Culture’s influence on used control mechanisms

(1)

Culture’s influence on

used control mechanisms

Master Thesis Msc. International Business and

Management

Faculty of Economics and Business

State University of Groningen

Date: 09-05-2008

“Believing, with Max Weber, that man is an animal suspended in webs of significance he himself has spun, I take culture to be those webs, and the analysis of it to be therefore not an experimental science in search of law but

an interpretive one in search of meaning.” (Geertz, 1973) Jan-Willem Hoekstra Nachtwachtlaan 255 1058 EJ Amsterdam janwillem.hoekstra@gmail.com Student number: 1334689

Supervisor Faculty of Management and Organisation: Dr. Bram Neuijen Second supervisor Faculty of Management and Organisation: Dr. Bartjan Pennink

(2)

ABSTRACT

(3)

FOREWORD

After all, the writing of my Master Thesis has been a tough journey. I started with the research methodology course in September 2006, which I began right after returning from an exciting summer in Egypt. Soon, I developed the idea of writing my thesis about internatio nal strategic alliances, and the control relationships surrounding those alliances. This finally evolved into the management control mechanisms used by multinationals to control foreign subsidiaries. Meanwhile, I had decided to join the International Study Project to Brazil, which meant I could not finish my thesis before the summer of 2007. Next to the preparations for my research in Brazil, I read as many articles as I could find about management control in an international context. After I finished most of this literature review, I decided that I wanted to combine the writing of my thesis with an internship within a multinational company. The most important reason for this is that I wanted to see the context in which management control mechanisms are actually used. ABN AMRO provided me with the opportunity to research the management control mechanisms used to control the off shored operational activities in Chennai, India. I conducted my research in two departments of Treasury Operations: the CLS Centre and the Client Service Group. I would like to thank my colleagues at both departments for their willingness to answer all my questions and the support they have given me during my research.

(6)

INTRODUCTION

During the summer of 2006, I worked in Egypt, where I did an internship at Egyptian International Motors in Cairo, a large international Egyptian company. The months I spent there certainly strengthened my interest in cultural differences, which already awakened during my studies at the Johannes Kepler University in Linz, Austria, in 2005. Although I worked for an international company, during my daily work I recognized the large difference between the management within an Egyptian company compared to the management of most Western companies. The large status differences between management and lower personnel and the enormous power of the CEO were two of the most striking features I encountered in Egypt. These two variables (the status differences between higher and lower management and the power of the CEO) are examples of management control mechanisms used by a company. The CEO, as I just mentioned, had a great amount of power: when he was abroad, important decisions were postponed, and everybody within the company spoke with great respect about him. Most employees, even those with university degrees and years of experience in the field, did relatively simple work and had a low amount of autonomy. As a consequence, their work was relatively easy to monitor. Furthermore, documentation of results was poor, and management mostly took decisions on “gut feelings”. These experiences made me curious about the influence of culture on the way companies are managed and employees are controlled.

(7)

In 1968, Tannenbaum already stated that the importance of control is to ensure “achievement of the ultimate purpose of an organization”. Control is also a good measure of cultural sensitivity since it is, actually, “the way one deals with people”. Numerous studies have proved that it is important to take cultural differences into account when implementing control (Lammers, 1983). One example is Crozier’s study of the implementation of Weber’s bureaucracy model. It became obvious that this model was a lot more difficult to implement effectively in Latin countries than it was in Northern and Anglo-Saxon countries, due to major cultural differences (Lammers, 1983). In recent years, the search for direct links between control and performance has faded as a major issue (Inkpen, 2001), ma inly because it is fairly impossible to establish such a link. However, culture, or to what extent cultural differences are managed, is an important determinant of an international venture’s success (Steensma, Marino and Weaver, 2000).

(8)

The literature I shortly discussed above here, and which I will discuss more extensively in the first chapter, lead me to the following research question:

How can the relationship between, on the one hand, national culture and, on the other hand, management control mechanisms used by multinational companies be conceptualized?

(9)

The remainder of my thesis thus involves my research within ABN AMRO. First, I aimed to test my hypotheses on a sample of multiple companies in order to be able to come up with more significant results. However, in order to come up with a significant result it would have been necessary to either involve a few companies, active in the same industry, or using a sample of multiple companies from multiple industries. Both options appeared to be rather difficult to realize. In the end, I decided to approach ABN AMRO to do an internship there, because I knew that the company was internationally integrated to a great extent, and therefore probably had to deal with management control of foreign activities on lower levels of the organization as well. This indeed appeared to be the case.

The main goal of this case study within ABN AMRO has been to put the national culture-management control relationship in a practical perspective, and, subsequently, to enlarge my insight in actual interaction between national culture and used management control mechanisms. During the four months I spent at the Treasury Operations departme nt of ABN AMRO, I have researched to what extent the propositions derived from literature are valid with regard to the management control mechanisms used by this department to control its off shored activities in India. However, the main purpose of my empirical research within ABN AMRO has been to find out to what extent managers within a multinational company, in this case ABN AMRO, are aware of the cultural differences they encounter when they control activities performed in culturally and geographically distant locations, and to what extent they adjust the management control mechanisms they use to cultural differences. The off shoring of operational activities to India has been an excellent case for this purpose. First, India is culturally rather distant from The Netherlands, which makes culture an important variable. Secondly, the off shoring of operational activities requires a severe amount of management control because the off shored processes are critical for business continuity.

(10)

LITERATURE REVIEW

Definition of management control

“The importance of control as an integrating mechanism within organizations stems from the fact that it reduces uncertainty, and ensures that behaviours originating in separate parts of the organization are compatible and support common organizational goals” (Egelhoff, 1984). Similar definitions of control were already brought up by authors like Weber and Fayol, who viewed control as a process of monitoring and regulating activities through hierarchical authority (Challagalla and Shervani, 1997). Then, organizational theorists like Thompson (1967) and Ouchi (1979) developed theories about (management) control in organizations. Thompson viewed the structure of an organisation as a fundamental vehicle by which organizations achieve bounded rationality. By delimiting responsibilities, control over resources, and other matters, organizations provide their participating members with boundaries. Within those boundaries, efficiency may be a reasonable expectation (Thompson, 1967). However, an organization’s structure also has to facilitate the coordinated action of the

interdependent elements in an organization. Thompson names three kinds of

interdependencies: pooled interdependence, sequential interdependence and reciprocal interdependence. Pooled interdependence is the least extended form of interdependence: the different elements in an organisation operate separately and are only connected because they are simply part of the same organisation. Sequential interdependence involves limited interdependence; elements are just dependent on the elements who supply them with inputs. Therefore, Thompson calls this kind of interdependence sequential. When there is a situation of reciprocal interdependence, all elements within an organization are somehow related. However, when there is reciprocal interdependence within an organisation, this means there is also sequential and pooled interdependence.

(11)

This means that, when organizations encounter this kind of reciprocal interdependencies, they implement a more inclusive clustering, or combination of interdependent groups, to handle those aspects of coordination which are beyond the scope of any of its components (Thompson, 1967). The usual practices of multinationals, who often arrange their business according to geographical boundaries, is a good example of this clustering.

Thompson (1967) also argues that organizations rely either on an efficiency test, an instrumental test or a social test for evaluating performance. An efficiency test can be seen, in terms of Ouchi (1979), as an equivalent of output control, where an instrumental test is comparable to behaviour control and a social test to social control. Thompson, however, also clarified that a substitution of both forms of control is in many situations rather difficult (Ouchi and Maguire, 1975). Many authors (Rushing, 1969; Blau and Scott 1962) argued that behaviour control and output control are substitutes of each other. To be able to use output control effectively, Thompson argues, goals have to be perfectly clear. In order to use behaviour control effectively, means-end relationships have to be completely understood (Thompson, 1967). Ouchi and Maguire (1975) therefore argue that the use of behaviour control and output control is not dependent on each other, but depends on the environment. They found that the use of output control increased and the use of behaviour control decreased at a higher level in the hierarchy of an organisation. This is explained by the fact that, when moving up in the hierarchy of an organisation, means-end relationships become less and less evident.

(12)

Each employee can be rewarded in direct proportion to his contribution, and the only control measure used is output control. Arbitrary rules, that characterize behavio ur control, are unnecessary. In practice, however, a market mechanism is often accompanied by a bureaucratic mechanism. After all, when the market mechanism functions perfectly, there is little reason for organizations to exist at all (Ouchi, 1979).

Ouc hi’s bureaucratic mechanism conforms quite closely to Weber’s bureaucratic model we touched upon before. This mechanism involves close personal surveillance and direction of subordinates by superiors. Somewhat arbitrary rules form the basis of a bureaucracy. Within a bureaucracy, in contradiction to a market, behaviour control is the main control mechanism. In order to use a rule, one must observe some actual performance, assess this performance and compare it to the rule in order to determine if the performance was satisfactory or not. Companies merely install the bureaucracy mechanism when the organization has to deal with a rather complex internal and external environment (Ouchi, 1979). The third mechanism Ouchi names, is the clan. Members of a “clan orga nization” have a high internal commitment to the firm’s objectives, what eliminates the need for intensive behaviour control such as that what can be observed in bureaucracies (Ouchi, 1979). A well-known example of such an organization with high internal commitment is Mintzberg’s professional bureaucracy. Some form of an informal social system exists in almost every organization, as the Hawthorne studies, among others, showed (Roethlisberger and Dickson, 1939). It differs, however, to what extent such an informal social system can serve control purposes. The following table shows the social and informational prerequisites for all three control mechanisms:

(13)

“A norm of reciprocity assures that, should one party in a market transaction attempt to cheat another, that the cheater, if discovered, will be punished by all members of the social system, not only by the victim and his partners” (Ouchi, 1979). Thus, a norm of reciprocity is essential for any control mechanism. For a bureaucracy to function, it is necessary for subordinates to obey to certain rules within the bureaucracy and to give up some autonomy. Therefore, legitimate authority is added in the table for a bureaucracy. A clan requires not only the two aforementioned prerequisites, but also social agreement on a broad range of values and beliefs (Ouchi, 1979). Traditions most importantly serve the socialization of new members.

The next table shows which conditions determine whether output or behaviour measurement is possible:

Table 2.2: Conditions determining the Measurement of Behaviour and of Output (Ouchi, 1979)

(14)

Except specialised sub-units of companies, Ouchi names organizations that stem from a culture with a strong community sense to use the clan mechanism of control more regularly.

Cultural influences on management control practices

Egelhoff (1984) based his research on control patterns in U.S, U.K, and European multinational corporations on Ouchi’s distinction between output control and behaviour control. He found that U.S. multinational corporations used relatively more output control, but less behaviour control than European companies (Egelhoff, 1984).

Contingency theories propose that the appropriate organizational structure and management style were dependent upon a set of “contingency” factors, usually the uncertainty and instability of the environment (Tosi and Slocum, 1984). “Earlier, authors like Pfeffer, Evan and Aldrich have argued that the structure of most orga nizations (and as well the control mechanisms employed) is determined more by their environment than by any purposive, technologically- motived managerial strategy” (Ouchi, 1979). Based on Ouchi’s definition of “clan control” (Ouchi, 1979), Birnberg and Snodgrass (1988) conducted a study on culture’s influence on Management Control Systems (MCS) used in American and Japanese companies. They found very straightforward relationships between national culture and the emphasis placed on the “enforcing” of management wishes (Birnberg and Snodgrass, 1988). However, multinationals have a very complex environment, since they are active in multiple cultural environments that often differ in many respects. In order to be able to control all its activities, multinationals often rely on a general Management Control System. As Van der Stede (2003) indicates, this often even leads to conformism to the home-country culture in lower levels of the organisation abroad. Van der Stede indicates that pressures towards

intracorporate isomorphism are mostly stronger than (cultural) pressures to suit local

(15)

This kind of control is more likely to be influenced by international convergence. So, it is doubtful to what extent his results can be extended to other areas of control. In my research, however, I will constrain myself to a higher level of management control: the control of a multinational’s foreign subsidiaries. Earlier in this paper I already discussed several operationalizations of management control. These contributions, however, did mostly not distinguish between higher and lower levels of management control. Sorge and Harzing (2003), however, focused explicitly on the management control of foreign subsidiaries in their research. They developed a framework that operationalizes the management control variable, which will be explained under here. One of their main findings was that the home-country culture was an important determinant of management control practices used by multinationals (Sorge and Harzing, 2003). This means that the control that is exercised from Head Quarters to the local subsidiaries is merely based on home-country practices. Sorge and Harzing conclude that internationalization strategy is merely related to industrial structures and size, whereas control is first and foremost directly related to country of origin (Sorge and Harzing, 2003). They used indirect en direct personal and impersonal control as their dependent variables, and national culture as the independent variable. Direct and indirect control are merely reflections of the output-behaviour control continuum, as discussed before. Personal and impersonal control reflect the management control practices that are founded on social interaction on the one hand (personal control), and instrumental artefacts on the other hand (Sorge and Harzing, 2003). In the table under here, one can find a summary of the management control mechanisms involved:

Personal/Cultural Impersonal/Technocratic Direct/Explicit Centralization, Direct

Supervision, Expatriate Control

Standardization, Formalization Indirect/Implicit Socialization, Informal

Communication, Management Training

Output Control, Planning

Table 2.3: a classification of Control Mechanisms on two dimensions1 (Sorge and Harzing, 2003)

1

(16)

Chow, Shields and Wu (1999) hypothesized that the controls used by firms operating in the same foreign country would not differ from each other or those of the locally-owned firms. They researched US-, Japanese- and Taiwanese-based firms active in the Taiwanese electronics industry. Most control mechanisms were similar for companies from those three countries operating in the Taiwanese electronics industry. These results indicate that the cultural environment of a (host)country is a very important determinant of the control modes companies use. Aulak, Kotabe and Sahay (1996: p. 1025) indicate that the modes of control -which is, they argue, interdependent with the level of trust- vary greatly depending on the macro-cultural environment. For the use of control mechanisms, this means that the choice of which mechanisms to employ is most likely dependent on the cultural environment as well. This aim is supported by Chow et al. (1999) and Birnberg and Snodgrass (1988). Van der Stede, however, indicates that pressures towards intracorporate isomorphism are mostly stronger than (cultural) pressures to suit local business-unit circumstances (Van der Stede, 2003). Van der Stede researched the extent to which Belgian multinational companies adjusted their budget control to local cultural business-circumstances. In contradiction to the main findings of Chow et al. (1999) and Birnberg and Snodgrass (1988) he found the pressures towards intra-corporate isomorphism to be stronger than the cultural pressures to adapt to local business-circumstances. Van der Stede, though, constrained his research to budget control. This kind of control is more likely to be influenced by international convergence. It is therefore doubtful to what extent his results can be extended to other areas of control. In summary, the influence of culture on management control can be exercised in two ways: by the host- or the home cultural environment. Initially, I will focus on the home-country influence, since this influence is proved to be the strongest with regard to the management control towards foreign activities or subsidiaries.

(17)

Hypotheses for used management control mechanisms, based on the Hofstede

dimensions (Hofstede, 1991)

I operationalize the variable national culture with the help of the work of Ho fstede (1991) and Trompenaars and Woolliams (2003). Hofstede conducted a research among IBM employees in 64 countries. As a result, he found for each country scores on five different dimensions of culture: Uncertainty Avoidance, Power Distance, Masculinity and Individualism and Long- versus Short Term Orientation. I use Masculinity, Uncertainty Avoidance, Power Distance and Individualism as operationalizations of culture. I do not use the Long Term Orientation dimension since it is less straightforward than the other dimensions and therefore more difficult to use. Secondly, it is mainly directed towards the differences between the Western and the East-Asian culture, which makes it less useful for my research.

Another famous author in the field of culture is Trompenaars. Trompenaars and Woolliams (2003) wrote an article about change across cultures. In this article, they give a framework consisting of two axes: the degree of formalisation and the degree of centralisation. They indicate that companies are organized merely according to one out of the four archetypes2 they give. In each country, one of the archetypes is most of the time prevailing. This means that, according to Trompenaars and Woolliams, when companies enter a foreign country they have to adjust to the prevailing archetype and somehow have to adopt the corresponding modes of control. However, Sorge and Harzing (2003), whose research has been discussed in the previous section, argue that home-country culture is more important in this respect. When one applies their findings on the framework developed by Trompenaars and Woolliams, one could conclude that one of the four (culturally determined) archetypes is used by multinational companies in their international ventures as well. A similar type of reasoning is used with regard to the management control of foreign subsidiaries. This means that, for example, a German multinational (generally characterized by high formalization and low centralization) organizes the management control of its foreign subsidiaries in a highly formalized and lowly centralized way.

2

(18)

However, since the dimensions used by Trompenaars and Woolliams (formalization and centralisation) are rather similar to the Uncertainty Avoidance and Power Distance dimensions used by Hofstede (1991), I prefer to limit myself to the dimensions used by Hofstede. Under here, I will discuss all these aforementioned aspects of national culture and hypothesize the supposed effects of national culture dimensions on the control mechanisms employed. First, I will discuss the hypotheses regarding direct vs. indirect control. This difference is quite straightforward in the sense that most companies chose a direct or indirect way of management control. I will discuss the influence of each of the four dimensions separately. The difference between personal and impersonal control, however, is less clear-cut, since most (international) companies use both ways of control. We will deal with the implications the Hofstede dimensions suggest on the differences between the use of personal and impersonal control across cultures at the end of this section.

Masculinity

Hofstede (1991) states that cultures differ, among others, in the extent to which assertiveness is valued, when compared to modesty. He calls this the Masculinity/Femininity dimension. Countries where modesty is regularly valued are Feminine countries, where cultures that value assertiveness more, are considered to be Masculine. The Netherlands is an example of a Feminine country, the US of a Masculine country. In Masculine countries the following values are considered to be the most important working goals: income, recognition, promotion possibilities and challenge. In Feminine countries, however, other goals were found to be more important: a good relationship with the boss and colleagues, a nice living area, and (job)security. As a consequence, financial rewards are proved to be less effective in feminine cultures than in more masculine countries (Lere and Portz, 2005). Hofstede also establishes a relationship between Masculinity and centralization. In more masculine countries like the US, “strong leaders” are preferred above consultation of subordinates, which is a more popular management style in feminine countries like Sweden, Denmark and The Netherlands.

(19)

When one relates this to Sorge and Harzing, it seems that direct personal and impersonal control is more common in masculine than in feminine countries. Therefore, I propose the following hypothesis:

• Hypothesis 1: the more masculine a country is, the more direct personal and impersonal control will be used

Individualism

The Individualism/Collectivism dimension refers to the question if people firstly identify themselves with a group or just with their own self. Hofstede (1991) defines this dimension as follows: A society is Individualistic when the mutual ties between individuals are loose:

everyone is supposed to care for his own or his or her close family. A society is Collectivist when individuals from their birth are received in strong, tight groups, who offer them lifelong protection in exchange for unconditional loyalty (Hofstede, 1991). In companies in

Collectivist countries, employment- for- life is more common than in Individualistic countries. Consequences of failure are also less rigid than in Individualistic countries. Because individuals tend to act for their best interest in Individualistic countries, work as well as incentives should be designed on an individual level (Lere and Portz, 2005). As a consequence, individual- level rewards and pay- for-performance are more common. Furthermore, training may also be more individual-based (Lere and Portz, 2005). In order to be able to be able to provide subordinates with more individual-based rewards, individual performance should be measured more accurately. Therefore, I suggest that output control is used more in Individualistic than in Collectivist cultures. This, however, does not mean that behaviour control is not used at all. Output control is, however, likely to be the most important measure that is used by companies from Individualistic countries.

(20)

Steensma, Marino and Weaver (2000) found that in collectivist and high-context cultures personalized control is emphasized. They conclude that behaviour control is more common in these cultures than output control, what also corresponds with “the fear to lose face” that is more prevalent in these cultures than in others. This fear makes objective output control less feasible. The research of Birnberg and Snodgrass (1988), however, raised the suspicion that in more collectivist cultures, direct control is used less because in a collectivist culture workers form a homogeneous group, what makes (direct) control less necessary. In Japanese firms, mana gement just needs to provide information to aid in decisions, while management in the U.S., where Individualism scores are high, must provide the same information and additionally the information needed to support the control system (Birnberg and Snodgrass, 1988). They suggest that in collectivist cultures behaviour control will be present in a more subtle and persona l form, such as social control. This discussion leads the following hypothesis:

• Hypothesis 2: the more Individualistic a country is, the more indirect personal and impersonal control is used

Power Distance

Hofstede (1991) gives the following definition of Power Distance: Power Distance is the

extent to which the less powerful members of institutions or organizations in a country expect and accept that power is unequally divided. In Arabic countries, for example, power is

(21)

Therefore I suggest the following hypothesis:

• Hypothesis 3: the higher the Power Distance in a country is, the more direct personal and impersonal control is used

Uncertainty Avoidance

A high score on Uncertainty Avoidance implies that a people are less willing to take risks, and more eager to control the risks they take (Hofstede, 1991). A high score on Uncertainty Avoidance implies that members of a culture feel threatened by unsure or unknown situations. To counter these feelings, these cultures often have a strong desire for predictability: to formal and informal rules (Hofstede, 1991). According to Lere and Portz (2005), effective management control systems operating in a strong Uncertainty Avoidance culture tend to have many informal and formal rules that limit the decision making of managers. Individuals in a strong Uncertainty Avoidance culture are more uncomfortable with uncertainty and ambiguity. Therefore, managers are probably more likely to use more behaviour control. Furthermore, in weak Uncertainty Avoidance cultures, a personalized form of behaviour control is more important (Hofstede, 1991). In strong Uncertainty Avoidance cultures, contrarily, rules and standardization of organizational processes are believed to be the best way to emp loy behaviour control. The same as or Uncertainty Avoidance also holds for centralization. Managers in countries with a high score for Uncertainty Avoidance are less likely to delegate responsibility, since it causes ambiguity. Therefore, I suggest the following hypothesis:

(22)

Personal vs. impersonal control

As already discussed before, virtually every company in the world uses a combinatio n of personal and impersonal control. Companies that do not use a form of personal control simply do not exist. Perhaps one could argue that a small family company may just use personal control, but even this is a questionable statement since even these companies use a form of planning and impersonal financial control. However, in this research our focus is at multinational companies. They definitely use both forms of control. However, the extent to which companies use personal compared to impersonal control differs. When we view this difference in a cultural context, Japanese companies clearly use more personal control than for example French companies, because the emphasis in France is more on bureaucratic (and thus impersonal) control. In Japan, companies use more social (and thus personal) control. Now, the question is: do the scores on the different Hofstede dimensions (UA, PD, MAS and IND) have any implications for the relative use of personal vs. impersonal control?

(23)

This preference is illustrated by the importance of written contracts and financial controls. These controls are most effective in the American context, because they are seen as the most objective and straightforward. Therefore, I propose the following hypothesis:

• Hypothesis 5a: the more Individualistic a country is, the more direct and indirect impersonal control is used

• Hypothesis 5b: the higher the Uncertainty Avoidance score of a country is, the more direct and indirect impersonal control is used

In figure 2.1 below here, one can find a summary of the hypotheses discussed earlier. The black lines indicate a positive relationship between both variables; the red line indicates a negative relationship.

Figure 2.1: the hypothesized relationships between national culture and management control mechanisms, based on Hofstede (1991) and Sorge and Harzing (2003)

(24)

The conceptual model on the previous page displays the hypothesized relationships between the different cultural dimensions as described by Hofstede (1991) and manage ment control mechanisms. Table 2.4 shows a similar picture, but then with explications for the preference for one of the four management control mechanisms named by Sorge and Harzing (2003). Generally, one can assume that the propositions summarized in figure 2.1 and table 2.4 have at least some explaining value because they are well rooted in literature.

However, some authors (Harrison and McKinnon, 1999; Bhimani, 1999) have recently cast doubt on the relevance of the Hofstede dimensions for the use of management control practices. In the next paragraph, alternative explanations for the different preferences for management controls across cultures will be explored thoroughly.

Cultural dimension Leads to a higher probability of Uncertainty Avoidance 1. Direct impersonal control

2. Direct personal control Power Distance 1a. Dir ect personal control

1b. Direct impersonal control Masculinity 1a. Direct personal control

1b. Direct impersonal control Individualism 1. Indirect impersonal control

2. Indirect personal control

Table 2.4: the hypothesized relationships between Hofstede’s cultural dimensions and the four management control mechanisms

Moving towards a deeper explanation of culture

(25)

Harrison and McKinnon (1999) point out the weaknesses in cross-cultural researches to date. Their main argument is the simplistic way in which culture is treated. Culture is restricted to a few dimensions, and as a consequence it loses much of its richness and, consequently, explaining power. This conceptualization of culture is almost entirely restricted to the typology of Hofstede. One of the critics Harrison and McKinnon have on the explaining power of the Hofstede dimensions is the centrality of cultural values, which differs across cultures. For example, Uncertainty Avoidance is less central in Asian cultures than in Western (Harrison and McKinnon, 1999). Another shortcoming of the use of cultural dimensions for explaining other variables is the lack of attention for interaction between these dimensions. The impact of, for example, a low score on masculinity is very different when one combines it with a high score on individualism, as is the case in for example The Netherlands. An example of the lack of explaining power the Hofstede dimensions alone have to explain preferences for management control mechanisms across cultures is the comparison between The Netherlands and the US. Both countries have rather similar scores on each dimension, except Masculinity, on which the US scores significantly higher. Looking at the hypotheses formulated earlier, this difference in masculinity should be visible in a higher preference for direct control in the US. However, the higher score on Individualism (91 against 80) suggest a higher preference for indirect controls….

(26)

Finally, I will conduct a critical review for each of the four countries, comparing the implications on management control suggested by Hofstede and other researchers who used his typology with the implications suggested by my research. As I explained above here, I conducted an analysis of these four business systems to gain a more thorough understanding of national culture and the possible influences this variable can have on used management control mechanisms by multinational companies. I chose to use Whitley’s framework, because it offers a complete overview of a country’s business system. His framework is displayed in figure 2.2 under here:

A. Ownership relations

1 Primary means of owner control (direct, alliance, market contracting)

2 Extent to which unified ownership is associated with systematic managerial integration of economic activities

3 Extent of ownership integration of production chains 4 Extent of ownership integration across sectors

B. Non-ownership coordination

1 Extent of alliance coordination of production chains 2 Extent of collaboration between competitors

3 Extent of alliance coordination across sectors

C. Employment relations and work management

1 Employer–employee interdependence

2 Delegation to, and trust of, employees (Taylorism, task performance discretion, task organization discretion)

Figure 2.2: Business system characteristics (Whitley, 1998)

In the next section, I will describe the Japanese, Dutch, French and German business system with help of Whitley’s framework (see figure 2.2). In the end, I will compare the influences of these business systems on preferred management control mechanisms within these systems. These relationships will be displayed in a similar conceptual model as I constructed for the hypothesized relationships between the Hofstede dimensions of national culture and management control mechanisms3. It is important to note here that I did not use the Indian business system for theoretical development purposes, because I just added this analysis after I decided to complement my theoretical research with an empirical research at ABN AMRO.

3

(27)

BUSINESS SYSTEMS

In this section, I provide an analysis of the business systems of The Netherlands, Japan, France, Germany and India. When I decided to conduct a case study at ABN AMRO, as a supplement to my theoretical research, I added the analysis of the Indian business system. As mentioned earlier, all business systems are described by means of Whitley’s business system typology4. At the end of this chapter, I first present a comparison between the five business systems discussed in this chapter. Then, I present a conceptual model that shows the hypothesized relationships between business system characteristics and used management control mechanisms, with business system characteristics as independent variable and management control mechanisms as dependent variable.

The Dutch business system

The Dutch economy has been mainly dominated by five multinationals the last decades (Shell, Unilever, Philips, Akzo and DSM). Another distinctive characteristic of the Dutch business system is the lack of medium sized companies with between 500 and 25000 employees, and the huge amount of small companies with less than 500 employees. Next to this “waisted” structure, Dutch business is also characterised by large scale industrial firms (Akzo, DSM, Philips, and Shell) and small- scale agricultural firms. Under here, I will give a description of the Dutch business system by using Whitley’s framework5 (van Iterson and Olie, 1991).

Ownership

Most large companies in The Netherlands are still organized similar to the first Dutch multinational: the VOC (United East Indian Company). This company, established in 1602, was characterized by dispersed ownership and strict separation of management and ownership. Family-owned companies are not common in The Netherlands, on the other hand. However, shareholders do not have many rights in The Netherlands. Most power rests with the management board of directors, who are controlled by the supervisory board. Exemplary for the power of the members of the management board is that they are appointed for an undefined period, most of the time.

4

See figure 2.2 on page 20

5

(28)

Recently, this habit is slightly changing, under influence of Anglo-Saxon practices. The Dutch business system mostly resembles the market type of owner control, although the influence of shareholders is considerably less than in the US or the UK.

Network s

Dutch business is not characterized by interlocking relationships or cross-share holdings, such as in Japan and Germany, although there exist many ties between the supervisory boards of companies, which are characterized by a practice of co-optation. But there does not exist a cooperative culture in Dutch business: most companies are independent. In some sectors (e.g. dairy and the chemical sector) there is some extent of vertical integration of production chains, but this is not common (van Iterson and Olie, 1991).

Management

Consensus is very important in the Dutch business system. In boards of directors of companies, where decisions are always a form of consensus, as well as in determining social economic policy, reaching consensus has always been very important. Wages are generally determined by means of bargaining between the government, the unions and the employers. Mostly, the general interest of society prevails above the own interest of the involved parties (van Iterson and Olie, 1999).

Institutional context

(29)

Implications for Dutch management control practices

(30)

Dutch multinationals generally use similar management practices. Their foreign subsidiaries are relatively autonomous, but are tight together by a strong corporate culture, a form of indirect personal control. The other form of control that Dutch multinationals are most likely to use is indirect impersonal control: output control and planning. In these two kinds of (indirect) control, you can clearly trace the influence of the consensus culture that characterizes the Dutch business system. Hofstede found that The Netherlands is remarkable with respect to its combination of a high Individua lism and a low Masculinity score (see table 3.1). When one follows the hypothesized effects of these scores on the common management control practices, less direct personal and impersonal control and more indirect personal and impersonal control has to be expected. Conclusively, the hypothesized effect of the Hofstede dimensions on management control is similar to the effect of the business system. In general, one could conclude that the Hofstede dimensions possess enough explaining power with regard to the commonly used management control practices in The Netherlands.

Dimension Score Power Distance 38 Individualism 80 Masculinity 14 Uncertainty Avoidance 53

(31)

The Japanese business system

The Japanese business system is dominated by the horizontal and, especially, the vertical keiretsu system (Westney, 1996). Under here I will shortly discuss this keiretsu system in terms of Whitley. Thereafter, I will consider the implications that the Japanese business system has for management control practices.

Ownership

Japanese companies are characterised by a strict separation of ownership and control. The Japanese economy is dominated by a couple of large vertical keiretsu (large industrial conglomerates, such as Matshushita, Toyota, Sony and Mazda), especially in the auto and electronics industry. These keiretsu are controlled by the lead-firm, which has minority shares in many first- and second tier suppliers and distribution channels. In turn, the lead-firm herself is owned by, often allied, banks and a combination of other vertical keiretsu (Westney, 1996; Tezuka, 1997; Nakamura, 2004). The Japanese business system is thus characterized by a clear form of alliance control. The integration of product chains is very high in the Japanese business system. This integration, however, just partly exists under ownership (Westney, 1996). The integration is merely facilitated by dense information exchange between the different parties in the production chain. For the integration of units in different sectors, ownership is much more important though. This is the so-called horizontal keiretsu, which will be dealt with in the next section.

Networks

(32)

First of all, this is the case with R&D. Secondly, suppliers in the same (vertical) keiretsu group compete with each other and with outside suppliers to excel in quality delivery, reliability and cost performance (Tezuka, 1997). Competing suppliers often cooperate with each other as well, in order to compete against other keiretsu groups.

Management

The interdependence between employees in Japan is very high. Within vertical keiretsu, the flow of personnel from the parent (or lead-firm) to subsidiaries and related companies is overwhelming (Westney, 1996). This employee-transferring maintains the strong links between the parent company and its subsidiaries and simultaneously facilitates the vertical inter- group communication. The importance of information exchange and interaction in Japan is characterized by the ringisei, a system whereby a single document detailing a decision is circulated to every department affected, collecting the seals of managers to indicate assent (Yoshina, 1968). The ringisei symbolises the characteristic Japanese “middle-up-down” management (Westney, 1996). Japanese management is furthermore characterized by decentralization and a high level of delegation to employees. Japanese employees generally have vague job descriptions and work in groups (Tezuka, 1997). However, as indicated above here, this relative autonomy is accompanied by dense information exchange and interaction in the decision- making process (Westney, 1996).

Institutional context

The Japanese institutional environment is based on the following three pillars:

• Industrial policy

• The keiretsu

• Life-long employment

(33)

Implications for Japanese management control practices

The keiretsu system, which incorporates many relatively small firms that are connected through a dense network structure, fosters control mechanisms with emphasis on a more personal form of control (Westney, 1996). Japanese firms today do heavily rely on face-to-face interactions. These clearly function less effective in very large, vertically integrated firms than in smaller organisations. Direct supervision, informal communication and expatriate control are control mechanisms frequently used by Japanese companies. The control mechanism expatriate control as described by Sorge and Harzing (2003) can be found in Japanese companies in a national context, as well. The practice of secondment of employees from the parent firm to subsidiaries within the keiretsu is probably the best example. Apparently, Japanese companies use similar control modes abroad, noting the high amount of expatriates in Japanese foreign subsidiaries. Besides these forms of direct personal control, the Japanese business system induces companies to use indirect personal control as well, especially in the form of informal communication, and the company’s educational system. These forms of indirect personal control are so well-developed within most Japanese companies, that they also replace indirect impersonal management control practices such as output control and planning. When one takes a look at the score of Japan on the Hofstede dimensions, it appears that Japan has especially high scores on Masculinity and Uncertainty Avoidance. High scores on both of these dimensions imply more direct personal and impersonal control. Also the relatively high score on Power Distance supports this supposition. However, as we just noticed while taking a closer look to the Japanese business system, the most remarkable feature of Japanese control mechanisms is probably the emphasis on personal forms of control, instead of impersonal. Because of Japan’s average score on Individualism (46), one should not directly expect such a strong preference for personal control. Dimension Score Power Distance 54 Individualism 46 Masculinity 95 Uncertainty Avoidance 92

(34)

The French business system

The French business system is one in radical transition (Boyer 1997; Redding 2005). Recently, the state-domination, which was so characteristic for the French economy, has decreased. However, the French business system is still one with a very own identity, which will be shown under here.

Ownership

The prevalent mechanism of ownership is direct control. In 1993, 28.8 percent of the Domestic Industrial Top 100 companies in France were under personal ownership and control (Mayer and Whittington, 1999). Dispersed ownership is not (yet) a widespread practice in France, with a stock- market less than half the size of the UK (Redding, 2005). French business has traditionally been characterized by (state-led) large vertically integrated companies, combined with ownership via a complex web of cross-share holdings. Within these holdings are blocs of investment by government, banks, and pension funds as well as holdings by fellow industrial firms (Redding, 2005).

Networks

French business has merely been dominated by the dominant elite of the enarques (Redding, 2005), consisting of graduates from a couple of grandes écoles. The role of these enarques has not really changed after the shifting role of the state (Mayer and Whittington, 1999). The state’s more reserved stance in the economic arena has given business people influence and negotiating power in matters of industrial development policy (Redding, 2005). Furthermore, the new more market-driven business system enabled French companies to attract more international capital, due to productivity rises. Finally, the still existing system of cross-shareholdings protects large companies in France from state intervention as well as from the rigidities of the stock markets (Redding, 2005).

Management

(35)

Institutions

The stable cross-shareholdings, where the government, banks, pension funds and major firms do participate in form the most important source of finance for most companies. International capital is an important source of finance for French companies as well, with foreign investors holding 35% of equity in the top 40 companies, measured by turnover (Redding, 2005). The French educational system is an important pillar in the maintenance of the most remarkable features of the French business system. The grandes écoles, and particularly the afore-mentioned enarques, produce the French business elite, which therefore remains a relatively close and stable group. Often, higher managers in French companies used to work for the government as well, which maintains the government’s influence on French business (Redding, 2005).

In summary, one could say that French companies consist of a small layer of managers recruited from the grandes écoles and, consequently, concentration of power at the top, with the rest of the organisation characterised by social stratification.

Implications for French management control practices

(36)

Hofstede found relatively high scores for Uncertainty Avoidance, Individualism and Power Distance for France. The high scores for Power Distance and Uncertainty Avoidance imply preferences for direct personal and impersonal control, where a high individualism score implies a preference for (indirect) impersonal control. However, when one takes a look at the French business system and the actual implications for management control practices, the influence of the high scores on Uncertainty Avoidance and Power Distance seems to be prevalent. Somewhat contrasting is the important role of personal control, something that is not expected in a country with a relatively high score on Individualism. In the case of France, the Hofstede variables have partial explaining power for the use of management control mechanisms, but an analysis of the business system clearly reveals more accurate results.

Dimension Score

Power Distance 68

Individualism 71

Masculinity 43

Uncertainty Avoidance 86

(37)

The German business system

The German business system is traditionally characterized by nationally specific social institutions, strong internationally competitive manufacturing firms and the provision of long-term capital, together with a strong emphasis on stakeholder management (Streeck, 1997). Under here, I will discuss the German business system using the framework of Whitley. Extra attention will be devoted to the German management practices, which differ in quite a few respects from those in other European countries.

Ownership

The German business system is, just like the French, characterized by a direct form of owner control. The stock market is relatively unimportant, although this is changing gradually with the growing importance of the international capital market (Streeck, 1997). Most large industrial firms are owned by a select group of shareholders, mainly (corporate) banks. Firms finance themselves less through equity than through long-term bank-credit (Streeck, 1997). Because banks generally have voting rights in excess of their actual shares, they can effectively monitor management. Labour has similar power in organizatio ns, exercised through work-councils and the supervisory board. These practices guarantee a long-term vision by all stakeholders, including the government.

Networks

(38)

Management

German higher management is characterized by a strong preference for technical experts. Other important features of German companies are the short working hours and a qualification-based organisation of work (Streeck, 1997). However, the most remarkable feature of German management is the German apprentice system. Three institutions (the government, the employers’ association and the unions) work closely together in the coordination and design of this training system. In most large German firms, at least five percent of the workforce consists of apprentices (Glunk, Wilderom and Ogilvie, 1997). This apprentice system provides a well-trained work force, and most higher managers in German firms once started as an apprentice. With, next to the apprentices, their teachers (Meisters) having an important role in German companies, interdependency among employees and between employees and employers is high. Next to the (highly- valued) technical expertise of German managers, and the great respect for competence, German managers use to have a wide span of control (Glunk, Wilderom and Ogilvie, 1997). As a consequence, employees in German organizations are relatively autonomous.

With regard to day-to-day tasks, German managers consider job descriptions and clear-cut procedures of great importance, in contrast with, for example, their Japanese counterparts. Where some researchers argue that formalization in large German organizations has led to excessive bureaucracy, poor communication, and, ultimately, inefficiency; other authors find a merely personal management style (Glunk, Wilderom and Ogilvie, 1997). This last finding is especially supported by the importance of training in German companies.

Institutions

(39)

Implications for German management control practices

When one takes a look at the development of the German economy, and the current German business system, responsibility, decentralization, commitment and interconnectedness seem to be key principles. For management control, this means that several practices are probably popular among German (international) companies. Decentralization and responsibility are central characteristics of indirect impersonal control. The direct form of impersonal control, which is so abundantly found in French organizations, is to be found in German organizations as well. However, this is accompanied by indirect personal control in the form of extensive internal training programs, a wide span of control and the importance of communication among workers. These two forms of control exemplify the interconnectedness and commitment within German organizations. The high level of education of the work force, together with the general strong commitment among German employees generally makes direct or explicit personal control (e.g. direct supervision) redundant in German companies. Compared with the management control mechanisms that the scores on the Hofstede dimensions imply for Germany, we see some similarities with the mechanisms mentioned above here. The only large deviation from the profile deduced from the Ho fstede scores is the low use of direct personal control. This may well be caused by the importance and high level of development of the German educational system.

Power Distance 35

Individualism 67

Masculinity 66

(40)

The Indian business system

In this section, I will describe the Indian culture in terms of Hofstede, and the Indian business system in terms of Whitley, just as I have done with The Netherlands, Japan, France and Germany earlier. Since India is a very large and diverse country, that has undergone tremendous societal transformation the last decades, the information I present below here is only valid on a rather general level. As mentioned earlier, I did not use the Indian business system for theoretical development purposes, because I just added this analysis after I decided to complement my theoretical research with an empirical research at ABN AMRO.

Ownership

The Indian economy is still to a great extent dominated by the state. In the beginning of the 1990s, various fiscal, investment and trade reforms were carried out and, most importantly, the role of foreign direct investment has been enlarged (Becker-Ritterspach, 2000). From $ 165 million in 1992, FDI raised to $ 4,222 million in 2002. The prevalent ownership pattern in India is family ownership (Becker-Ritterspach, 2007), but the large role of foreign direct investment in India also lead to the growing importance of market control via dispersed ownership in India.

Networks

Besides the still important role of the state, political parties and unions (that are strongly affiliated) and highly adversarial employer-union relationships are main characteristics of the Indian economy. The role of the state is mainly reflected in the strict labour laws in India, which date back to the colonial times (Gupta and Sett, 2000). These laws6 lead to an inflexible labour market, especially when it comes to recruiting, deploying, disciplining, and dismissing labour. After the liberalization of the Indian economy in the beginning of the 1990s, the rise independent labour unions led to less conflict, particularly in private sector companies (Bhattacherjee, 2001).

6

(41)

Management

In India, the (extended) family is still the main cadre of reference. The typical Indian family is patriarchal and hierarchical, with the father and, to a lesser extent, oldest son as the principal family members who generally guide the lives of the other family members (Kakar, 1997). This kind of family structure also penetrates into the work context. Like the father in the family, the superior is consulted on all major issues, and there is little scope for arguments (Becker-Ritterspach, 2007). The reluctance of Indian employees to work independently makes it rather difficult to implement downward delegation. Especially in the large cities, the pattern described above here is changing, which is mainly caused by the growing importance of Western multinationals. However, for most Indian employees the image that I just sketched is still valid (Becker-Ritterspach, 2007).

Another characteristic of Indian management is the large status difference between white- and blue collar workers, which partly stems from the Indian caste-system (Becker-Ritterspach, 2007). This distance also exists between the different categories of white collar workers, although to a lesser extent. This distance has its cause in the hierarchical nature of the Indian education system, which comprises of many different schools varying in status.

Implications for Indian management control practices

(42)

But it may be likely to have this influence in the future, which would in turn lead to higher preferences for more indirect management controls. The Individualism and Masculinity scores for India are both moderate, which leads us to the conclusion that Power Distance and Uncertainty Avoidance are the main dimensions that deserve a closer look. Table 3.5 shows that India especially has a high score on Power Distance. The main implication of suc h a high score is a high probability of direct personal management control mechanisms, which also resulted from the analysis of the Indian business system. A high Power Distance score also leads to a relative high preference for direct impersonal management control mechanisms. However, the relatively low score for Uncertainty Avoidance (India is only 45th in the ranking) implies a low preference for management control mechanisms such as standardization and formalization.

Power Distance 77

Individualism 48

Masculinity 56

Table 3.5: Scores of India on the PD, IDV, MAS and UA Dimensions (Source: Hofstede, 1991)

(43)

Whitley and Sorge combined: towards a new conceptual culture-management

control model

On the previous pages, I provided an analysis of the business sys tems and Hofstede scores of five countries. Under here, I compare these five countries with each other:

Business system characteristics

The

Netherlands

Japan France Germany India

Owner control Market Alliance Direct Direct Direct/Market Ownership

integration of production chains

Some Some High High High

Ownership

integration of sectors

Low Limited Some Some High

Alliance

coordination of production chains

Low High Low Low Low

Collaboration between competitors

Low High Low High Low

Alliance

coordination of sectors

Low Some Low Low Low

Employer-employee interdependence

High High Low Some Low

Delegation to employees

Some High Low High Low

(44)

Control mechanisms The Netherlands Business system Hofstede Japan Business system Hofstede

Direct personal control Low Low High High Direct impersonal control Low Low Low High Indirect impersonal control High High Low Low Indirect personal control High High High Some

Control mechanisms France

Business

system Hofstede

Germany

Business

system Hofstede

Direct personal control High High Low High Direct impersonal control High High High High Indirect impersonal control Low Some High Some Indirect personal control Some Some High Some Control mechanisms India

Business

system Hofstede

Direct personal control High High Direct impersonal control High Medium Indirect impersonal control Low Low Indirect personal control Some Some

(45)

The comparisons made above here lead us to the identification of differences and similarities between these five countries’ business systems and cultures, and the possible implications that these differences and similarities have for the management control mechanisms most likely used in these countries. In general, the Hofstede dimensions certainly have some explaining value for the management control mechanisms that are most likely used in different cultural environments. However, my analysis of the five business systems discussed in the previous section reveals that, if one wants to measure the influence of culture on management control more specifically, one has to take a deeper look into the context of specific cultures. Interactions between the different dimensions and the centrality of those dimensions seem to be decisive in the relationship between culture and management control.

In this chapter we have thus seen how the prevalent management control mechanisms used in five different countries correspond with these countries’ business systems. The ultimate goal of this chapter is, however, to unravel the relationship between these two variables, and to establish a conceptual model that explains the variety of management control mechanisms used around the world. I have done this by combining the models of Whitley (1998)7 and Sorge and Harzing (2003)8. Under here, I will discuss each of the three main variables (ownership, networks and management) in Whitley’s business system model, and their consequences for management control practices.

7

See figure 2.2 on page 21

8

(46)

Ownership

The Netherlands is the only country with a market form of owner control. One should not be surprised to see this combined with merely indirect control mechanisms. On the contrary, France and Germany are generally characterized by a direct form of owner control. In both countries we see a preference for direct impersonal control (i.e. standardization, formalization). The preference for either direct or indirect personal control seems to be dependent on softer variables, such as employer-employee interdependence and delegation to employees, as we will discuss later. With the ownership integration of production chains and sectors, we can observe the same pattern.

Networks

We observed that the collaboration between competitors is high in Japan and Germany, and low in The Netherlands and France. A high extent of collaboration (mainly in product development) generally leads to much indirect personal control. This relationship can be clearly observed in Japan and Germany. In France and The Netherlands, though, collaboration between competitors tends to be low. In France, direct personal control is used more often than indirect personal control, although the latter is often used additionally on a horizontal level. Dutch companies usually have a preference for indirect impersonal control (output control, planning). However, indirect personal control is frequently used in The Netherlands as well. This issue will be elaborated in the next paragraph.

Management

(47)

The analysis of the relationships between the models of Whitley and Sorge and Harzing made above here eventually led to the construction of a conceptual model that displays the relationships between business system characteristics and used management control mechanisms:

Figure 3.1: a conceptual model of the relationship between a business system and management control

Market control High value of _networks

(48)

In the preceding chapters, I have conducted a thorough literature review on management control, and I have conceptualized the relationships between national culture and management control mechanisms. By means of this analysis, I already answered my main question9 to a great extent. In the next chapters, I will discuss the findings of my empirical research within ABN AMRO. The goal of this case study within ABN AMRO has been to put the national culture- management control relationship in a practical perspective, and, subsequently, to enlarge my insight in actual interaction between national culture and used management control mechanisms. However, the main purpose of my empirical research within ABN AMRO has been to find out to what extent managers within a multinational company, in this case ABN AMRO, are aware of the cultural differences they encounter when they control activities performed in culturally and geographically distant locations. The off shoring of operational activities to India has been an excellent case for this purpose. In the next chapter, I will discuss the methods I used for my empirical research.

9

The main question I put forward in my Introduction was the following: How can the relationship between, on

(49)

METHODS

The context where my research took place in is the Treasury Operations department of ABN AMRO, and specifically the CSG (Client Service Group) and the CLS (Continuous Linked Settlement) Centre in Amsterdam. All departments in Chennai, India, that perform operational activities for ABN AMRO departments are part of ACES, a 100% daughter of ABN AMRO. The research I describe in this section targets the management control mechanisms that are used towards the activities performed in CLS Centre Chennai and the Client Service Group (CSG). The goal of this research is to retrieve a clear picture of the management control mechanisms that these two departments put in place to control the operational activities performed in Chennai, India.

The empirical part of my research focuses on the off-shoring of the operational activities of the above mentioned departments to Chennai, India. I will research which management control mechanisms as defined by Sorge and Harzing (2003)10 are used by ABN AMRO in The Netherlands to control the operational activities in India. The ultimate goal of my research is to identify the relationships between national culture and used management control practices. I already described the links between both variables above here, and retrieved a couple of hypotheses regarding these relationships. The goal of the practical part of my research is to gather empirical data on the management control processes that ABN AMRO in Amsterdam uses to control its foreign operations in India. Under here I will explain which research methods I chose in order to gather this data.

Yin (2003) suggests a couple of case study designs that researchers can use when they want to find answers to a “how or why” question, do not require control of behavioural events (such as in an experiment) and focus on contemporary events. He distinguishes multiple-case and single-case designs, and holistic and embedded case study designs. This results in four possible case study research designs. Since I dealt with two different cases independent from each other, I chose a multiple case design, but with a holistic approach, with a single unit of analysis: the management control of the off shored operational activities in Chennai, India.

10

Culture’s influence on used control mechanisms