” At the end of the year a balance will be made about the product you made, what did not happen according to the planning and what is the reason for what? A sort of performance appraisal.”

The interviewed team managers and subordinates value the conversation cycle because of giving grip on individual progress during the year. Subordinates indicate that these three meetings are a very

(32)

31 attention to the progress regarding the expectation and to discuss personal development or career development. This cycle is mentioned by all of the interviewees and the ideas about the conversation cycle are shared among the teams.

Distinctiveness

The strategic HR policy is originally an organization-wide policy and is designed to be applicable for all organizations. The conversation cycle is very personal. The team manager has a crucial role in managing the subordinates’ ambitions and meeting the organizations’ expectations. The strategic policy is focused on long term challenges of the organization and the conversation cycle focuses on a annual planning of one specific individual. This indicates the difference in focus between the two control systems. In terms of use, the strategic HR planning complements the conversation cycle. The HR policy is a constant factor, while the conversation cycle is relevant at three specific points in a year. Regarding the timing of the two control systems there is a difference. Components should be the same, but reality differs because team managers have different interpretations and thereby only use some components that function well to their own perception. This is also visualized in the next quote. Overall there tends to be a medium to high distinctiveness between the two management control systems.

Responsiveness

In general the conversation cycle became operational from the renewed strategic HR policy. In that sense there is an extent of responsiveness. In practice the intentions seem to differ sometimes from the actions. In the following quote a team manager describes his experiences with the hierarchical influence on the conversation cycle:

“..we have very strict formats. I have to admit that I do not always stick to those formats. There are subjects on those format that are not relevant for my team members. In those situations it is more like “what shall we fill in here?” and then it is okay. On the other hand there are subjects that are discussed in depth and those are put in to other boxes. In that sense we did not always stick to the strict formats.” Although there is input from the HR policy, team managers tend to bring their own ideas and

interpretations in practice about relevancy of the policy. This influences the level of dependence, directness and strength of the relationship. Because of this independent attitude, the responsiveness tends to be low to medium.

Type of coupling

The coupling between the two management control systems tends to be decoupled. This is

(33)

32 there is deviation between the intentions and input from the strategic HR policy. Regarding the

conversation cycle, managers tend to do what is best and most relevant to their own perception, instead of following the line of top management. This phenomenon creates a clear distinction between what is meant by top management and what is done in practice. In addition, the manager does respond partly to the HR direction but also determines his own path and

direction according to own norms, values and level of importance. This coupling is not extremely decoupled, but according to the coupling matrix it tends to be decoupled. In the following picture, the position in the matrix is visualized. It indicates that there is some responsiveness and a medium to high distinctiveness.

Figure 5. Position of coupling in the coupling theory-matrix

4.3.4 Results out of coupling descriptions

In this section, the results of the described couplings will be visualized in an overview (figure 6) in order to make it more clear how the management control systems are coupled and what causes these couplings.

(34)

33

Coupling Distinctiveness Responsiveness

Loosely coupled systems: Budgeting program and individual work plan.

High: Difference in process- and people-focus. Difference in components: other management controls influence design and composition.

High: Budgeting program serves as input for individual work plan. Context of P&D is sensitive for political influence. Budgeting program does not influence entire content of individual work plan.

Tightly coupled systems: A3 annual scheme concern and team level.

Low: Same focus, design and components.

High: The concern scheme serves as input for the team scheme. The only difference is the adjustment of abstract terms to contextual concrete terms.

Decoupled systems: Strategic HR direction and conversation cycle.

High: A difference in focus. The organizational challenges are translated to personal plans. Components should be equal, but do differ because of difference in actions and intentions.

Low: There is partly responsiveness, but managers let their own influence dominate the content of the conversation cycle. In that sense the content of the conversation cycle is independent from the hierarchical influence.

Figure 6. Overview of couplings between management control systems. (1) the extent in which a control system is build up out of different control systems or elements.

It seems that if more control systems give input to a certain control system, the control system needs to deal with different interests. This makes it harder for the control system to have congruence with all influences, for example because of conflict of interests. In this situation, the control systems have a higher distinctiveness. When a control system represents the only source of input for another control system, it creates a different situation. The hierarchical control system creates organizational-wide expectations. These expectations are translated to the departmental control system in order to make to make it concrete in the contextual situation. In this situation the control systems have a low

distinctiveness.

(2) the extent in which managers or subordinates are accountable for their actions.