Masterthesis Master of Science in Business Administration Organizational & Management Control

Master of Science in Business Administration

Organizational & Management Control

The structure of the Finance Function within companies

with a multidivsional structure

Amsterdam, August 2009

2

The structure of the Finance Function within companies

with a multidivsional structure

-A field study within large Dutch multinationals-

Master Thesis

Master of Science in Business Administration

Faculty of Economics and Business

Specialisation Organizational & Management Control

Name: H.W. van Zelderen

Student number: 1662694

Telephone: +31(0)644225056

E-mail: haroldvz@hotmail.com

Rijksuniversiteit Groningen

Preface

The following report is my final master thesis for the Master of Science degree in Business Administration at the University of Groningen. After seven years Groningen, I could leave this city with a satisfied feeling, with two diploma’s in my bag. Firstly, I graduated the study business econom-ics on the HanzehogeSchool Groningen. Three years ago, I started with my premaster year of the pro-gram Business Administartion at the University of Groningen. That year, I followed several courses in preparation for the master Organizational & Management Control, including the specialisation course. Hereafter, I enrolled the master. Within the master I followed several courses regarding Management Accounting, Management Control, Financial Accounting and Corporate Governance. These courses helped me to identify relevant information for this master thesis.

Now I can look back at seven beatiful years in Groningen. A time where I learned a lot about Business Administration, but also a time where I learned to took care for my self. Came in as an eigh-teen year old boy, and leaving as a 25 year old man with great memories.

I wrote this master thesis during an internship at PricewaterhouseCoopers Advisory. Within PricewaterhouseCoopers they were interested how companies could structure their financial processes, and why companies choose for this way of organising. This agrees with my interests regarding the structure of the finance function within large companies. Moreover, the subject was in line with courses of my study, for example Financial Accounting, Corporate Governance and the Field Course Organizational and Management Control.

I would like to thank several number of people for their support. Without these people, I would not have been able to succeed. Firstly, I would like to thank PricewaterhouseCoopers for giving me the opportunity to write my thesis in combination with an internship. Furthermore, I would like to thank my supervisiors at PricewaterhouseCoopers, namely Gerben Kraak en Evert van Os, who gave me usefull insights about the subject. I also would like to thank the other colleagues of Pricewaterhou-seCoopers who helped me when I needed them. Furthermore, I would like to express gratitude to the intervieewees for their time, input and feedback. I would also like to thank Sandra Tillema, my super-visor of the university, for her feedback, suggestions and usefull comments. Last but not least, I would thank my girlfriend Rina, my family and friends for their support and interests.

4

Management Summary

Challenged by the idea that less is known about the structure of the finance function within multidivisional companies and the fact that companies struggle how to structure their finance function, this research focuses on the different possibilities to structure the finance function. A field study was chosen to examine how companies structure their finance function and why these companies choose this way of organising. The field study comprises interviews with managers who are active within the finance function of nine large Dutch multinationals.

Table of Contents

Chapter 1.

Introduction to the research

7

1.1 introduction 7

1.2 Objective and Research Questions 7

1.3 Relevance of the research 8

1.4 structure of this thesis 9

Chapter 2. Theorethical framework

10

2.1 Introduction 10 2.2 Management Accounting 12 2.3 Financial Accounting 16 2.4 Treasury 17 2.5 Tax 19 2.6 Concluding remarks 20

Chapter 3. Methodology

21

Chapter 4. Analysis of the case results

26

4.1 The multidivisional structure 26

4.2 Management Accounting 27

4.3 Financial Accounting 33

6

4.5 Tax 40

4.6 Concluding remarks 42

Chapter 5. Conclusion

43

5.1 The finance function within large multinationals 43

5.2 Limitations 43

5.3 Recommendations for further research 43

References

47

Appendix A

Interviewquestions

53

Appendix B

E-mail

56

1. Introduction to the research

Though much is known about the multidivisional structure (e.g. Chandler, 1991) and the struc-ture of some financial processes (e.g. Granlund & Lukka, 1998; Belk, 2002), little is known about the structure of the finance function within multidivisional companies. Neither, there is a proper oversight of the structure of the most important financial processes. Such an overview is important because this will help financial managers to think about the structure of the finance function within their compa-nies.

The role and the position of the finance manager changed a lot during the last century. In the beginning of the twentieth century financial managers primarily raised funds and managed the cash position of the firm. In the 1950s, the present value concept became more accepted. Financial manag-ers were now encouraged to expand their responsibilities and to become concerned with the capital investment projects (van Horne & Wachowich, 2005 pp. 2).

External factors like the raised competition, technological changes, volatility in inflation and interest rates, worldwide economic uncertainty, fluctuating exchange rates, tax law changes, and ethi-cal concerns, resultes in a business partner role for the financial manager, which means that the finan-cial manager is involved in the decision making protocol. Nowadays, there is much more pressure on the finance function. This is partly due to the scandals in the beginning of this century, such as the Enron scandal and the Ahold scandal. The finance function still needs to present reliable figures, but it also has to take care that the company is in control (van Horne & Wachowich, 2005 pp. 2-7).

1.2 Objective and Research Questions

The increasing number of companies with a multidivisional structure and the growing impor-tance of the structure of the finance function, lead to the following objective and research question:

Objective

8

Research Question

How can the finance function of companies with a multidivisional structure be organised, and which arguments do such companies give for their way of organising?

To give an answer to this question, severable questions must be asked. Firstly, it is important to understand what the finance function is, and what processes it comprises. Then we need to find out what possibilities there are to structure the finance function of companies with a multidivisional struc-ture according to the literastruc-ture, and what the literastruc-ture says about their (dis)advantages? The academic literature must provide a framework, which includes the most important financial processes and some important variables for the structure of these processes. This framework will be used to analyse the structure of the financial processes within several companies.

The empirical research consists of a field study within nine large Dutch multinationals, with a multidivisional structure. Employees who are active within the finance function of these companies will be interviewed to find out how their companies structure their finance function, and what their arguments are for choosing this structure. There will also be a comparison between the results from the academic literature and the empirical results to find out what the differences are between these two results, and if recommendations could be made to the literature or organisations in practice.

1.3 Relevance of the research

With this research a contribution will be made to the literature concerning the finance func-tion. An overview will be provided of the structure of the finance funcfunc-tion. Many authors investigated (parts of) the structure of a financial process. For example Granlund & Lukka (1998) investigated the role of employees within the management accounting process and the financial accounting process, and Belk (2002) conducted a research about the structure of the treasury process. Nevertheless, this study will be a contribution because it will go into the structure of four financial processes, and will present findings of four variables within each process. The four variables which will be discussed within this research are:

 Degree of centralisation  Reporting lines

 Hiring and firing procedure

 Scorekeeper role versus the business partner role.

1.4 structure of this thesis

10

2. Theorethical framework

This chapter provides an overview of the relevant theories that exist regarding the problem statement outlined earlier, and provides the theoretical framework, which will be confronted with the practices of large multinationals with a multidivisional structure. The conceptual background is dis-cussed in five sections, which will form the basis of this study. Section 2.1 starts with an introduction on the subject. This section defines the multidivisional structure and the finance function. Section 2.2 until 2.5 describes four processes of the finance function, namely:

 Management accounting  Financial accounting  Treasury

 Tax.

Within these sections, variables will be discussed which are important for the design of these processes. Section 2.6 presents a short conclusion.

2.1 Introduction

Most large companies have a multidivisional structure (Chenhall, 1979, pp. 5; Hill & Pickering, 1986, pp. 26; Hoskisson, Hill & Kim, 1993, pp. 296; Suzuki, 1980, pp. 270). The multidivi-sional structure is seen as the most important organisational innovation of the twentieth century (Wil-liamson, 1985, pp. 279). A multidivisional structure is an organisational structure with several levels. Most authors define three levels within an organisation; business units, divisions, and corporate level (Chandler, 1991, pp. 34). A business unit is an entity with a product / market combination and an own strategy (Hall, 1978, pp. 18). A division is responsible for the day-to-day operating decisions of a sin-gle business and is self-contained. Corporate level is responsible for the long-term strategic decisions within the company and they exercise financial control over the divisions. Concluding, this means that there is a separation of strategic and operational control within the company; however the divisions still make strategic decisions for the business units (Hoskisson, Hill & Kim, 1993, pp. 272; Chandler, 1992).

Within these structures, the companies are organised based on respectively the functions, the areas and the different products which a company produces or delivers.

Williamson (1970) named two other reasons for the exisitence of the multidivisional structure. The first reason is the distinction between strategic and operational decision-making. The structure makes it simpler to observe the profitability of each division, and because of this, corporate level is better able to optimise the resource allocation process. The second reason is that a company is better able to handle multiple product-market combinations using this structure, because business units are in many companies responsible for one (or some) of these product-market combinations.

Besides a substantial number of authors who are positive about the multidivisional structure, there are also authors who are less positive about the multidivisional structure and some of them even criticise on it (Bettis, 1991; Hill, 1985; Hoskisson, Hill & Kim, 1993). They conclude that the transi-tion to a multidivisional structure could lead to worse leadership. Bettis (1991, pp. 317) named the multidivisional structure an ‘organisational fossil’ which becomes irrelevant in the economic world the coming years.

Corporate level is able to structure the organisation, to design the control framework, and to structure the different processes at an optimal way (Chandler, 1991; 1992). There are many processes and functions within an organisation, like sales, production and Human Resource Management. The decision for a particular organisational structure has impact on the structure of the processes. For ex-ample, the degree of centralisation differs. The multidivisional structure was invented to change the degree of centralisation within some processes. Another logical variable which differ in the different organisational structures are the reporting lines within the company. For the purpose of this research, the focus will be on the finance function of companies with a multidivisional structure.

Finance function

The finance function consists of the people, technology, processes and policies that dictate tasks and decisions related to financial resources of a company (Mateevich, 2007, pp.1). The classical functions of the finance function are divided into three parts; the factory, the powerhouse and the counting house (Sheridan, 1997, pp. 7). The factory consists of the books of account. This contains the ledgers, the consolidation, and the statutory reporting. This is the financial accounting process of the company. The powerhouse is the management accounting process of the company. This contains budgets, management information and costing. The last function is the counting house, which includes cash management, tax, and bank relations. These are the treasury process and the tax process (Sheri-dan, 1997, pp. 7).

control-12 ler was mostly the leader of this function (Desai, 2008, pp. 112). Nowadays, the CFO is in many com-panies the right hand of the Chief Executive Officer (CEO) and responsible for the control, the costs and the performance of the company (Atkinson et al., 1987, pp. 3; The Economist, 2008, pp. 5).

As mentioned before, four processes of the finance function will be discussed, namely man-agement accounting, financial accounting, treasury and tax. Companies could have more processes within the finance function (for example; pension administration, or salary records), but these four are the most important processes (Sheridan, 1997, pp. 7). The first process which will be discusses is the management accounting process.

2.2 Management Accounting

There are many definitions of accounting. For this study the definition of the American Ac-counting Association (1966) is used:

“The process of identifying, measuring and communicating economic information to permit informed judgements and decisions by users of the information”

The objective of accounting is to provide sufficient economic information to people who have interest in a company. This can be managers, (potential) investors, the government, employees and creditors. The people with interest in the company can be divided into two groups, internal parties (employees, managers) and external parties (government, creditors, investors). Management account-ing provides information for the internal parties, while financial accountaccount-ing is concerned with the pro-vision of information to external parties (Atkinson et al., 2006, pp. 3). This section describes the man-agement accounting process, while section 2.3 discusses the financial accounting process.

The institute of management accountants defines management accounting as a continuous im-provement process for the planning, design, testing and employment of financial and non-financial information systems, which creates value for the company. Management accounting information makes it is also possible to measure the performance of the business units and is the main source for companies to make decisions (Atkinson et al., 2006, pp. 3).

Variables which are discussed in this section are the degree of centralisation within a com-pany, the reporting lines, the hiring and firing procedure and the business partner role versus the scorekeeper role.

Degree of centralisation

When a company decentralises the management accounting process, they can better react to local changes in the market. It also enables corporate level to focus more on the long term strategic decisions of the management accounting process (Dooms, 2005, pp. 2).

However, decentralisation also implies a loss of control for corporate level and increases the agency costs within a company (Christie, et al., 2003, pp. 4; Maas & Matejka, 2007, pp. 15). Agency costs are costs taht are made because the goals of the business unit manager and the person who is responsible for the management accounting process (mostly a business unit controller), are not always congruent with the goals of the company (Mallin, 2007, pp. 12). It is more likely that agency costs increase, when business units have more authorities within the management accounting process.

Hopper (1980, pp. 404 - 405) discussed three forms of centralisation within the accounting process. A decentralised structure, a centralised structure, and a combination between the decentral-ised and the centraldecentral-ised structure. Within the fully decentraldecentral-ised structure, the business unit controller shares the office with the business unit manager, and the business unit manager is also the boss of the business unit controller. Within the combined structure, the business unit controllers share the office with the business unit manager, but the CFO / corporate controller is the boss of the business unit con-troller. The primary reporting line is also to the CFO / corporate concon-troller. Within the fully centralised structure share the business unit controllers one office and the CFO / corporate controller is the boss of the management accountant.

Recently there is a trend to decentralise the management accounting process. The most tant reason for this is that the business unit controllers, who acts closely to the business, have impor-tant information about the business which is not available at corporate level. This could be specific country information, which is not known by corporate level which is located in another country (Burns & Baldvinsdottir, 2007 pp. 128; Granlund & Lukka, 1998, pp. 194).

Reporting lines

Another variable when structuring the finance function are the reporting lines within a com-pany. Within the management accounting process, the reporting lines of the business unit controller will be discussed.

Two different reporting lines are mentioned within the literature regarding the management accounting process, namely the dotted reporting line and the solid reporting line. When the primary reporting line is between the business unit controller and the functional manager, than it is called a solid line. When the primary line is between the business unit controller and the business unit man-ager, it is called the dotted line (Merchant & Van der Stede, 2007 pp. 633-634).

14 function at corporate level or at the division, for example a corporate controller or division CFO). The division controller needs to report fairly and objectively on the economic situation of their business unit. On the other hand, they have the local responsibility to the business unit manager to provide stra-tegic and operational decision-making information (Maas & Matejka, 2007, pp. 3; Merchant & Van der Stede, 2007, pp. 633; Ten Rouwelaar, 2006, pp.12).

The dual allegiance of the business unit controller corresponds to the two possibilities of re-porting. Figure 2.1 presents the two possible reporting lines (solid and dotted lines). Within both mod-els, two lines can be seen. The solid line represents the primary reporting line. The dotted line repre-sents the secondary reporting line (Anthony & Govindarajan, 2001 pp. 73-74; Merchant & Van der Stede, 2007 pp. 633-634). Sathe conducted a survey about this subject. He concluded that 15% uses the solid line reporting. Because this research was executed in 1982, it is not known if this is still the main approach.

Dotted line Solid line

Figure 2.1: possible reporting lines

Hiring and Firing procedure

Some authors find the hiring and firing procedure one of the key indicators of the degree of centralisation (Chia, 1995, pp. 817; Chow, Shields and Wu, 1999, pp. 458). However, this research will discuss these variables separately to find out what the relation contains.

Within the management accounting process, it is interesting to see who is responsible to hire and fire the business unit controller. The literature mentions two persons who can hire the business unit controller, the hierarchical manager (business unit manager), or the functional manager (division controller). The person who is the primary receiver of the reports is in most cases also the ‘boss’ of the controller (Anthony & Govindarajan, 2001 pp. 73-74). Because this variable depends on the degree of centralisation and the reporting lines, it is expected that the hierarchical manager is responsible to hire the business unit controller within most companies, because there is a trend to decentralise this proc-ess, with the primary reporting line between the business unit controller and the business unit man-ager. When the business unit manager hires the business unit controller, and he is also the primary receiver of the management report of the business unit controller, than there is a loose of control for corporate level.

When companies centralise their management accounting process, the division controller is be responsible to hire the business unit manager (Anthony & Govindarajan, 2001 pp. 73-74). When the division controller is responsible to hire the business unit controller, and corporate level increases the emphasis on the functional responsibility, the business unit controller could be seen as a ‘corporate watchdog’ (Maas & Matejka, 2007, pp. 3).

Scorekeeper role versus business partner role

The role of the business unit controllers can be a scorekeeper or a business partner. A score-keeper is someone who writes accurate financial reports, based on the past (Granlund & Lukka, 1997, 213). A business partner is an active forward thinking business analyst who is involved in the deci-sion-making process and provides strategic information for the management (Burns & Baldvinsdottir, 2005, pp. 725; Desai, 2008; Granlund & Lukka, 1997, pp. 213; 1998, pp. 185; Järvenpää, 2007, pp. 99; Kaplan, 1995, pp. 7; Lukka, 1998, pp. 333; Maas & Matejka, 2007, pp. 2;).

16 Because of this, a company must react quickly on demands of the clients, and a business unit control-ler with a focus on the business can react faster on the demands of the clients, and the trends in the local market, than someone from corporate level.

Concluding, two different structures can be seen within the management accounting process, according to the literature. The first structure is the decentralised structure. The business unit control-ler within this structure, reports to the business unit manager, and is hired by the business unit manag-er, with a business partner role for the business unit controller. The second structure is a more centra-lised structure. The business unit controller reports to the division controller within this structure and is hired by the division controller, with a scorekeeper role for the business unit controller. The first structure is the most familiar one, according the literature.

2.3 Financial Accounting

As mentioned before, the financial accounting process is the second part of the accounting process. The most important output of the financial accounting process is the statutory financial ac-counts and the annual budgets. The financial accounting process is mostly outward facing, which means that the financial department has much contact with shareholders and auditors (Sheridan, 1997, pp. 12). The literature does not provide much information about the structure of the financial account-ing process, but discusses the accountaccount-ing process in general or only the management accountaccount-ing proc-ess. However, some information about the variables concerning the management accounting process can be used within this process too, because the structure of these processes has many similarities. The management accounting process and financial accounting process could even be integrated processes (Granlund & Lukka, 1997, pp. 232).

Degree of centralisation

Reporting lines

Just as in the management accounting process, a relation can be seen between the reporting lines and the degree of centralisation (Hopper, 1980, pp. 404-405). The possibilities for the reporting lines within the financial accounting process are the same as within the management accounting proc-ess. These are the solid line method, and the dotted line method. Just like within the management ac-counting process, there is also a dual allegiance within the financial acac-counting process. The financial accountant has a functional responsibility to the division controller, and a hierarchical responsibility to the business unit manager. However, because there is a trend to centralise the financial accounting process, and the figures are made fore the external reports which are made by corporate (Granlund & Lukka, 1998, pp. 194), it seems logical that most companies have solid reporting lines within the fi-nancial accounting process.

Hiring and Firing procedure

The literature regarding the management accounting process named two functions who can hire a business unit controller. These are the hierarchical manager (business unit manager), and the functional manager (division controller) (see for example Merchant & van der Stede, 2007). These are probably the two functions who can hire the person who is responsible for the financial accounting process. The business unit manager was named as the most obvious manager to hire the business unit controller within the management accounting process. This is because the person, who receives the monthly report, is also the person who hires the business unit controller (Anthony & Govindarajan, 2001). Because the financial accounting process is a more centralised process (Burns & Baldvinsdot-tir, 2007, pp. 124), where the person who is responsible for this process reports mostly to the division controller, it could be suspected that the division controller is most likely the person who hires the financial accountant.

Scorekeeper role versus Business partner

As mentioned within section 2.2, managers within the accounting process can be a scokeeper or a business partner. However, within the financial accounting process the person who is re-sponsible for this process within the business unit will normally be a scorekeeper. This is because the aim of this process is to deliver accurate figures for the annual report (Granlund & Lukka, 1998, pp. 199).

18 business unit manager, and is hired by the business unit manager. Within both structures, there is a scorekeeper role for the financial accountants.

2.4 Treasury

Many authors have given a definition of the treasury process. Judged by some definitions, the treasury process is responsible for cash management, efficient use of capital, financial risk manage-ment, borrowing cash and the dividend policy (Brealey & Myers, 2000, pp. 9).

The present treasury process is relatively new, compare to the other processes. The treasury process became globally important since the 1980s (Sweeney, 1997, pp. 22). The growing importance of the treasury process is first of all because of the globalisation of economies, since corporations with an international focus run more risks, like for example currency risk. Also, the financial deregulation was important for the growing importance in the treasury process. Important deregulations were the interest rate deregulation in 1981 and the liberalisation of capital flows in 1980 (Collier, Cooke & Glinn, 1988, pp. 190; Gual, 1999, pp. 372).

Since no specific information could be found regarding the remaining variables, this section only provides information about the degree of centralisation. A judgement about the structure of the other variables will be made after the empirical research.

Degree of centralisation

Sweeney (1997, pp. 22 - 23) describes four possibilities in the degree of centralisation of the treasury process. The detailed knowledge about the financial resources of the company could be com-pletely centralised at corporate level. The second possibility is that the operational activities are exe-cuted at the business units, but with centralised decision authorities. The third possibility is to contract out the treasury function. The fourth possible structure for the treasury process is that all the business units have their own treasury departments where all the decision-authorities are decentralised. The last structure is used in multinational companies and conglomerates.

The introduction of the euro was important for the structure of the treasury process of multina-tionals. Before the introduction of the euro, the treasury process was simply organised. Every country with a substantial turnover had a treasury department (Soenen & Aggarwal, 1989). After the intro-duction of the euro, more and more companies centralise there treasury process, because the euro re-duces the financial market imperfections and therefore it is more efficient to centralise this process (Von Eije & Westerman, 2002, pp. 15). This suggests that there is a causal relationship between the number of different currencies and the degree of centralisation of the treasury process.

der Stede, 2007). Many advantages are named for a centralised treasury process. Firstly, companies with a centralised treasury process are, for example, more able to gain economies of scale by manag-ing the various currencies. Secondly, there is more control for corporate level because they have a better view of the liquid resources of the company. Thirdly, the number of employees within the treas-ury process can be reduced (Brown, 1997; Collier & Davies, 1985, pp. 329; Von Eije & Westerman, 2002; Miles, 1997).

Partly centralising the treasury process often results in a motivation and initiative problem for local treasury management to control the cash flows adequately, because they are only allowed to exe-cute the policy of corporate level (Von Eije & Westerman, 2002, pp. 16; Merchant & van der Stede, 2007).

Belk (2002, pp. 50) conducted an in depth study within 65 multinational companies from the United States of America, the United Kingdom or Germany. He concludes that 83% of the companies within his sample have a centralised treasury process.

2.5 Tax

The tax process is defined as the process which deals with tax planning, tax management and compliance (Sheridan, 1997, pp. 13). The tax process is important with raising finance, especially the choice between issuing debt rather than equity capital (Pike & Neale, 1993, pp. 615; Horne & Wa-chowizc, 2005). Some types of fixed assets can attract a tax relief. “These are tax incentives to stimu-late certain types of investment such as in industrial plant and machinery.” The tax experts could give an advice here, because they are familiar with these rules (Pike and Neale, 1993).

Also the structure of the tax function differs between companies, however not much informa-tion can be found regarding the tax process, compared to the accounting processes. Because of this, only the degree of centralisation are discussed in this section.

Degree of centralisation

20 A fundamental objective of the tax process is to collect revenues with lowest possible costs. The primary advantage of a centralised tax process is that tax managers are able to operate more effi-cient, because of economies of scale and greater specialisation. However a decentralised tax process leads to more local knowledge about the local rules. Companies must make a choice between a more efficient process and a process with more local knowledge about the foreign tax laws (Martinez-Vazquez & Timofeev, 2005, pp.5).

2.6 Concluding remarks

The structure of the finance function in a multidivisional company comprises many variables. This chapter defined four variables within four different processes. However, no information could be found regarding the reporting lines, the hiring and firing policy, and the scorekeeper versus business partner role within the treasury process and the tax process.

It is not known if the four variables are the most important variables. Other processes could be important as well. What we do know is that these variables are an issue at most companies. There are many authors who have written about this subject. However they were mainly concentrating on the management accountant process. Moreover, they not discuss all the variables very extensively. It can be seen that the degree of centralisation is the most discussed variable within the literature. Table 2.1 shows a summary of this literature review. It showed what which alternatives are mentioned, and which are the most preferable in the different processes.

The empirical research can give more answers about the four variables and will test the pre-liminary answers given in this chapter.

Degree of centralisation Reporting lines Hiring and firing policy

Scorekeeper role versus business partner role Management account-ing Decentralised, Combination Dotted lines, Solid lines Hierarchical man-ager, Functional manager Business partner, Scorekeeper

Financial accounting Centralised, Combination Solid lines, Dotted lines Functional manager, Hierarchical manager Scorekeeper Treasury Centralised, Combination, Decentralised, Contract out Tax Centralised, Decentralised

3. Methodology

The focus of this chapter is to describe how this research has been conducted. Section 3.1 de-scribes the research design. Section 3.2 dede-scribes which companies have been selected for this re-search, and the criteria for the selection. Section 3.3 presents more information about how the data has been collected. Hereafter, section 3.4 describes how the data has been analysed. The last section dis-cusses three quality indicators; namely, internal validity, external validity and reliability.

3.1 Research design

The field of research regarding the finance function at multidivisional companies appears to be relatively undiscovered. Various studies have explained the processes within the finance function and many authors have evaluated the multidivisional structure. However, to be the best of our knowledge, there has been no study about the design of the finance function at multidivisional companies, or about the differences of the finance function between the financial service companies and the commercial sector.

For this reason, a field study has been conducted. A field study is closely related to a case study and a multiple case study. A case study is: “an empirical inquiry that investigates a contempo-rary phenomenon within its real-life context when the boundaries between phenomena are not clearly evident and in which multiple sources of evidence are used” (Yin, 2003, pp. 13). The difference of a case study and a field study is that a field study comprises multiple used which are carefully chosen. The difference of a field study and a multiple case study is that the cases are carefully chosen within field studies. This is in contrast with a multiple case study, where the cases are randomly chosen (Meredith, 1998, pp. 443).

Advantages of field studies are that (1) the research can be conducted in its natural setting, (2) that the important why question can be raised instead of just the how and what questions and a last advantage is that (3) a case / field study is better able to explore variables which are still unknown (Benbasat, Goldstein & Meat, 1987, pp. 370).

However, there are also disadvantages of field research. For example the costs and the access to the data could be a problem. Another disadvantage is the generalisation of the data (Meredith, 1998, pp. 443). The generalisation of the data will be discussed in section 3.5.

22 explain causal relationships. Because this research will describe the structure of the finance function within some companies, this is a descriptive research.

3.2 Case selection

In this research, nine1 finance functions at large Dutch multinationals with a multidivisional structure, have been investigated. The case studies were selected carefully to ensure that the compa-nies met the criteria2. Three criteria were essential in choosing cases for this thesis. First it had to be large companies. The companies all have a high amount of turnover (between 747 million and 26,345 million) and many employees (between 11,714 and 126,459 ). Six of the nine companies are quoted on the AEX, while two companies are quoted on the AMX. The other company is a cooperation. The second criterion is that the companies must have a multidivisional structure. The third criterion is that the main office must be located within the Netherlands, so the cultural differences are reduced as much as possible. The following companies are selected as representative case studies, because they meet the three requirements:

 AEGON  ASMI  Company X  Company Y  DSM  ING  Philips  Rabobank  Wolters Kluwer

Company X and Y wanted to cooperate with this study, but they requested to stay anonymous.

3.3 Data collection

Information has been gathered by interviewing participants who are working within the fi-nance function, and have the oversight of the different processes of the fifi-nance function. Most partici-pants were located at corporate level. However, some participartici-pants who have an important function at

1

Eisenhardt (1989, pp. 545) mentioned that a proper sample size for a field study lies between four and ten cases.

2

the division / business units were asked to shed their light on the propositions made in the literature review. Semi structured open-ended interviews were conducted to elicit information about important processes of the finance function within companies with a multidivisional structure. Using this method helped to create a better understanding about the structure of the finance function.

The aim of the interviews was to find out how large companies with a multidivisional struc-ture organised the different financial processes. The interviews were strucstruc-tured per variable. The ques-tionnaire used during the interviews can be found in Appendix A.

Employees we interviewed were contacted by e-mail. The purpose of this e-mail was to in-form the employees about the contents and the purpose of the research and to ask for their cooperation. We assured them that the provided information is only be used for this study and will be handled con-fidentially. The email which was sent to the interviewees can be found in Appendix B.

Every interview lasted between 1 and 1.5 hour which was enough time to handle every ques-tion of the quesques-tionnaire. Three persons were present during the interviews; the interviewee, the inter-viewer and a co-interinter-viewer to help the author when necessary. Each interview was tape-recorded. The interviews were summarised by the interviewee, discussed with the co-interviewer, and sent back to the interviewee for comments and approval. After receiving approval from the interviewee, the data was analysed.

3.4 Data Analysis

The data is analysed in chapter 4. No specific information of a particular company is men-tioned because the data is confidential. The objective of chapter 4 (data analysis) and chapter 5 (con-clusion) is to show how companies structure their finance function how companies deal with the dif-ferent possibilities of the variables and the impact of these possibilities on the process. Another objec-tive is to compare these results with the findings from the academic literature.

The data are ordered by the four defined financial processes, because within this rank, it is eas-ier to identify similarities between the processes. However, the first section is about the multidivi-sional structure. The processes are subdivided per variable. Each section starts with a table with the most important outcomes of the interviews and the literature review. The order of the companies is randomly chosen. The objective of the data analysis is to find out what the similarities and differences are between the findings from the empirical research and the literature review, and to discuss why companies choose a particalur way of organising.

24 identify the degree of centralisation and the scorekeeper role versus the business partner role. The tables of chapter 4, shows that these two variables are valued with a number of one till five, giving an indication of the degree of centralisation and the scorekeeper role versus business partner role. These figures are assigned by the author of this thesis, on the basis of the interviews.

The data are analysed by using the principles of cross-case analysis (Eisenhardt, 1989, pp. 540). Similarities, deviations, and reasons for particular choices are discussed. Company A is named twice in the analysis because it uses various structures within the divisions for the financial processes. Because of this, ten companies are discussed, while these are actually nine different companies.

3.5 Quality Indicators

Internal validity, external validity and reliability are important quality indicators (Miles & Huberman, 1994, pp. 277 - 279). Internal validity refers to the degree in which the results of an em-pirical investigation could be interpreted adequately and tells us if we can trust those interpretations. External validity refers to the degree in which the results of an empirical research can be generalised to a wider population. A reliable research means that the same results are achieved when other re-searchers repeat the same procedures used for this research (Riege, 2003, pp. 80 – 81).

A few actions were undertaken to achieve a high internal validity. Firstly the results of this research were discussed with key informants and experts. Another solution to increase the internal validity is to tape record the interviews and to give the interviewees the possibility to react on the in-terview report (Riege, 2003, pp. 81; Maso & Smaling, pp. 71 - 72). For this research, we inin-terviewed only one person in each company. In order to improve the internal validity, it would be better when multiple persons in each of the selected companies had been interviewed to obtain a better view about the content. Because of the little amount of time of both the researcher and the companies just one person per company was interviewed.

4 describes the comparisons and differences between the literature and the evidence found during the interviews.

26

4. Analysis of the case results

In this chapter, the results of the case study research will be presented and a comparison will be made with the academic literature. The chapter illustrates how companies have structured their finance function, and how this relates to the findings from the literature review. The first section pre-sents a short summary of the multidivisional structure within the companies. Sections 4.2 until 4.5 present the analysis of the financial processes per variable. Relevant literature is used to provide a framework for the structure of the financial processes, and an empirical research was executed to test this framework within ten Dutch multinationals with a multidivisional structure. The results of the empirical research provide insight in the reasons why companies have chosen for their current struc-ture of the financial prosesses, and what the (dis)advantages are of the strucstruc-ture.

Every section shows tables which summarise the structure of the variables within the different processes. The results are based on the interviews with employees who are active in the finance func-tion of large Dutch multinafunc-tionals, and the academic literature. Appendix C shows the analysis per company.

4.1 The multidivisional structure

All companies that participated in this research were selected because they have a multidivi-sional structure. However, there are many differences between the structure of these companies. These are for example (1) the way companies make a distinction between divisions, (2) the name of the func-tion of the managers within the finance funcfunc-tion, and (3) the number of hierarchical levels in the struc-ture.

Six companies have divisions based on product lines, while the other four companies have di-visions with a geographical focus. It seems that this difference does not influence the structure of the finance function.

The name of the functions of the managers differ between the companies. For example, some companies have business unit managers, while other companies named this person a business unit CEO. Another example is that some companies have a business unit controller and a business unit CFO, while other companies only have a business unit controller.

Figure 4.1: a standard structure of (a part) of the finance function within a multidivisional company, based on the interviews.

4.2 Management Accounting

The management accounting process is the process with most differences between the compa-nies. Because of this, this process was the main issue in every interview. Many relations could be seen, between the variables. The first variable which will be discussed is the degree of centralisation.

Corporate CEO

Corporate CFO

Division CEO Division CEO Division CEO

28

Degree of centralisation

Within the degree of centralisation, the decision authorities of the business unit controller were examined. Table 4.1 shows the degree of centralisation, and the reporting lines within the management accounting process, regarding the empirical results and the findings from the academic literature.

Company Degree of

centrali-sation

Reporting lines

A1 2 Solid line

E 2 Solid line

G 2 Both solid and dotted

lines

C 4 Dotted lines

D 4 Dotted lines

F 4 Both solid and dotted

lines

H 4 Both solid and dotted

lines A2 5 Dotted line B 5 Dotted line I 5 Dotted line Literature Decentralised (mf) or combination Dotted lines (mf) or solid lines

Table 4.1: a summary of the structure of the degree of centralisation and the reporting lines within the management accounting process

The literature mentioned two degrees of centralisation for the management accounting proc-ess. These were the decentralised structure (most favourable) and the combination of the centralised and the decentralised structure.

The empirical research show the same possibilities to structure the management accounting process. The business unit controllers within companies with a decentralised management accounting process have many decision-authorities regarding among others price setting, investments and system of reporting. The business unit controller is the right-arm of the business unit manager. Together they are allowed to decide some KPIs of their own, and have the freedom to decide how they want to attain the KPIs from corporate level. As can be seen in table 4.1, seven companies have decentralised their management accounting process.

Legend

Company A1, E and G have a combination between the centralised and the decentralised structure. Company E and G are two of the smallest companies in this sample. These companies have a centralised the decision authorities within the management accounting process, because they believe that corporate level are able to have a clear overview of the market. Corporate level within company A1 has also the possibility to operate closely to the market, because corporate level (company A) is located next to company A1.

The results of the field study correspond with the academic literature about this subject. None of the Dutch multinationals that participated in this research have a fully centralised management ac-counting process. Main reason is that the management acac-counting process is very important for the decision making process in these companies. However, not all the companies decentralise the man-agement accounting process, while this was expected by the literature. It can be seen that smaller companies centralise the decision rights. This is because smaller companies are better able to have a view over all the markets they are active. This is not possible for larger companies.

Reporting Lines

Table 4.1 shows the reporting lines within the ten interviewed companies in combination with the results from the academic literature. The literature mentioned two possible reporting lines for the business unit controllers. These are the dotted reporting line and the solid reporting line. Furthermore, the literature mentioned that the dotted reporting line is the most favourable reporting line.

Table 4.1 shows three different possibilities for the reporting lines within the interviewed companies. These are the solid reporting lines, the dotted reporting lines, or both solid and dotted re-porting lines. The three companies with both dotted and solid lines, have an integrated management accounting and financial accounting system, where both the hierarchical manager and the functional manager needs the information included in the management report.

Furthermore, there seems to be a relation between the reporting lines and the degree of cen-tralisation, according to the interviewees. Both companies with solid reporting lines (A1, E) have a centralised management accounting process. As mentioned in the last sub-section, these are companies where corporate level have has decision rights within the management accounting process. There are solid lines because corporate needs information to make decisions, according to the interviewees.

30 One company recently centralised the management accounting process. In the old situation, the business unit controller was authorised to determine some KPIs, and he had the authority to decide how the business unit should attain those KPIs. The business unit controller reported to the business unit manager. Nowadays, the business unit controller has fewer authorities regarding the KPIs, and the business unit controller reports to the division controller. The corporate controller of this company mentioned that the centralisation of the management accounting process, affected the relation between the business unit managers and the business unit controllers:

“In the past, the business unit manager and the business unit controller worked like a tandem. Nowa-days, the business unit manager miss their economic right-hand”

Company A illustrates perfectly the relation between the degree of centralisation and the re-porting lines, as was mentioned within the literature. The structure of the management accounting process differs between the divisions. The structure of the management accounting process within A1 is centralised, while A2 has a decentralised structure. Because corporate level carries out many finan-cial activities for A1, they need the information to make operational decisions. For this reason, there is a solid line relationship between the business unit controller and the division controller.

The other divisions have a decentralised management accounting process. The business units within these divisions have more authorities, so the business unit controller reports to the business unit manager. This is because the business unit managers need the information to make decisions, accord-ing to the interviewee.

When we compare the empirical results with the findings from the literature, these partly cor-respond with each other. A relation can be seen in the literature between the degree of centralisation and the reporting lines in large multinationals with a multidivisional structure. As expected by the literature, most companies have a dotted reporting line. However, three companies have both dotted and solid lines. The existing literature assumed that there are solid or dotted reporting lines, however even large multinationals integrated the management accounting and financial accounting process what means that there are both dotted and solid reporting lines. Two companies completely deviate from the findings of the literature. Company A1 and E have solid reporting lines, because they believe that the division controller must receive the management report, because the division controller in these companies have the decision authorities within this process.

Hiring and Firing procedure

The focus in the interviews was on the person who is responsible for hiring and firing the business unit controller. Table 4.2 presents a summary of the findings from the interview.

Hiring and firing proce-dure

A2 Hierarchical responsibility 5 Dotted line

B Hierarchical responsibility 5 Dotted line

I Hierarchical responsibility 5 Dotted line

C Hierarchical responsibility 4 Dotted line

D Hierarchical responsibility 4 Dotted line

G Hierarchical responsibility 2 Both solid and dotted lines

A1 Functional responsibility 2 Solid line

E Functional responsibility 2 Solid line

F Central management development

4 Both solid and dotted lines

H Central management development

4 Both solid and dotted lines

Literature Hierarchical manager (mf)

or functional manager

Decentralised (mf) or combination

Dotted lines (mf) or solid lines

Table 4.2: a summary of the structure of the hiring and firing within the management accounting process

The literature mentioned two possible managers who can hire the business unit controller. The functional manager (division or corporate controller) and the hierarchical manager (business unit man-ager). Furthermore, the literature mentioned a relation between the degree of centralisation, the report-ing lines and the hirreport-ing and firreport-ing procedure. The hierarchical manager is seen as the most favourable manager to hire the business unit controller because he has to cooperate with this person, and because the business unit controller reports to the business unit manager, according the literature.

32 manager is responsible to hire the business unit controller. Company G deviate from this relation. They have a centralised management accounting process and corporate level have all the decision rights regarding the settlement and the execution of the realisation of the KPIs, the setting of the report and the data of reporting. However, the business unit manager is responsible to hire the business unit controller.

Company F and H have a central management development team. This team, which is located at corporate have the responisbility for the hiring and firing procedure of all the relevant employees within the finance function. The first job interviews with a candidate occurs with the central manage-ment developmanage-ment team.When this team decides that the candidate is capable enough, there will be a final job interview between the candidate and the business unit manager. The candidate is hired when the business unit manager and corporate agrees. The central management development team is also responsible for the internal transfers of the relevant employees within the finance function. Both inter-viewees mentioned within the companies with a central management development team agrees that this program works effectively.

The empirical findings deviate from the academic literature. Firstly, some companies have a central management development team, which was not mentioned within the literature. Secondly, the relation between the reporting lines and the hiring and firing procedure is important within companies. This relation can also be seen within the literature. However, some companies have solid reporting lines, which means that the business unit controller is hired by the division controller.

Scorekeeper versus business partner role

During the interviews the role of the business unit controller is discussed. This variable is re-lated to the degree of centralisation and the reporting lines. Table 4.3 presents the findings of the in-terviews in combination with the academic literature.

Scorekeeper role versus business partner role

Degree of centralisation Reporting lines

D 5 4 Dotted line

I 5 5 Dotted line

A2 4 5 Dotted line

C 4 4 Dotted line

F 4 4 Both solid and dotted lines

H 3 4 Both solid and dotted lines

A1 2 2 Solid line

G 1 2 Both solid and dotted lines

B NA 5 Dotted line

Literature Business partner (mf) or

Scorekeeper

Decentralised (mf) or combination

Dotted lines (mf) or solid lines

Table 4.3. The scorekeeper role versus the business partner role within the management accounting process

The literature mentioned that there is a trend that companies design a business partner role for their business unit controllers, because the business unit controller is better able to react on changes in the local market.

As shown in table 4.3, not all the companies agree with the academic literature. Three compa-nies with a centralised management accounting process and solid reporting lines (or both solid and dotted lines) created a scorekeeper role for their business unit controller (company A1, E and G).

Six companies have a business partner role for their business unit manager. This finds expres-sion in the relation between the controller and the business unit manager. Within these companies, the business unit controller is the right-hand of the business unit manager. The controller is a proactive employee with a focus on the business. The interviewees mentioned that the business unit controller must act closely to the business, so he is able to observe certain trends in the market. However, this structure causes a loss of control for corporate level, according to the interviewees. For company B, it was not possible to determine the role of the controller. This is because the business units are allowed to decide the role of the business unit controller themselves.

It can be seen that if a company wants to create a business partner role for their business unit controller, they must have a decentralised management accounting where the business unit controller is responsible to the business unit manager. Three companies deviate from this structure. These com-panies find the efficiency of a centralised managent accounting process more important than a busi-ness unit controller who can add value for the company.

4.3 Financial Accounting

There are fewer differences in the structure of the financial accounting process compared to the management accounting process of the interviewed companies. However, there are some interest-ing differences and motivations highlighted below.

Legend:

34

Degree of centralisation

Within the degree of centralisation of the financial accounting process, the decision authorities of the financial accountant were examined. Table 4.4 shows the degree of centralisation and the re-porting lines (which will be discussed hereafter) of the financial accounting process.

Company Degree of

centralisa-tion Reporting lines A1 1 Solid line C 1 Solid line E 1 Solid line D 2 Solid line

F 2 Both solid and dotted

lines

G 2 Both solid and dotted

lines

H 2 Both solid and dotted

lines Solid line

B 3 Solid line A2 4 Dotted line I 4 Dotted line Literature Centralised (mf) or combination Solid lines (mf) or dotted lines

Table 4.4. The degree of centralisation and the reporting lines within the financial accounting process

The literature mentioned two possible degrees of centralisation for the financial accounting processes. These are the centralised structure, and the combination between the centralised and decen-tralised structure, whereby the cendecen-tralised structure is used mostly.

As shown in table 4.4, three companies have centralised all the administrative tasks for the fi-nancial accounting process (A1, C and E), while the other companies (partly) decentralise some ad-ministrative tasks to the business units. These three companies are relatively small companies. For larger companies, it is not possible to centralise the whole administration, according to some inter-viewees.

There are four companies which decentralise the decison authorities, but the business units of these company have their own administration. The administration and some decision authorities of company B are decentralised, but the financial accountant send their report to the division controller.

since the local business units have more information about these rules, it is better to decentralise the decision authorities. Company I decentralised this process, because they believe that the result of the business units is the responsibility of the business unit manager, so he should also have decision au-thorities within this process.

The interviewees mentioned that various systems decrease the efficiency. An example is the consolidation process. When all the administrative tasks are executed in one system, it makes it easier to consolidate the figures. When all the units have their own administration, it makes the consolidation process harder. This occurs at company H and I. Both companies took over (parts of) other companies, but the purchased companies still use their own different systems. The interviewees mentioned that it would be more efficient to report in one system, but there was no time or money to change this system into the system used by the other units.

The results from the literature review seem to partly correspond with the empirical results. Most companies (partly) centralise the financial accounting process. However, two companies still decentralise the financial accounting process while they know that this is less efficient.

Reporting Lines

The literature mentioned two possible reporting lines, namely the solid and the dotted line. Furthermore, most companies use solid lines, because the aim of the process is to deliver accurate figures for the external reports, which are made by corporate. Another reason is that it is expected that most companies centralise the decision authorities, which means that corporate needs the information to make decisions.

A relation between the degree of centralisation and the reporting lines within the financial ac-counting process, as mentioned in the literature, can also be seen within the interviewed companies. Five companies have solid reporting lines. Of these five companies, three companies have completely centralised the financial accounting process. The other two companies (B and D) decentralised the administrative activities within the financial accounting process, but (most of) the decision authorities are centralised. Companies use solid reporting lines because the objective of this process is to deliver figures for the external report, which are made by corporate level.

The three companies with both solid and dotted reporting lines use these lines because they have an integrated financial accounting and management accounting process, and because of this, both the business unit managers as well as the division controller must be informed.

36 The empirical results partly correspond with the literature. The relation between the degree of centralisation and the reporting lines mentioned in the literature comes back in the empirical results. However, the literature mentioned a trend to centralise this process as much as possible, which should mean that there are solid reporting lines. It seems rather strange that information which is meant for external parties is firstly sent to the business unit manager instead of the division controller / corporate controller.

Hiring and Firing procedure

Table 4.5 presents a summary of the structure of empirical findings in combination with the literature review regarding the hiring and firing procedure and the scorekeeper role versus business partner role in the financial accounting process.

Company Hiring and firing

proce-dure

Scorekeeper role versus business partner role

Degree of centralisa-tion

Reporting lines

A1 Functional manager 1 1 Solid line

C Functional manager 1 1 Solid line

E Functional manager 1 2 Solid line

A2 Hierarchical manager 1 4 Dotted line

B Hierarchical manager 1 3 Dotted line

D Hierarchical manager 1 2 Solid lines

G Hierarchical manager 1 2 Both solid and dotted

lines

I Hierarchical manager 1 3 Dotted line

F Central management development team

1 2 Both solid and dotted

lines

H Central management development team

1 1 Both solid and dotted

lines

Literature Functional manager (mf)

or hierarchical manager

Scorekeeper Centralised (mf) or

combination

Solid lines (mf) or dotted lines

Table 4.5: The structure of the hiring and firing procedure and the scorekeeper role versus business partner within the financial accounting process.

The functional manager is the most favourable manager to hire the financial accountant, ac-cording to the literature, because this person is seen as the manager who receives the monthly report from the financial accountant, and because of this, he is also responsible to hire the business unit con-troller. Hence, a relation between the reporting lines and the hiring and firing procedure.