Management Control Packages

(1)

Management Control Packages

Explaining configuration variety

Master thesis

by

Pier Cornelis Peter Hager

University of Groningen

Faculty of Economics and Business

MSc Business Administration:

Organizational and Management Control

August 2012

Name: P.C.P. Hager

Address: Anna Paulownastraat 50a 9725 JV, Groningen Phone: +31 (0)6 – 20 45 32 61 E-mail: p.c.p.hager@student.rug.nl Student number: 1557262

Supervisor: B. Van der Kolk MScBA

(2)

Preface

With this master thesis my studies at the business faculty of the RuG come to an end. I consider this period of my life as one of exploring and learning that such intangible concepts as business and economics can be approached in structured and scientific ways, which I found always very fascinating. I have also learned that the end of these studies does not mean the end of learning, as there are so many things left to learn and worth looking at.

I would like to thank my father for his continuous support during my studies and the writing of this thesis. Furthermore I would like to thank my girlfriend for giving me continuous feedback while writing this thesis. I specially like to thank my supervisor mr. B. van der Kolk for his positive and constructive conversations, ideas and feedback which I have experienced as very pleasant. Finally I want to thank my co-supervisor prof. dr. H.J. ter Bogt for taking the time to read this thesis.

(3)

Abstract

Within an organization various Management Control Systems (MCSs) can be enacted, which implies that MCSs do not operate on stand-alone basis. The Management Control Package (MCP) is defined as the metaphor for the simultaneous existence of a number of Management Control Systems in an organization. But what drives the configuration of these Management Control Packages? This concept as a whole received minimal attention in the research genre thus far. Therefore the research question that needs to be answered is:

Which factors influence the configuration variety of Management Control Packages?

This insight is gained based on the systematic categorization of developed Management Control literature during the past decades. 122 publications in the timespan 1978 to 2011 from the journal ‘Accounting, Organizations and Society’ were analysed to develop a model including control typologies based on the framework of Malmi and Brown (2008). Furthermore for each publication the configuration factors and their influences were recorded. This literature analysis resulted in 147 Management Control System observations leading to a list of 68 factors of influence in a total of 289 factor occurrences.

This thesis proposes a model to define factor antecedents that includes a ‘technically and/or economically required’ antecedent and a ‘psychologically and/or culturally determined’ antecedent. Antecedents of the factors provide insight into configuration variety of MCPs as the majority of the factors (67%) share a technically and/or economically required antecedent. Patterns of structural co-existence of Management Control Systems are analysed and found, and most specifically for the Cybernetic, Planning and Administrative types of control systems. Detailed models are presented on the factors and their influence underlying every control typology in the framework. Out of the 68 factors 19 are found to be of linear influence on more than one type of Management Control System, therefore able to explain something about Management Control Packages. Though all these 19 factors have influence on the package, ‘Strategic Focus’ has the largest shaping effect within the observed literature. The factors ‘Competitive forces’, ‘Electronic/information integration’ and ‘Environmental Uncertainty’ follow respectively as the most important in explaining configuration variety of Management Control Packages.

The results combined allow making some conclusions on the factors underlying management control package configuration variety, however this research should be considered as a starting point and more research is needed to deepen insights and develop explanatory theories.

Key words: Management Control Packages, Management Control Systems, Management

(4)

1 Introduction

In recent decades the amount of empirical research and research directions on Management Control Systems (MCS) has grown. Consequently, various interpretations emerged of what constitutes a Management Control System. For example, Merchant and Van der Stede (2007) define management control systems in terms of systems that deal with employees’ behaviour. Anthony (1965) however, presented an interpretation of management control that encompassed more than solely employee behaviour and defined management control systems in terms of a system in which managers attempt to obtain and effectively and efficiently use resources to accomplish organizational objectives. Though these two examples of definitions overlap, an issue remains regarding the precise definition of content and span of control of a MCS. Does a MCS focus solely on employee behaviour, or should it include more widespread organizational resources such as materials and liquid assets? This variation in definition of what constitutes a MCS left researchers with some issues with regards to the interpretation of research results and the design of MCS. Malmi and Brown (2008) suggest a clarification of this issue by starting with the managerial problem of directing employee behaviour. “Those systems, rules, practices, values and other activities management put in

place in order to direct employee behaviour should then be called MCSs (Malmi & Brown, 2008, p. 290)”. This thesis will follow this interpretation as it puts a clear and specific focus

on directing employee behaviour, while allowing for interpretation of various systems, rules practices and values within this definition.

Within an organization various management control systems can be enacted, which implies that MCSs do not operate on stand-alone basis. This is represented in the idea that there is no single Management Control System, but that various systems may work as a related set of concepts in the metaphor of a package. For decades this idea of a Management Control Systems as a Package (MCP) exists (Chenhall, 2003; Dent, 1990; Flamholtz et al. 1985; Otley, 1980; Malmi and Brown, 2008), however, the question that remains unanswered is how the different research streams relate to each other and which antecedents of management control package configuration the literature as a whole provides.

(7)

While the management control system in itself performs the actual controlling of the behaviour of employees within an organization, it is important to understand how these systems function and interact within the context of other organizational control systems. This could provide new insights in what an effective management control should look like. Gaining this broader understanding of MCS may facilitate the development of better theory of how to design a range of controls to support organizational objectives, control activities, and drive organizational performance. MCS research has shown that organizations configure Management Control Systems in distinct and unique ways, which in this thesis will be labelled ‘configuration variety of Management Control Packages’.

A method to gain these insights into configuration variety is to analyse published management control literature and classify this literature according to its characteristics and possible position within a management control package. It is also important to take a look at the factors in the literature that were shaping that particular MCS configuration. Existing literature is considered essential here as the stand-alone results that this literature offers, when systematically categorized and analysed, could provide valuable insights in generic paths organizations follow when designing, configuring and implementing their management control systems. This will lead to an overview of literature into distinct typologies by the use of a classification framework which, combined with identification of the shaping factors behind that literature, could provide valuable insights into the causes of configuration variety within MCS.

The goal of this research is to analyse management control literature and develop these insights into these factors that underlie MCS configuration. This insight will be gained based on the systematic categorization of developed Management Control literature during the past decades. Malmi and Brown (2008) have proposed a conceptual typology to classify MCS systems that is specifically suited for Management Control Packages. The categorization of the literature will take place based using this classification scheme. For each study, results will be examined to identify the main configuration factors, which will then be linked to the typology. This research approach will lead to an integrated overview of a set of typologies and related factors that are the foundation of MCSs selection, and thus form the basis for selection and configuration of systems within a Management Control Package as such.

Based on the results of this research, the main research question that this thesis aims to answer is:

(8)

Sub questions:

What is a Management Control System? What is a Management Control Package?

Which factors can be identified in existing accounting literature, that influence the

configuration variety of Management Control Systems?

How could the findings from the literature research potentially explain configuration

variety of Management Control Packages?

This research will start with defining Management Control Systems and Management control packages. Next the methodological approach for the research is discussed. Afterwards the results of the literature review will be presented, followed by a section on potential implications of the results on management control as a package. Finally a conclusion and some limitations and directions for further research are given.

(9)

2 Management Control Systems

As Management Control Systems form the principal components of a Management Control System Package, considerable attention is paid to this subject by this thesis. In this chapter the concept and foundation of Management Control Systems is addressed by discussing a selection of various relevant management control research during the past decades. Consequently, the first sub-research question of this thesis will be answered: ‘What is a management control system?’

2.1 – Initial definitions of Management Control Systems

Anthony (1965) defined management control in terms of a system in which managers attempt to obtain and effectively and efficiently use resources to accomplish organizational objectives. During the 1960s and 1970s researchers built on Anthony’s work and that of others by attempting to develop the best way to design and use formal systems to help organizations implement their strategies and objectives (Simons, 1990).

Ouchi (1979) defines management control as a process by which an organization tries to obtain cooperation between individuals who might share partial congruence in their objectives. Based on this assumption this research presented three control mechanisms: market, bureaucracy and clan controls, each suitable for a specific situation of control needs. According to Merchant (1982) until then most organizational control research included only feedback and adjustment processes. Merchant proposes that if individual goals would be congruent with organizational goals, control would not be needed. However as congruence due to unwillingness or inability of individuals might differ, a set of controls must be implemented to encourage desirable actions and to protect for undesirable behaviour. Merchant proposes control archetypes, based on the output measurability of the organization, with a focus on people, action or results control. All archetypes are created based on the functional definition that the purpose of a control system is to achieve organizational goals, with the help of individuals (Merchant, 1982).

In 1988 Anthony differentiated between the nature of controls as formal and informal controls. Formal controls include rules, standard operating procedures and budgeting systems. This type is the more visible, objective components of the control system. Informal controls are more intangible control mechanisms such as corporate culture, norms and values.

(10)

2.2 - Agency theory

Agency theory also focuses on goal congruence but has its routes in information economics. It states that principals and agents might attain potentially differing goals and also attain different risk preferences in their contracts (Jensen, 1983). Eisenhardt (1989) based a control study on agency theory and discussed how principals can exercise control over the activities of agents and how the optimal contract to approach control over the agent’s activities can be created. A distinction between two types of optimal management control on the contract is made, namely a behavioural focus and a results focus. The right type of control should allow for a realignment of goals between the principal and agent, so both parties desire the same outcome. The choice between the types of control depends on the amount of information about the activities of the agent that is available to the management. When no complete and accurate information is present, the results control is the most suitable control approach. When complete and accurate information is available, the behavioural control is the more optimal approach. The essential aspect for this approach is the addition and specific call for information availability within Management Control configuration.

2.3 - Organizational and strategic control

It could be argued that the research discussed until this point builds on relatively generic approaches and is less based on the specific analysis of organizational practices and needs for organizations with a more flexible character that possess specific needs.

Otley (1994) types these kind of companies as contemporary organisations, which are based on relatively new management concepts such as 'de-layering', 'right-sizing', 'business process re-engineering', 'out-sourcing' and 'value chain analysis'.

Some researchers made the call that a management control system is related to organizational and strategic control. Simons (1990) for example states that Management Control Systems can not only be used to implement strategy, but also can be seen as systems which can assist in the formulation process of organizational strategies. His central argument is that Management Control Systems become an actual competitive advantage of an organization by allowing for strategic control. Langfield-Smith (1997) critically reviewed this relationship between MCS and strategy. They conclude that it is not clear what role MCS can play to bring intended strategies to realization, or whether MCS can absorb negative side effects caused by strategic change, particularly when changes are undertaken over a considerable period of time.

2.4 - Contingency approaches to management control

(11)

conflicts were a result of both the differences in research designs and the way in which control and effectiveness were operationalized and measured (Langfield-Smith, 1997).

Together with the development of contingent research increased attention is paid to the behavioural aspect of management control, for example by the Balanced Scorecard theory (Kaplan and Norton, 1992). This control proposition empowered non-financial performance measures and included an organizational learning and growth perspective in a control environment. The contingent approach is incorporated in this respect by arguing that an organization should design a situational specific control system guided by a theoretical line of reasoning in four pre-defined causal perspectives: financial, customer, internal business processes, and organizational learning and growth. Later however this causal-deterministic relationship was criticised by Nørreklit (2000), arguing that the perspectives are not causal but inter-related.

An important underlying element of this thesis is the investigation of management control from a holistic perspective that encompass controlling aspects of organizations in its entirety, in this thesis in the form of the concept of Management Control Package. A lot of MCS theory however deals only with partial elements of controls, meaning not every element of a control setting within organizations is explained or researched, and provides therefore only a fragmented approach on how control should be exercised. This raises question and increases interest for insight in the range of controls that support organizational objectives, control activities, and drive organizational performance, and why these controls vary between organizations.

2.5 - Transaction cost approach to control

Speklé (2001) discusses control variation by questioning why some organizations use extensive, formal systems to direct their operations, whereas other organizations seem much less planning their aims and actions. The problem of contingency theory as an explanation according to Speklé is that it represents more a general idea than a well-defined theory based on relevant factors and their likely consequences. Furthermore he argues that rather similar problems exist in agency theory in the sense that agency theorists focus mostly a limited number of well-speficied MCS elements rather than capturing control in its entirety. Transaction Cost Economics (TCE) central aim is to explain why some transactions are more likely to be executed within one form of organization, whereas other transactions tend to be associated with different organizational modes (Williamson, 1979). Speklé (2001) transforms this concept to Management Control literature by arguing that organizations implement differing control structures, based on their control and situational requirements and the problem-solving ability and costs of the control mechanism. This resulted in a framework of nine control archetypes that are primarily based on ‘market’, ‘arm’s length’, ‘machine’, ‘exploratory’ or ‘boundary’ control.

2.6 – Towards a holistic perspective on Management Control Systems

(12)

Furthermore the scope of the management control system extended from purely financial elements towards the inclusion of non-financial measures within control systems. Some theories present overlap in their assumptions while others are conflicting or impose a fragmented effect to the field.

In my opinion management control is about directing employees, however, including a more multi-disciplinary focus in control systems has altered the ingredients for the provision of direction. The increasingly complex and heterogeneous environments, and thus diverse control situations, organizations are facing nowadays combined with aim of this thesis to investigate Management Control from a complete control perspective creates a need for a broad definition and classicisation of Management Control Systems.

Malmi and Brown (2008) presented a Management Control System as: “those systems, rules,

practices, values and other activities management put in place in order to direct employee behaviour (Malmi & Brown, 2008, p. 290).” This thesis will adopt this MCS definition. The

actual configuration of these systems, rules and practices depends on control requirements that are based on organizational environments, control costs, and problem solving abilities. This forms the basis for the concept of configuration variety in this thesis. The latter of this thesis will focus on explaining this configuration variety within management control packages, but first we will take a look at how multiple control systems function simultaneously in organizations in terms of Management Control as a Package.

Table 2.1 – Theory overview

Authors Control definition Control typologies

Ouchi (1979)

System by which an organization tries to obtain cooperation between individuals who might share partial congruence in their objectives.

Market – Bureaucratic – Clans

Merchant (1982)

Controls must be implemented to encourage desirable actions and to protect for undesirable behaviour of individuals.

People – Actions – Results

Anthony (1988)

Process by which managers influence other members of the organization to implement the organization's strategies

Formal – Informal

Eisenhardt (1989)

Control should allow for a realignment of goals between the principal and agent, so both parties desire the same outcome.

Behavioural focus - Results focus

Speklé (2001)

Control is about focus on the behaviour of organizational participants and a concern with the effect of this behaviour on organizational

outcomes.

Market - Arm’s length – Machine -

Explorartory - Boundary

Malmi and Brown (2008)

Systems, rules, practices, values and other activities management put in place in order to

Planning – Cybernetic – Administrative – Reward

(13)

3 Management Control Systems as a Package

This chapter will elaborate on the Management Control as a Package by presenting the ideas behind the concept. Furthermore a definition and various theoretical ideas about the concept will be presented.

3.1 - Theoretical development of Management Control as a Package

For decades the idea of a Management Control Systems as a Package (MCP) exists and multiple authors have written about the concept (Chenhall, 2003; Dent, 1990; Flamholtz et al. 1985, 1987; Otley, 1980; Malmi and Brown, 2008).

Flamholtz (1983) presented a theory that budgeting and even an accounting system should not be handled as independently operating systems. Instead, those systems should be managed as components of a total system of organizational control. When these linkages between the systems and other necessary ingredients of a control system are not sufficient, then the system may not achieve its intended functionality. This research led to the development of an integrative framework of control (Flamholtz et al., 1987). This model combined views of administrative management, organizational sociology and organizational psychology. The model consisted of a core control system established within the organizational structure, culture and the external environment. Flamholtz defines control as the effort of an organization to improve the probability of individuals achieving desired organizational goals (Flamholtz et al., 1987). This definition of management control is very similar to the definition handled by this thesis, as described in the previous chapter.

According to Flamholtz et al. (1987), control of work behaviour is accomplished by the four core control mechanisms of planning, measurement, feedback and evaluation-reward. When taken together, the model and the propositions comprised a basis for an integrative framework of organizational control, in this thesis discussed in the metaphor of a management control package.

In the same period Macintosh and Daft (1987) also provided an approach to the concept of management control as a package. In their research they investigated the relation between more general control systems and departmental interdependence. They suggest that a management control system is one element in a package of control systems. Their research investigated the relationship between the organizational characteristic of departmental interdependence and the design and use of elements in a package of management controls. The findings support the hypothesis that departmental interdependence is related to the emphasis placed on each management control system.

(14)

3.2 – Reasons to study MCS as a Package

Studying Management Control Systems as a Package is important as within control literature to an increasing level awareness is raised that Management Control Systems do no operate on stand-alone basis. However, much of the MCS literature presents theories as single solutions based on specific contexts, but many control systems are part of broader control systems (Chenhall, 2003). If the interacting effects between these systems are not considered, then we can make potentially invalid conclusions about effects of the single MCS elements (Fisher, 1998). These conclusions can turn invalid because the theories and conclusions are drawn with the goal of explaining the development, adoption, use and impact of individual systems (Malmi and Brown, 2008). These conclusions however can be subject to situational contingent variables and factors that are shaped based on a potential influence of the package of control systems within an organization.

Furthermore, a major focus of the MCS theory is on the formal design of MCS systems in order to produce desired outcomes. The knowledge on the impact of other, more intangible types of control such as cultural controls is limited (Malmi and Brown, 2008). Therefore, the improvement of knowledge on the interacting effects between the more intangible and tangible control systems within contexts of different Management Control Packages is important. If more insight could be gained in the position of and factors and effects that shape Management Control Systems within the entirety of an organizational domain, it could facilitate the development of improved theory of how to develop a range of controls to support organizational objectives, control activities, and drive organizational performance.

3.3 - Package control typology

(15)

Table 3.1 – Management Control Typologies (Malmi and Brown, 2008)

Typology

Sub-typology Definition

Planning Provides direction by providing goals to be achieved across the functional areas of the organizations.

Long-term planning Longer term more general goals, usually for one year and more.

Action planning Tactical focus. Detailed goals, focused on short term (less than one year).

Cybernetic

Measures to enable quantification of a phenomenon, together with performance standards and or targets that is regularly compared with feedback after which an analysis follows.

Budgets Planning acceptable levels of behaviour and evaluating performance against

those plans.

Financial measurement More narrow measures that can be used in financial target setting.

Non-financial measurement Overcome some of the perceived limitations in financial measures and to

identify the drivers of performance.

Hybrid measurement Contains both financial and non-financial measures.

Reward and compensation Motivating and increasing the performance of individuals and groups through attaching rewards to control effort direction, effort duration, and effort intensity. Administrative Control and direct employee behaviour, trough:

Governance structure Monitoring and defining accountability and responsibility.

Organisation structure Organizing individuals and functional areas.

Policies and procedures Specifying how tasks and procedures should be performed.

Cultural The values, beliefs and social norms which are established that influence employees behaviour.

Value-based controls Definitions of basic values, purpose, and direction for the organization.

Clan controls May relate to cultural controls on group level within the organization such as an

organisational unit or division.

Symbolic controls Visible expressions, such as building/workspace design and dress codes, to

(16)

3.4 Definition of Management control as a Package

By combining the insights from the studies discussed (Flamholtz, 1983; Flamholtz et al., 1985; Macintosh and Daft, 1987; Malmi and Brown, 2008) a definition of management control systems as a package is defined that answers for the second research question ‘What is

a Management Control Package?’. This thesis will define a Management Control Package as

the metaphor for the simultaneous existence of a number of Management Control Systems in an organization. These Management Control Systems combined together form the entire control package of the organization. In case a single Management Control System is designed in an integrated matter that entails the entire organization, this thesis also refers to that system as a Management Control Package.

(17)

4 Research method and domain

This thesis is concerned with explaining configuration variety between control systems of organizations. To be able to gain a comprehensive insight into the factors that explain this phenomenon an extensive literature analysis will be performed. Similar methods were chosen in accounting studies performed by authors such as Luft and Shields (2003) and Van Helden (2005). The analysis will be based on management control literature stemming from the journal of ‘Accounting, Organizations and Society’ (AOS) that will be analysed for factors explaining control structure variety of management control systems. The journal ‘Accounting, Organizations and Society’ is chosen for this purpose based on the impact factor of 2.377 for the journal (Elsevier, 2010), which is one of the highest factors for accounting journals available. The level of the impact factor indicates that this journal is considered relevant by and has significant qualitative influence in the accounting field.

An alternative approach of selecting literature from multiple accounting and management journals with a limitation in terms of publication dates (e.g. last 15 years) was considered as well. However, the argument of creating a complete picture of the development of Management Control literature during a larger period of time, which shapes the opportunity to create an holistic insight in configuration factor developments during the last decades of research, led to the choice for this single-journal approach as described before.

Performing an all-fields database query on the ScienceDirect database for the terms “Management+control” in the Accounting, Organizations and Society journal yielded a literature sample of 507 potential relevant articles in the time period 1976 till 2011. The year 2012 was left out of the sample as that year is only due for some limited months at the time of writing of this thesis, providing no option to present a complete insight in the literature of that year. No other constrains were made in the selection of literature, neither in subject nor any other delimiting arguments. The sample of 507 potentially relevant articles was manually screened for relevance for this research, by performing an analysis of the abstract of the article. An article was considered relevant when the management control system could be regarded as the dependent variable in the publication. This implied that as soon as any form of management control system was clearly shaped to a certain degree by the researched or explained phenomena, the article was considered relevant and added to the research sample for further analysis.

An example of a relevant abstract comes from the article of Angelo Ditillo (2004):

(18)

double coordination and knowledge integration role played by management control systems. A field research conducted in three project teams of a software firm supports the relevance of knowledge complexity in explaining the variation of management control systems. The paper concludes with some avenues for future research. (Angelo Ditillo, 2004, p. 1)

An example of an irrelevant abstract, in which the MCS is not the dependent variable, comes from Shultz et al. (2010):

Auditing standards direct auditors to consider business risk and other risk factors when they evaluate the overall risk of material misstatement during the planning phase of an audit. Large audit firms generally use either a strategic-systems approach (SSA) or a transaction-focused approach (TFA) to evaluate misstatement risk. This study used data from a laboratory experiment to examine whether (1) being trained to use either SSA or TFA and (2) analysing information organized in an SSA or TFA format influence the extent to which auditors integrate knowledge of business risk into their judgment about the likelihood of financial misstatement. Only auditors trained to use SSA who analysed information provided in an SSA format effectively integrated business risk assessments with their assessment of the risk of material misstatement.” (Shutz et al., 2012, p. 1)

After screening the initial sample of potentially relevant articles, 122 articles were considered relevant for the research. These 122 publications covered a time period of nearly 35 years of management control research and literature, starting at 1978. For each article in the sample a more extensive analysis was performed, by evaluating primarily the abstract, introduction and conclusion/discussion of the article. In certain situations where no direct consensus on research factors could be established, more in-depth investigation was performed. An example of potentially relevant article is the research made by Chapman and Kihn (2009). This research starts by focussing on just one aspect of information system integration, namely in terms of data architecture, commonly referred to as the single database concept. This seems irrelevant for control research, as the focus is solely on IT system integration, however later in the article they argue that the level of information system integration fosters the four design characteristics that make up an enabling approach to management control. This implied that after having analysed the article it became relevant for the study.

(19)

The research type of (mathematical) analysis was found as not occurring and is therefore left from the table. After having established the research type, the management control typologies relevant for the article were established and attached to the database. Using the management control package typology framework of Malmi and Brown (2008), five main typologies (planning, cybernetic, reward/compensation, administrative and culture) and their 13 ‘sub-typologies’ (long range planning, action planning, budgets, hybrid measurement systems, financial measurement systems, non financial measurement systems, reward/compensation, policies and procedures, governance structure, organisation structure, clans, symbols, values) were used as classification scheme for the analysed literature. The definition and explanation of the typologies used can be found in chapter three of this thesis. In certain situations multiple typologies were attached to the database record of the publication due to multi-perspective approaches of the authors. This led to an increased number of typology occurrences (146) compared to the number of articles studied (122).

Next, the actual factors that influenced the design of the management control system were investigated and reported in the database. Only factors explicitly stated by the authors as influential were considered relevant. Some factors have been combined or merged due to their similar nature and definitions. For example the factors ‘historical background’, ‘historical tendency’ and ‘constitutional tendency’ were merged into the factor ‘historical/constitutional tendency’. This led to a pool of 68 factors influencing the design of management control systems in the literature reviewed. In the appendix a full list of factors can be found.

Table 4.1 - Classification of Papers According to Research Methods, adopted from Shields (1997)

Research method Number of _Papers _PercentagesIn

Theory 33 27% Surveys 29 24% Interviews 6 5% Case/field study 40 32% Archival study 13 11% Simulation/experiments 1 1% Total 122 100%

Table 4.2 - Classification of Papers According to Control Typologies

Typology _OccurrencesNumber of _PercentagesIn

(20)

The amount of factors created a paradox-effect in the research. Further merging the factors to gain a more comprehensive overview would lower explanatory insights of the study and exclude potential interacting effects between factors. However the need for a clear and comprehensive overview remains. To solve for this problem a categorization of the factors was performed. This categorization was made based on two levels of detail.

As this thesis aims to explain control variety in management control packages an integrative approach towards factor categorization had to be performed, which created the first level of categorization of the factors. This approach was based on the Management Control Elements (MCEs) within a package and lead to a model of antecedents for a management control package. This model consists of two sections that can have a potential overlap, and thereby the MCEs within the package can influence each other. The first section is the 'technical and/or economically required' section and consists of the MCEs that follow from theories in the field of economics and management. The second section is the 'psychologically and/or culturally determined' section and consists of MCEs that are used for other reasons, e.g. managerial preferences. Based on this model, a primary distinction between the factors was made between technical and/or economically required or psychologically and/or culturally determined. This approach allows for a global categorization of the factors, but still allows for interaction and overlapping effects between the variables to occur.

The second level of categorization and classification of the factors is built on the work on contingent control variables of Fisher (1998). The factors considered to be ‘Technical and/or economical required’ were categorized towards their contingent nature. The five sub-categories are ‘Uncertainty’, ‘Technology and Interdependence’, ‘Industry, Firm and Unit Variables’, ‘Competitive Strategy and Mission’ and ‘Observability Factors’ (Fisher, 1998). Factors within the ‘psychologically and/or culturally determined’ domain were not sub-categorized as limited relevant research has been performed on the existence and influence of psychological and/or cultural influences on MCS design. Consequently no potential relevant categorization schemes were found in existing literature.

(21)

5 Results of the literature research

This chapter will present and interpret the results found in the literature analysis of 122 articles on management control in the journal of ‘Accounting, Organizations and Society’. Therefore this chapter will answer the third research question of this thesis: ‘Which factors

can be identified in existing accounting literature, that influence the configuration variety of Management Control Systems?’

Firstly the general and overall results on management control antecedents and categories are discussed. Secondly a deeper look is taken into the effects of control factors by examining the results of the classification within the different typologies and categories. Thirdly an analysis is made on a potential effect of influential differences in management control variety between the periods 1978 – 1995 and 1995 – 2011. Afterwards an overview of the direction of influence of the factors on MCS design/adoption is given. Followed by a presentation of the results on simultaneous observations of Management Control System existence, and a list of factors of influence on multiple MCS sub typologies.

5.1 - General and overall results

In the total examination of the 122 articles, and after a merging process as described in the methodology section of this thesis, an amount of 68 factors was considered relevant for the study. These 68 factors had an accumulated total of 283 occurrences in the 122 publications studied. A full list of observed factors and their occurrences can be found in the appendices. Table 5.1 presents and overview of the classification of the antecedents and categories in the study in general.

Table 5.1 – Classification of antecedents and categories of the factors found in the study

Antecedent

Category Occurrences Number of Percentages In

Psychologically and/or culturally determined factors 92 33%

Technically and/or economically required factors 191 67%

Industry, Firm and Unit variables 94 33%

Uncertainty 46 16%

Competitive Strategy and Mission 25 9%

Technology and Interdependence 15 5%

Observability Factors 11 4%

Total 283 100%

(22)

As table 5.1 outlines, the antecedent with highest dominance in factor definition is the technically and/or economically required control element. Furthermore it can be seen that from the technically and/or economically required control elements the majority of the elements is grounded on industry, firm and unit variables. This category explains for 94 (33%) of total factor occurrences in the study. Subsequently the elements related to the Uncertainty category account for 46 (16%) of total factor occurrences in the study. To conclude: Industry, Firm and Unit characteristics combined with Uncertainty influences explain nearly 49% of configuration variety in the management control system research that was investigated.

5.2 - Results per typology

(23)

(24)

Table 5.3 (which was also presented in the methodology section) gives an overview of the number of occurrences of the typologies studied. In the next sections the results per sub typology are briefly discussed including three important factors of influence. A full list of influential factors on each sub-typology can be found in the appendices.

5.3 - Results on Cultural controls

The Cultural typology was among the least discussed typologies in the management control literature studied. Only 7% of the literature addressed any kind of Cultural control typology. Due to this limited attention in research, any potential predictive capabilities of the results on this typology might be weakened. What can be considered interesting is that the technical/economical required antecedent attained the overweight in the Cultural typology, while the psychological/cultural antecedent at first hand could be expected to be of more significant importance. However, as explained in the methodology section, the two antecedents can overlap and interact in our model, therefore making the results not mutually exclusive. Within the technical/economical antecedent the observed variety is dominantly explained by Industry, Firm and Unit variables.

5.3.1 - Clans

For the Clan sub-typology only 2 occurrences were found, with the factors ‘Shared norms/understandings’, ‘Organizational Capabilities’ and ‘Existing knowledge/ skills/ experience’ as three factors influencing the design of Clan systems. The technical/economical antecedent was with 67% of leading influence in the design of Clan control systems in the literature reviewed.

5.3.2 - Symbols

The Symbols sub-typology only had a single occurrence in the research sample, with the factors ‘Socio-political effects’, ‘Historical/constitutional tendency’ and ‘National culture’ as the most important factors on a total of four factors influencing the design of symbolic systems. Within this observed Symbols sub-typology the psychological/cultural antecedent was the single influence in the design of the Symbol control systems in the literature reviewed.

5.3.4 - Values

Table 5.3 - Classification of Papers According to Control Typologies

Typology _OccurrencesNumber of _PercentagesIn

(25)

‘National culture’, ‘Shared norms/understandings’ and ‘Internal consistency’ as the most observed factors out of a total of 8 identified factors within the design of Value systems. The psychological/cultural and technically and/or economically required antecedents exactly shared (50%/50%) the influence in the design of Value control systems.

5.4 - Results on Planning

With 13 observations (8% of the total) the Planning typology is, similar to the Cultural typology, not extensively discussed in the management control system literature studied. Both sub-typologies possess quite an overlap in explaining their configuration variety, as the technically and/or economically required antecedent attained the overweight (74% and 81%) in the factor observations.

5.4.1 - Long range planning

The factors ‘Strategic focus’, ‘Competitive forces’ and ‘Organization size’ were 3 important influential factors out of 12 factors that explained long-range planning in the 5 observations of this control sub-typology. The technically and/or economically required antecedent accounted dominantly for 81% of the variety within these 16 factors.

5.4.2 - Action planning

The 8 observations of Action planning revealed that, similar to Long range planning, in this sub-typology the ‘Strategic focus’ is the most important factor explaining the configuration variety. ‘Decision-making processes’ and ‘Task/Knowledge complexity’ follow this factor in the list of important factors out of a total of 18 identified factors. The technically and/or economically required antecedent therefore explains 74% of variety within the design of Action planning in the literature sample observed.

5.5 - Results on Reward and compensation

Malmi and Brown (2008) did not identify any sub-typology within the Reward and Compensation typology. Therefore this typology is discussed in its entirety. Perhaps contrary to expectations, the literature reviewed on Reward and Compensation systems was largely (68%) explained by technically and/or economically required elements. 17 factors were found to have influence on the configuration, with ‘Environmental uncertainty’ and ‘National culture’ leading the list of these influential factors.

5.6 - Results on Cybernetic controls

(26)

5.6.1 - Budgets

29 factors have been found of influence on the design on Budgets. ‘Organizational culture’, ‘Budget flexibility allowance’ and ‘Manager structure of interest’ are the most important factors. What might be contrary to expectations is that technically and/or economically required antecedent is only slightly dominant in the determination of Budgets (53%).

5.6.2 - Financial measurement systems

The factors ‘Historical/Constitutional tendency’, ‘Institutional environment’ and ‘Environmental uncertainty’ shape the top three of the factor list which consist of 14 factors in total. Similarly to the Budgets sub-typology, what might be contrary to expectations is that technically and/or economically required antecedent is only slightly dominant in the determination of Financial Measurement Systems (55%).

5.6.3 - Non-financial measurement systems

Within the sub-typology of the Non-financial Measurement Systems only three factors have been found of influence: ‘Departmental Interdependence’, ‘Organizational Characteristics’ and ‘Strategic Focus’, leaving for a 100% explanation of variety through the Technically and/or Economically required antecedent.

5.6.4 - Hybrid measurement systems

Hybrid measurement systems form a substantial portion (13%) of the control systems observed in the literature sample. The largest amount of factors in the study has been found within this typology, leading to a list of 30 factors of influence. ‘Electronic/information integration’, ‘National Culture’ and ‘Organization Size’ are forming 3 important factors of influence.

5.7 - Results on Administrative controls

With 62 total occurrences (42%) in the study, the administrative control is the most observed control system in the literature analysis. Also within this typology the technically and/or economically required antecedent is dominating the factors of influence with occurrences of 72, 81 and 81%.

5.7.1 - Governance structure

The Governance Structure is the most observed sub-typology in the study with 33 occurrences (22%). 37 factors have been discovered to have an influential effect on the design of governance structures, with ‘Decision-making processes’, ‘Joint-venture partner relationships’ and ‘Existing knowledge/skills/experience’ as 3 important factors.

5.7.2 - Organisation structure

(27)

5.7.3 - Policies and procedures

The factors ‘Strategic focus’, ‘Competitive forces’ and ‘Degree of owner management’ lead the factor list for this sub-typology which consist of 30 factors. Policies and procedural based control systems have been found 14 times in the literature.

Overall it can be concluded that the technically and/or economically required factors are dominating the control typologies in the literature analysed to the largest degree (68%), with the factors based on Industry, Firm and Unit variables defining this antecedent most dominantly (34%). In the next section of this thesis, an investigation is made whether this dominance shifted within the time spanned by the literature sample.

5.8 - Historical development of factor dominance

Table 5.4 shows an overview of a comparison of the overall results of the study for the periods 1978 till 1995 (period 1) and 1995 till 2011 (period 2). Much contemporary research in MCS theory started to evolve around the mid 1990’s. As described before Otley (1994) for example typed contemporary organisations, which were based on relatively new management concepts such as 'de-layering', 'right-sizing', 'business process re-engineering', 'outsourcing' and 'value chain analysis'. To observe whether similar theories led to a shift in practices in MCS design and configuration the sample is divided in the two periods so potential differences between ‘classical’ and ‘contemporary’ research periods in the literature sample can be observed.

Perhaps contrary to any potential expectations is that the overall results at first hand do not significantly differ between the two time periods taken in investigation. However in appendices two tables with detailed data on the two time periods is depicted, and the results of these tables will be briefly discussed per control typology to show that some interesting changes did occur.

5.8.1 – Developments in cultural controls

For cultural controls no data was available in period 1 on Clan and Symbol control types. For the values control a shift in dominance from psychologically/culturally determined in the

Table 5.4 – Classification of antecedents and categories of the factors found in the study – differences between the periods 1978 – 1995 and 1995 – 2011

Antecedent

Category 1978 – 1995 period 1 1995 – 2011 period 2

Psychologically and/or culturally determined 51 (34%) 41 (31%)

Technically and/or economically required 99 (66%) 92 (69%)

Industry, Firm and Unit variables 48 (32%) 46 (35%)

Uncertainty 28 (18%) 18 (13%)

Competitive Strategy and Mission 11 (7%) 16 (12%)

Technology and Interdependence 9 (6%) 8 (6%)

Observability Factors 3 (2%) 4 (3%)

Total 150 (100%) 133 (100%)

(28)

period 1978 – 1995 to technically/economically required control elements in the period 1995 - 2011.

5.8.2 – Developments in planning controls

For planning controls no significant changes or shifts in dominance in control antecedents took place between the two periods.

5.8.3 – Developments in reward and compensation

In the period 1978 – 1995 the single dominant antecedent was technically/ economically required. In the period 1995 – 2011 this dominance was reduced till 60%, implying psychologically/culturally determined factors gained importance.

5.8.4 – Developments in Cybernetic controls

From the tables it can be observed that for the development of Budgets the balanced psychological/cultural and technical/economically focus (48/52%) shifted from period 1 till period 2 towards a large dominance of the technically/economically required control elements (14/86%). Within the technical/economical antecedent of Budgets the categorical focus shifted from Industry, Firm and Unit Variables to Uncertainty related factors.

For Financial Measurement systems similar effects occurred on the antecedents, but the main technical/economical categories changed here from a focus on Uncertainty in period 1 to Industry, Firm and Unit variables in period 2, which is opposite of the effects on the Budget sub-typology. Within the Non-Financial Measurement systems, no change in dominance of antecedents occurred. Only the Industry, Firm and Unit variables lost their influence on the technical/economical antecedent in period 2. For the Hybrid Measurement Systems no significant changes occurred in the role of the antecedents and composition of the relative weight of the categories within the antecedent.

5.8.5 – Developments in Administrative controls

Within the sub-typologies ‘Governance structure’ and ‘Policies and procedures’ no significant changes occurred in the role of the antecedents and composition of the relative weight of the categories within the antecedent either. Within the Governance structures, the ‘Observability’ factors only acquired a slightly more dominant position. For Policies and Procedures this was the case with the ‘Uncertainty’ related factors, which also acquired a slightly higher dominant position. For the Organisation Structure the effects of antecedents of factors remained almost the same, however within the technical and/or economical antecedent the focus shifted entirely from a combination of Uncertainty and Industry, Firm and Unit variables towards a dominant focus on Industry, Firm and Unit variables.

5.9 – Factor influence

(29)

overview of the observed factors, study numbers in which the factor occurred and their possible negative, non-linear and positive influence on the design and/or adoption of MCSs. The factors are sorted in alphabetical order and displayed on the left column of the grid. The column to the right of the grid presents the study numbers in which the factor was found, which corresponds with the literature overview list in the appendices. Each direction of influence (negative, non-linear, positive) is presented with a bubble, with the size of the bubble representing the total weight of influence of the factor found in the study. In this way not the accumulated influence of the factor is displayed. For example an occurrence of negative, negative, non-linear (neutral) and positive influence would count for a positive bubble with a weight of -1 + 1 + 0 + 1 = 1. But in this table each time that a factor had a direction of influence it is displayed. In case of the previous example this would imply a display of a negative bubble with a weight of 1, a positive bubble with a weight of 2 and a neutral bubble with a weight of 1. This allows for straightforward observation of the existence of simultaneous and rather opposing influences of the factor.

Out of the total of the 283 total factor occurrences, in 131 (46%) of the occurrences a linear relation between the factor and MCS design and/or adoption was found. The absolute majority of these relations (126) were positively linear. The sum of all influence is 121 positively (126 pos. – 5 neg.), which will be used later in this chapter to compare relative influence of factors. A potential explanation for the domination of the positive influence might be that theories generally form about the reasons to adopt a MCS, therefore not investigating the factors that actually lead to decreased use of MCSs.

The factors ‘Strategic focus’ and ‘Electronic/Information integration’ are the factors with the highest positive linear relationship. ‘Electronic/Information integration’ is even in every single occurrence (12) observed as having a positive linear effect on the design and/or adoption of Management Control Systems. ‘Degree of Owner Management’ implied the largest negative influence, with 50% of its occurrence in the literature under analysis to be of negative linear influence on MCS design and/or adoption. However, as explained before, only 5 negative influences were found.

(30)

T ab le 5. 5 – F ac to r i nfl ue nc e o n M C S de sig n a nd /o r a do pt ion

(31)

Table 5.6 – Factor influence on the sub typlogies of Planning, Cultural and Reward/Compensation types of Management Control Systems

(32)

Table 5.7 – Factor influence on the sub typlogies of Cybernetic Management Control Systems

(33)

Table 5.8 – Factor influence on the sub typlogies of Administrative Management Control Systems

(34)

5.10 - Simultaneous observations of MCS factors and typologies

To be able to derive any conclusions on configuration variety of Management Control Systems within Packages it is important to analyse any structural co-existences of MCSs within the literature analysed by this study. Table 5.9 displays the analysis of simultaneous existence of MCS typologies by presenting a matrix with data on simultaneous existence of MCS typologies in the analysed literature. Simultaneous existence of MCSs here means that while analysing a single publication within the literature sample, multiple Management Control Typologies could be identified. In the appendices a matrix can be found with cross-data on every MCS sub typology.

As can be seen in the table, out of the total co-existences the Cybernetic (27), Administrative (27) and Planning (21) type of management control systems exist together to the largest degree. The percentage in the total row displays the relative amount of co-existence compared to the total count of management control systems in the literature (146 total observations, see also table 4.2).

5.11 Factors influencing multiple MCS sub typologies

As in this thesis the focus lies on the role of the factor within packages of control systems it is important to look at those factors that influence more than one element of a control package. Table 5.10 presents a list of factors that are found to have a linear influence on the design and/or adoption of Management Control Systems within more than MCS sub typology. A number equal of higher than two means that the factor can influence multiple elements of the Management Control Package. The factor occurrences that involved a linear influence on the MCS are selected for this table as occurrences with linear influence allow for most systematic explanation of configuration variety of Management Control Systems. This is important because if non-linear influence would be included, the factor list would be less generalizable across packages of systems as the effect of influence can differ between control situations.

Table 5.9 – Structural co-existence of MCS typologies

Culture Planning Reward/ _{Compensation Cybernetic} Administrative

(35)

The first column explains the number of sub typologies the factor has influence on, the second presents the total amount of influence the factor exercised. This number is calculated by accumulating the influence of the factor, making it a sum of both positive and negative influence. The percentage displays the relative amount of influence this factor accounts for compared to the total sum of influence (121) as described in paragraph 5.9. ‘Strategic focus’ for example explains for 12,4% of all linear influence found in this study and was found to influence eight sub typologies.

In the next chapter of this thesis the potential implications of the results, as presented in this chapter, on management control package design will be discussed.

Table 5.10 – Occurrence/influence of factors on multiple MCS sub typologies

Factor # Different sub typologies Accumulated influence _{the factor}

Strategic focus 8 15 (12,4%) Competitive forces 7 14 (11,5%) Electronic/information integration 5 12 (9,9%) Environmental uncertainty 5 7 (5,8%) Organization age/life-cycle 5 5 (4,1%) Task/Knowledge complexity 4 7 (5,8%) (External) consultants 4 4 (3,3%) Environmental complexity 3 6 (4,9%) Existing knowledge/skills/experience 3 4 (3,3%) Organization size 3 4 (3,3%) Departmental interdependence 3 4 (3,3%) Manufacturing flexibility 3 3 (2,5%) Decision-facilitating measures 3 3 (2,5%) Learning 3 3 (2,5%)

Transfer pricing tax compliance 3 3 (2,5%)

Existence of outside investors 2 3 (2,5%)

Decision-making processes 2 2 (1,6%)

Strategic risk 2 2 (1,6%)

Training 2 2 (1,6%)

Note: Number of different typologies represents the amount of simultaneous typologies on which the factor

had a linear influence on design and/or adoption of MCS’s within that typology. The accumulated influence is the sum of all the linear influence of the factor exposed on the MCS’s within the literature research. The percentage explains the relative amount of influence explained by this factor in terms of the sum of total

(36)

6 Analysis of the results and discussion of potential implications

on Management Control Package design

The results presented in this thesis tell something about configuration variety within individual Management Control System theories studied stemming from the journal of Accounting, Organizations and Society. However if the number of theories analysed (122) and the diversity and number of control system observations (146) is taken into account it allows to make some more general statements with regards to Management Control Packages. The last sub research question of this thesis quests how the findings from the literature research potentially explain configuration variety of Management Control Packages. Based on the results presented in the previous chapter, this question will be answered by discussing three coherent perspectives.

6.1 – Simultaneous existence of control typologies

First of all the configuration variety of Management Control Packages can be approached by looking at observed structural co-existence of different control systems. The results show that particular types of systems are found to have a pattern of co-existence together. Therefore these systems might operate regularly together within a management control system package. This is especially the case for Administrative, Cybernetic and Planning systems, which are found to exist in the largest degree together. These particular patterns of co-existence might be due to the similarities in fundamental characteristics of these systems, as they all include mostly formal, tangible and measurable aspects. Perhaps because of this similar nature it makes these systems more obviously visible and also more easily configurable together within the metaphor of the Management Control Package. An important note to make is that this co-existence is based on results of individual studies on MCS level, which implies that any lack of observed co-existence in this thesis does not imply that those particular systems do not successfully co-exist together in practice.

6.2 – Factor antecedents

(37)

literature was explicitly that of a cultural type. This would be consistent with the theory of Malmi and Brown (2008) who mentioned that cultural effects on controls have received less attention in empirical studies. Moreover, even though cultural influences have received less attention by the articles analysed in this study, the psychological/cultural antecedent still appeared to have influence of 33% on MCS design, providing indication that the concept can be considered relevant in the research genre.

6.3 – Control factors

Finally the perspective with the largest potential to explain configuration variety of Management Control Packages is on the actual factors underlying the design and/or adoption of management control systems. Especially those factors are important that are able to explain configuration variety to the largest degrees for multiple control typologies.

As can been observed from this table in appendix 9.1, 13 out of 68 factors account for 143 occurrences in the total literature observations. This is a rather remarkable result of the study and in relative terms this means that 19% of the factors are responsible for 51% of the total management control system variety in the literature observed in this research. This means that configuration variety of the management control systems discussed in the literature, is to large degrees defined by a limited set of factors.

The factors can explain more about the configuration variety within packages as a whole, if their actual influence and the amount of control elements the factor affects is also taken into account. Table 5.10 presents this list of 19 factors, and it becomes clear that ‘Strategic Focus’ has the largest shaping effect in MCS configuration of the systems within the observed literature, as this factor was found of influence in 8 sub typologies and explains 12,4% of total linear influence observed in the literature studied. The factors ‘Competitive forces’, ‘Electronic/information integration’ and ‘Environmental Uncertainty’ follow respectively. Only these four factors together explain for nearly 40% of total linear influence. Therefore it can be said that, within the literature studied, 40% of control systems configuration variety is explained by these four factors. As these four factors are found to have influence on at least 5 out of the 13 elements of the package, this allows for a generalization towards a Management Control Package. However not only these four factors are important as all the factors in this table influenced the design of management control systems within at least two sub typologies. Therefore table 5.10 presents a list of factors that, based on the literature studied, are most likely able to explain configuration variety of Management Control Packages. It has to be noted that here the assumption is made that the factors will linearly influence the configuration of Management Control Systems and thus the Management Control Package equally in different situations, however this assumption is solely based on the data of this research. If other journals or publications would be selected for analysis, different results can occur. More on this is written in the limitations chapter of this thesis.

Management Control Packages