• No results found

Developments in Management Control Systems research – A literature review

N/A
N/A
Protected

Academic year: 2021

Share "Developments in Management Control Systems research – A literature review"

Copied!
35
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Developments in Management Control

Systems research – A literature review

Tutor: C.P.A. Heijes

Secondary tutor: H.J. ter Bogt Author: Remko Gerbrandij Student number: 2597470

Master: Accountancy and Controlling Date: April 11th, 2015

(2)
(3)

Table of contents

Abstract ... 4 1 Introduction ... 5 2 Theoretical background ... 7 2.1 Culture ... 7 2.2 Management control ... 8 3 Methodology ... 9

3.1 Sample and data collection ... 9

3.2 Research method ... 10

4 Findings ... 11

4.1 Other examined variables in relation with MCS... 14

4.2 Effect of other variables on aspects of MCS... 17

4.3 Support for cultural hypotheses ... 18

4.4 Perception of culture ... 21

4.5 Reasons for using the value dimensions of Hofstede ... 23

4.6 Criticism mentioned on the value dimensions of Hofstede ... 23

4.7 Research method used ... 24

5 Discussion and Conclusion... 26

6 Limitations and future research ... 29

6.1 Limitations ... 29

6.2 Future research ... 29

(4)

Abstract

The research goal of this paper was to provide a review of developments in MCS literature, thereby analyzing studies on their theoretical and methodological strengths and weaknesses. More insight is provided on the inclusion, and the effect, of non-cultural variables on aspects of MCS, how national culture affects aspects of MCS, how is culture perceived, which cultural concepts are used, and how data is data gathered in the studies. This paper, taking the form of a literature review, incorporates studies from 1996 – 2014. The findings show that there is partial support for the effect of variables, other than national culture, on aspects of MCS, that a small majority of

cultural-hypotheses are supported, that culture is mainly perceived from the value dimensional

conceptualization of culture, that although criticized, studies excessively rely on Hofstede’s value dimensions, because it is almost exclusively used in previous literature, and finally, that the mail survey was the most often used method to gather data. Limitations and suggestions for future research are discussed.

(5)

1 Introduction

The design of management control systems (MCS) has been a mainstream issue in accounting research for many years (Harrison and McKinnon, 1999; Chenhall, 2003; Cugueró-Escofet and

Rosanas, 2013). However, despite some early recognition of the importance of culture in MCS design, MCS research has often neglected the influence of culture (Harrison and McKinnon, 1999). Regarding the influence of culture, researchers to date have argued that in designing MCSs, cultural influences cannot be neglected (e.g. Van der Stede, 2003; Lee et al., 2013). In their literature review, Harrison and McKinnon (1999) examine the influence of national culture (NC) on MCS design. Harrison and McKinnon (1999) find four major weaknesses that seem to apply to cross-cultural research in MCS. First, most studies only select some cultural dimensions for use in the theoretical specification while other cultural dimensions are often ignored. Second, the studies have tended to assume that each cultural dimension is maintained with equal intensity and importance in each of the examined nations. Third, there is a tendency to treat culture as simplistic which relates to ignoring the greater depth, richness and complexity of culture. Fourth, there is an excessive reliance on the value dimensional conceptualization of culture.

Based on their weaknesses, Harrison and McKinnon (1999) argue that we need to recognize culture interdependencies with other important variables. Harrison and McKinnon (1999) argue that it is important that future research focuses more on cultural and non-cultural variables in explaining control phenomena, independently and/or in interaction. Harrison and McKinnon (1999) opt for more research in three areas: first, more insight in how non-cultural variables affect MCSs, second, more insight in how national culture affects MCSs and third, the interdependency between culture and non-cultural factors in the culture-MCS relation.

Harrison and McKinnon (1999) mention that other research that examined the influence of culture on MCS has primarily investigated culture as a stand-alone factor, assuming that culture interacts with MCSs in isolation from other variables. Chenhall (2003) argues that it seems likely that other variables, such as markets and technologies, interact with culture to affect MCS design.

Because research has examined culture as stand-alone factor, it has limited our understanding of the effect of other influential factors in the culture-control relation (Harrison and McKinnon, 1999). Harrison and McKinnon (1999) argue that future MCS research should incorporate variables other than NC and that research should examine how such variables interact with culture in the culture-control relation. Since Harrison and McKinnon’s (1999) study, other studies incorporated factors, other than NC, in the culture-MCS relation. Henri (2006), Kasurinen (2002) and Bititci et al. (2008) have examined organizational culture as factor affecting MCS. Efferin and Hopper (2007) examine the effect of ethnic tensions, organizational- and economic factors on aspects of management controls and Douglas and Wier (2005) examine whether MCSs are affected by ethical values.

However, where studies incorporate other variables in MCS research, they find mixed results for the effect of these variables. Looking at the culture-MCS interdependency, Chenhall (2003) states that compared to studies of other contextual variables, research into culture has been limited and is somewhat exploratory. Furthermore, Chenhall (2003) finds that the research has provided mixed results as to whether culture does have effects across aspects of MCS.

(6)

Harrison and McKinnon (1999) find that there is an excessive reliance on the value

dimensional conceptualization of culture, which restricts culture-MCS research. By focusing on the value dimensional conceptualization, the research fails to recognize that values are only one aspect of culture, and only one way of conceiving of culture (Harrison and McKinnon, 1999). Within the value dimensional conceptualization, studies are even more restricted and narrowed in focus through an almost total adoption of the work of Hofstede (1980) (Harrison and McKinnon, 1999). Harrison and McKinnon (1999) argue that the excessive reliance on Hofstede (1980) has led research to ignore other relevant literatures and perspectives on culture, like those in sociology, anthropology and history. Moreover, since its introduction, criticism on the value dimensions of Hofstede (1980) is growing (Drogendijk and Slangen 2006; Baskerville, 2003; Taras et al., 2012; McSweeney, 2002). Harrison and McKinnon (1999) argue that we need to reconsider the way in how culture is approached in MCS research.

The Harrison and McKinnon (1999) study was conducted almost 20 years ago. Therefore, taking their study as a starting point, the research goal of this paper is to provide a review of developments in MCS literature, thereby analyzing studies on their theoretical and methodological strengths and weaknesses. More specifically, insight is provided on the inclusion, and the effect, of non-cultural variables on aspects of MCS, how national culture affects aspects of MCS, how is culture perceived, which cultural concepts are used, and how data is data gathered in the studies. This paper, taking the form of a literature review, incorporates studies from 1996 – 2014. I do not tend to fully replicate Harrisons and McKinnon’s (1999) study, however close comparisons can be made. This study goes a bit further in that it examines non-cultural variables in MCS research, whereas Harrison and McKinnon (1999) only focus on the cultural effects on MCS design.

This is particularly relevant for academics as this study investigates if, and which, non-cultural variables are incorporated in MCS research. Furthermore, this study shows if non-cultural variables affect MCSs and if culture is perceived differently, than from the value dimensional

conceptualization. The outcomes of this study will offer insight in the relation between cultural- and non-cultural variables in MCS research and will guide academics in conducting future MCS research. This study is also relevant for practitioners, as the outcomes of this study will guide them in their decisions regarding different aspects of MCSs.

The remainder of this paper is organized as follows. Section 2 presents some cultural concepts used within the value dimensional view. Section 3 presents the methodology. Section 4 presents the findings. Section 5 presents the conclusions and discussion, and finally, section 6 presents limitations and offers suggestions for future research.

(7)

2 Theoretical background

The research goal of this paper is to provide a review of developments in MCS literature, thereby analyzing studies on their theoretical and methodological strengths and weaknesses. This chapter does not serve as a typical theoretical framework, as it does not discuss theories. This chapter is divided in two sections. Harrison and McKinnon (1999) found that there is an excessive reliance on the value dimensional concept of culture, by Hofstede. Therefore, section 2.1 discusses Hofstede’s (1980) value dimensions. Section 2.2 will address MC and MCS.

2.1 Culture

Probably the most influential and popular work on describing cultures is the work by Geert Hofstede (Kirkman, et al., 2006; Leidner and Kayworth, 2006; Taras et al., 2011). Hofstede’s research was carried out between 1967 and 1978 (Hofstede, 1983). Hofstede’s objective was to develop a commonly, acceptable, well-defined, and empirically based terminology to describe cultures, and to use systematically collected data about a large number of cultures, rather than just impressions (Hofstede, 1983). Over 116,000 questionnaires about employees’ attitudes and values were obtained and virtually all employees of the corporation, stationed over 40 countries, were surveyed (Hofstede, 1983). In examining his outcomes, Hofstede (1980) identified four dimensions which were deemed important: individualism/collectivism, masculinity/femininity, power distance and uncertainty avoidance. It was only later, that long term orientation (Confucian dynamism) was added as a fifth factor (Hofstede and Bond, 1988). In 2010 indulgence versus restraint was added as sixth factor (Hofstede et al., 2010; Minkov and Hofstede, 2011).

Individualism refers to self-serving behavior, everybody is supposed to look after his or her own interest, whereas collectivism focuses on the group, everybody is supposed to look after the interest of his or her group. Masculinity cultures are dominated by factors as competition, success and performance, the motto seems to be: ‘big is beautiful’, whereas a feminine culture is more about nurturing, warm social relationships, quality of life, and care of the weak, where the motto is: ‘small is beautiful’ (Hofstede, 1983). Power distance is how society deals with the fact that people are unequal (Hofstede, 1983). Uncertainty avoidance refers to the degree in which member of a culture are uncomfortable with uncertainties in life (Hofstede, 1983). Low uncertainty avoidance can be linked to taking risks rather easily (Hofstede, 1983). Long term orientation refers to degree of a future-oriented mentality (Hofstede and Bond, 1988). Hofstede ranked his dimensions on a scale from 0-100 (de Mooij and Hofstede, 2010).

Hofstede (1980, p. 24) defines culture as “the collective programming of the human mind that distinguishes the members of one human group from those of another”. An example of a different mental programming is provided by Smith et al. (1996) who mention that the British people will form a neat queue whenever they have to wait, while the French do not form a neat queue. Such cultural programs are difficult to change, unless one detaches the individual from his or her culture (Smith et al. 1996). Relating to culture within nations, culture changes only slowly (Hofstede, 2001).

(8)

2.2 Management control

This section presents definitions of management controls (MC) and MCS in the extant literature. How to define MCS has been a debate in a number of studies (Malmi and Brown, 2008). I will not engage in this debate as it lies outside the scope of this paper, however, I will highlight how MCS are perceived in the literature and what is meant with MC.

A central concern of organizations is ensuring that employees’ decisions and actions are aligned with the interests of the company (Merchant, 1989). MC can be seen as tool to guide/control employees’ behavior (Merchant and Van der Stede, 2007). MCs are necessary to guard against the possibilities that people will do something the organization does not want them to do or fail to do something they should do (Merchant and Van der Stede, 2007). According to Malmi and Brown (2008, p. 290-291) “management controls include all the devices and systems managers use to ensure that the behaviors and decisions of their employees are consistent with the organization’s objectives and strategies, but exclude pure decision-support systems”. Harzing and Sorge (2003) argue that control mechanisms are the instruments that are used to make sure that all units of the organization strive towards common organization goals. Simons (1995) argues that MCS are routines and procedures that managers use to maintain or alter patterns in organizational activities.

Thus, MCS relate to systems and devices used to align employees’ interests with organization’s interest. There can be different systems contributing to the alignment between employees and organizations. Some examples of MCS are: budgeting or product costing (Chenhall, 2003), strategic development, strategic control and learning processes (Merchant and Otley, 2007), expatriate deployment (Richards, 2000), performance measurement (Jansen et al., 2009), and decentralization (Chow et al., 1999).

(9)

3 Methodology

The research goal of this paper is to provide a review of developments in MCS literature, thereby analyzing studies on their theoretical and methodological strengths and weaknesses. A literature review, including studies from 1996 till 2014, is applied to reach this goal. This chapter will highlight how data was collected and how the sample was established. The final part of this section will briefly explain why a literature review was applied.

3.1 Sample and data collection

Data was collected during a one month search via ScienceDirect and Business Source Premier. ScienceDirect 1 is a scientific database that offers journal articles and book chapters from nearly 2,500 journals and 26,000 books. Business Source Premier (EBSCOhost2) is a business research database, featuring the full text for more than 2,200 journals. I have searched these databases for cultural-MCS research papers. In doing so, the following search terms were used: ‘national culture

and management control’, ‘management control and culture’, ‘management accounting and culture’, ‘national culture’, ‘organizational culture’, ‘cross-cultural research’ and ‘cultural control’.

Only peer reviewed articles were used. The articles that I found during my search were also scanned on their used references. Because the literature review by Harrison & McKinnon (1999) includes articles till 1996, my search for articles starts in 19963, and includes articles till 2014.

The search resulted in a total of 116 articles. In order to determine the relevance of the articles, these articles were scanned on their title, abstract and conclusion. In the cases were there was doubt about the relevance of the article, I scanned the introduction and method of the article. Furthermore, the articles were judged on being published in top, or very good, journals. The rating for top, or very good, journals was made on the basis of a list with top, and very good, journals, provided by the school of management4. Articles that were not published in journals rated as top, or good journals, were judged on basis of the SJR Journal rankings. This is a ranking system by the SCImago Journal & Country Rank5. The site by SJR offers a portal that includes journals and country specific indicators developed from information from a database. The portal may guide in assessing the quality of journals, but also provides the average amount of cites per document and other information about citation. The ranking system of SJR ranks the quality of journals on a scale of Q1 - Q4, were Q1 means high journal quality and Q4 means low journal quality.

After the scanning process of articles and assessing their journal rating, a total of 85 articles were excluded from the list, leading to a sample of 31 articles. One article in this sample was already used by Harrison and McKinnon (1999), so this article was excluded from the sample. There were ten articles derived from journals that were not rated as top, or very good, journals. I judged these articles on their quality by applying the SRJ journal ranking. Articles published in journals with a Q4 rating were deleted from the sample. After excluding these articles from the sample, the total sample comprised 24 articles, mainly from top, or very good, journals. All journals, from which the articles in my sample are derived, are presented in table 1.

1 http://www.sciencedirect.com/ 2

http://www.ebscohost.com/academic/business-source-premier

3

Although the Harrison and McKinnon (1999) study is from 1999, their sample includes articles until 1996

4 http://www.rug.nl/research/som-ri/organization/performance-criteria/top-and-very-good-journals 5

(10)

Table 1 - Journals used in sample

Journal Top or very

good journal

Rating by SJR

Frequency

Accounting, Organizations and Society Yes 5

Advances in accounting, incorporating advances in international accounting

No Q3 1

European accounting review Yes 1

Information & Management Yes 1

International business review Yes 3

International journal of Research in Marketing Yes 1

Journal of Business Ethics Yes 1

Journal of International Accounting, Auditing and Taxation No Q3 2

Journal of International Business Studies Yes 2

Journal of International Management No Q2 1

Journal of World Business Yes 1

Management Accounting Research Yes 3

The International Journal of Accounting No Q3 2

Total articles used 24

Total top or very good journals 18

3.2 Research method

Because this paper has a close link with the literature review by Harrison and McKinnon (1999), it seems appropriate to also apply a literature review. This might increase the comparability of my result with the Harrison and McKinnon (1999) study. In order to conduct a literature review, a sufficient body of literature is required (Hodgkinson and Ford, 20156). Harrison and McKinnon (1999) argued that cross-cultural research is still in its infancy. However, since their study was conducted almost 20 years ago, one might argue that more research is conducted since their study.

6 The authors examine starting research projects in the psychology literature world. However, their paper is

(11)

4 Findings

Recall that this study’s goal is to provide a review of developments in MCS literature, thereby analyzing studies on their theoretical and methodological strengths and weaknesses. In order to reach this goal, the papers were scanned on different characteristics, which form the basis of the following sections. Section 4.1 presents variables, other than NC, that were examined or mentioned, as influential factors of aspects of MCS. Section 4.2 presents the support found for the effect of the other variables noted in section 4.1. Section 4.3 shows the support found for cultural hypotheses. Section 4.4 will highlight how culture is perceived in the literature, by specifying which cultural concept was used. Section 4.5 will provide reasons that authors name for using the value dimensions of Hofstede. Section 4.6 will highlight whether studies are mentioning criticism on Hofstede’s value dimensions. Finally, section 4.7 will present the data gathering method. A general overviewof the findings is presented in table 2. Throughout this chapter, additional tables are presented to provide a detailed overview.

Table 2 - Overview of the articles

Study Other variables examined 7 Support for cultural hypotheses 89 Cultural concept used Reason(s) for using this concept Mentions criticism on Hofstede Data gathering MCS charac-teristics examined10 Van der Stede (2003) No No

hypotheses stated

Hofstede11: UA, IDV, PD, MSC

Extensively used Yes Mail survey Budgetary control and incentive systems Williams and Seaman (2001) No No hypotheses stated Hofstede: PD, IDV

Extensively used No Questionnaire by envelop Planning, controlling, costing, directing, decision-making Jansen, Merchant

and Van der Stede (2009) No No hypotheses stated Concept of beliefs about the purpose of corporations and Hofstede: MSC, LTO For easy comparison of countries and/or with prior studies

Yes Mail survey, inter-views Perfor-mance measures and incentive compen-sation Efferin and Hopper

(2007) Yes No hypotheses stated Focus on cultural differences, no specific concept

N/a Yes Interviews,

documents, observation Results controls, action controls 7

Relates to: are variables other than NC, explicitly examined

8 Only national culture is included under the heading: ‘cultural hypotheses’ 9

‘Yes’ in this case means that all cultural hypotheses are supported. Where ‘Partially’ is indicated it means that at least one cultural hypothesis is supported and at least one is not supported. ‘No’ means that all cultural hypotheses are not supported

10

MCS characteristics are not explicitly examined in this paper, however, I do acknowledge that this might provide more insight on the validity of the findings. Therefore, they are included in the table

11 Throughout this chapter, were Hofstede is mentioned, we refer to the work of Hofstede (1980) about the

(12)

used and personal/ cultural controls

Chenhall (2003) N/a N/a Hofstede and

other concepts drawn from anthropology

N/a Yes Literature

review Budgeting, reward systems, ABC, BSC, interactive controls, diagnostic controls Dossi and Patelli

(2008)

Yes Supported Hofstede: IDV, PD, UA, MSC

Extensively used Yes Mail survey Perfor-mance measure-ment systems Awasthi, Chow and

Wu (1998) Yes Partially supported Hofstede: IDV, Trompenaars Chosen dimension(s) best suits the research No Laboratory ex-periment Individual vs team-based perfor-mance measures Brewer (1998) No Supported Hofstede: PD,

IDV

Extensively used No Case study ABC

Williams and van Triest (2009)

Yes Not

supported

Hofstede: IDV, PD, UA, MSC, LTO and Kogut & Singh

Extensively used Yes Questionnaire Decentra-lization

Drogendijk and Holm (2012)

No Partially supported

GLOBE: PD Provides more recent data

Yes Questionnaire HQ-subsidiary control Recht and Wilderom

(1998) Yes No hypotheses stated Hofstede: IDV, PD, UA, MSC, LTO and Hofstede’s organizational culture Differentiates between national and organization-al culture No Uses previous theory to explain transferability Kaizen oriented suggestion systems

Newman and Nollen (1996) No Partially supported Hofstede: IDV, PD, UA, MSC, LTO Chosen dimension(s) best suits the research No Data gathering from systems and surveys Use of merit-based rewards Dimitratos, Petrou, Plakoyiannaki and Johnson (2011) No Partially supported Hofstede: PD, IDV, UA Extensively used and Chosen dimension(s) best suits the research No Mail survey and follow-up phone calls Decentra-lization, lateral communi-cation, for-malization Brock, Shenkar, Shoham and Siscovick (2008) No Partially supported GLOBE: PD, UA, MSC, ASS, GE, Hofstede: PD, Three concepts chosen to enhance

No Mail survey Expatriate deploy-ment

(13)

UA, IDV, MSC and Schwartz reliability and to provide different measures. Hofstede: Extensively used Segalla, Rouzies,

Besson and Weitz (2006) Yes Partially supported Ronen & Shankar’s cultural representation of Europe and Hofstede: UA, IDV R&S: consistent with recent surveys. Hofstede: Chosen dimension(s) best suits the research No Questionnaire Incentive compensati on, including parity vs equity Choe (2004) No Partially supported Hofstede: IDV, PD, UA, MSC, LTO

Extensively used No Mail survey Manage-ment accounting informa-tion systems Carmona, Iyer and

Reckers (2011)

No Supported IDV Chosen

dimension(s) best suits the research No Laboratory ex-periment Decision-making process of BSC implemen-tation Tsui (2001) No Supported Hofstede: PD,

COLL, LTO

Chosen dimension(s) best suits the research No Questionnaire Manage-ment accounting systems and budgetary partici-pation Collins, Holzmann and Mendoza (2005) No Partially supported Lara-Cantu’s sex role inventory and ‘Machism’: ChM, ClM, AM Focus is on individuals

Yes Mail survey and visiting seminars Budget attitudes Douglas, HassabElnaby, Norman and Wier (2007) Yes No hypotheses stated Hofstede: UA, IDV, PD, MSC and Ethics No reason mentioned

No Mail survey Budgetary systems / budgeting behavior Awasthi, Chow and

Wu (2001)

No Supported Hofstede: IDV & PD

Chosen dimension(s) best suits the research No Laboratory ex-periment Perfor-mance evaluation and reward systems Richards (2000) Yes Supported Focus on

cultural differences, no specific concept

N/a N/a Interviews

HQ-subsidiary control and expatriate

(14)

used deploy-ment Douglas and Wier

(2005) No Partially supported Hofstede: IDV, PD, UA, MSC, Young: the concept of ‘face’, Forsyth: relativism, idealism

Extensively used Yes Questionnaire Budgeting systems

Chow, Shields and Wu (1999)

No Supported Hofstede: IDV, PD, UA, MSC, LTO For easy comparison of countries and/or with prior studies

Yes Mail survey Decentra-lization, structuring of activities, participa-tive budgeting, standard tightness, participa-tive perfor-mance evaluation, controlla-bility filters, perfor- mance-contingent financial rewards Legend

ABC: Activity Based Costing IDV: Individualism, were IDV is NL: the Netherlands AM: Aggressive Macho mentioned, reference is made to the PD: Power distance ASS: Assertiveness Individualism/Collectivism dimension Swi: Switzerland

BSC: Balanced Scorecard LTO: Long-term orientation UA: Uncertainty avoidance

ChM: Chauvinistic Macho MSC: Masculinity, were MSC is ClM: Classic Macho mentioned, reference is made to the COLL: Collectivism Masculinity/femininity dimension GE: Gender Egalitarianism N/a: Not applicable

4.1 Other examined variables in relation with MCS

The introduction of this paper stated that culture-MCS research should incorporate factors other than NC. This section examines if such factors12 are included in the research. I found that there is a clear separation between studies who directly examine other variables and studies who mention possible other influential variables or moderating factors in their future research part. Therefore, a distinction is made in this section between a direct investigation of variables, affecting aspects of MCS and other variables or moderating factors, mentioned in future research parts by the authors. The results are summed in table 3.

12 These are all factors who are incorporated in culture-MCS research, and are examined as having an effect on

(15)

Table 313 - Variables other than NC examined or mentioned

Variables examined14 Explicitly examined Mentioned in future research part15 Organizational culture 2 5 Ethics16 2 Size 2 Politics 1 Commercial consideration 1 History 1

Level of task interdependence 1 Subsidiaries’ participation 1 Measurement diversity 1 Linking performance measurement systems to reward mechanisms 1 Global pressure 1 Performance 1

Salespeople’s effort level17 1 Salespeople’s effort direction18 1

Industry 3 Technology 2 Other 14 Total19 17 24 13

This table only includes variables other than national culture

14 Studies can examine, or name, multiple variables. Also studies can both explicitly examine variables and

mention variables in their future research part

15

These are all possible influential factors affecting aspects of MCS, that are mentioned in the future research part of studies

16

Include ethnic tensions as mentioned by Efferin and Hopper (2007) and ethical position as mentioned by Douglas et al. (2007)

17 As measured by: salespeople’s- motivation, effort effectiveness, social control, equality and shirking behavior

prevention

18

As measured by: sales force’s harmony and salespeople’s long-term orientation

19 Carmona et al. (2011); Tsui (2001); and Williams and Seaman (2001) examine national culture itself as a

(16)

There are eight studies who explicitly examine the effect of variables, other than NC, on MCS. Efferin and Hopper (2007) state a research question which is about the effect of ethnic tensions, history, politics, and commercial consideration on shaping management controls. Dossi and Patelli (2008) examine the headquarters (HQ)-subsidiary relation and hypothesize that subsidiaries’ participation in the design of performance measurement systems (PMSs), measurement diversity in the structure of PMS, linking PMS to reward mechanisms, subsidiary size and global pressure are all positively correlated with the decision influencing use of PMS. Awasthi et al. (1998) hypothesize that as the level of task interdependence increases, people would have a greater preference for team-based over individual-team-based performance measures. Williams and van Triest (2009) state

propositions regarding the HQ-subsidiary relation and examine the extent to which corporate culture20 influences decentralization of the subsidiary. Recht and Wilderom (1998) state a research question and examine the importance of organizational culture on the transferability of Kaizen-oriented suggestion systems (KOSS). Segalla et al. (2006) hypothesize that, first managers

emphasizing salespeople’s motivations, effort effectiveness and shirking behavior prevention, are more likely to choose compensation plans featuring incentives (vs. fixed compensation plans), second, managers emphasizing salespeople’s social control, parity, sales force harmony and

salespeople’s long-term orientation, are less likely to choose compensation plans featuring incentives (vs. fixed compensation plans), third, managers emphasizing salespeople’s motivation or sales effort effectiveness, and shirking behavior prevention, are more likely to choose equity (vs. parity) as a rule to allocate incentive compensation, and fourth, managers emphasizing salespeople’s social control, parity, desirability of harmony, and long-term orientation, are less likely to choose equity (vs. parity) as a rule to incentive compensation. Douglas et al. (2007) state research questions and examine whether an Egyptian manager’s ethical position will mitigate slack-creation behavior. Finally, Richards (2000) examines the effect of size and performance on HQ control over their subsidiaries and hypothesizes that larger and higher performing affiliates enjoy more autonomy, which is related to the intensity of HQ’s control.

Eleven studies name other possible influential factors, in their future research part. Van der Stede (2003) mentions that industry effects and individual manager effects might be influencing design of MCS. Chenhall (2003) mentions other variables may interact with cultures and effect MCS design. Specifically, Chenhall (2003) argues that one variable that offers promise in cultural studies is organizational culture, and that is possible that a strong organizational culture may dominate national culture in the work situation. Dossi and Patelli (2008) argue that the area of corporate culture and international management style, in relation to the use of PMSs, might be a fruitful area to explore. Newman and Nollen (1996) conducted their survey in companies with a weak corporate culture. They argue that a strong corporate culture may weaken the effect between national culture and congruence effects. Dimitratos et al. (2011) name corporate culture as moderating factor since it can make firms in the same country, behave differently. Choe (2004) names increasing globalization and compensatory effects of cultural dimensions as moderating factors. Collins et al. (2005) mention that firm-, industry-, and economy-specific variables might have an effect on the outcomes of the study.

20

(17)

Jansen et al. (2009) name bank financing calculations as possible moderating factor. They argue that these calculations in the Netherlands often only consider an individuals fixed income, whereas in the U.S., bonuses are included and that this difference might moderate the effect of NC on choice for fixed or variable salary. Chow et al. (1999) name global environment, effect of contingency variables (uncertainty, technology, organizational size, regulation), organizational culture, competition and competitive strategy as possible moderating factors. Awasthi et al. (2001) mention that factors such as social pressure, mutual monitoring and the potential for future

retaliation/cooperation may affect each team member’s decisions and actions. Furthermore, Awasthi et al. (2001) mention that new management practices are often implemented in tandem with other methods and that these management packages may affect the individual versus collective interest conflict. Finally, Douglas and Wier (2005) mention that that differences in ethical ideology would either increase or mitigate the effects of opportunity and incentive to create budgetary slack.

Carmona et al. (2011); Tsui (2001); and Williams and Seaman (2001) examine national culture itself as moderating factor. Three studies are only examining the relation between NC and MCS and are not mentioning any other influential factors at all, which are Drogendijk and Holm (2012); Brock et al. (2008); and Brewer (1998).

4.2 Effect of other variables on aspects of MCS

In section 4.1 I discussed the variables, other than NC, that are examined in the literature. While this provides insight on which variables were explicitly examined, it does not say anything about finding support for the effect of these variables. This section provides more insight in the support found for the effect of the other variables. I made a distinction between studies which state hypotheses and studies which do not state hypotheses, for example, those using research questions or propositions. I find that ten hypotheses are supported, seven are partially supported and six are not supported. The results are summarized in table 4 and table 5.

Table 4 – Support found for the effect of other variables on aspects of MCS

Hypotheses21 Frequency

Supported 10

Partially supported 7

Not supported 6

Total 23

Starting with studies who state hypotheses, Dossi and Patelli (2008) find support for the positive correlation between subsidiaries’ participation in the design of PMS, subsidiary size and global pressure on the decision influencing use of PMS. However, no support is found for the positive effect of measurement diversity in the structure of PMS and linking PMS to reward mechanisms, on the decision influencing use of PMS. Awasthi et al. (1998) find support for the hypothesis that higher levels of perceived task interdependence lead to a greater preference for team-based performance measures.

21

(18)

Segalla et al. (2006), first, only find partial support for the hypothesis that managers emphasizing salespeople’s motivations, effort effectiveness and shirking behavior prevention, are more likely to choose compensation plans featuring incentives (vs. fixed compensation plans), second, only partial support is found for the hypothesis that managers emphasizing salespeople’s social control, and parity, are less likely to choose equity (vs. parity) as a rule to incentive

compensation, third, no support is found for the hypothesis that managers emphasizing

salespeople’s sales force harmony, and long-term orientation, are less likely to choose equity (vs. parity) as a rule to incentive compensation, and fourth, no support is found for the hypothesis that managers emphasizing salespeople’s long-term orientation are less likely to choose equity as a rule to allocate incentive compensation. Richards (2000) finds support for the hypothesis that higher performing affiliates enjoy more autonomy, however, no support is found for larger affiliates enjoying more autonomy.

Table 5 – Effect of other variables on aspects of MCS

Research questions22 Frequency

Factors/variables do have an effect 5 Factors/variables do not have an effect 1

Total 6

Examining studies who use research questions, Efferin and Hopper (2007) find that ethnic tensions and commercial considerations mitigated the owner’s ability to control and that ethnic tension was found to be fuelled by history. Recht and Wilderom (1998) find that when transferring KOSS abroad, in addition to national culture, organizational culture plays a far more relevant role than first assumed. Douglas et al. (2007) find that an Egyptian manager’s ethical position does mitigate slack-creation behavior. Finally, Williams and van Triest (2009) find that organizational culture partially influences decentralization of the subsidiary23.

4.3 Support for cultural hypotheses

In order to increase our understanding of national culture as influential factor of MCS, this section will present the support found for cultural hypotheses24. Of the total of 46 investigated national culture-MCS hypotheses, 28 are supported, nine are partially supported and nine are not supported. The results are summed in table 6.

22

If research questions involve multiple other variables, these are counted as multiple variables

23

Organizational culture is measured by degree of innovativeness and degree of shared values. Only support is found for the effect of innovativeness

24

(19)

Table 6 - Support found for cultural hypotheses25 Hypotheses Frequency26 Supported 28 Partially supported 9 Not supported 9 Total 4627

Dossi and Patelli (2008) examine if collectivism, power distance, masculinity, and uncertainty avoidance characterizing the national culture of HQ are negatively correlated with the influence of PMS on subsidiaries’ decisions. No support is found for this hypothesis. Awasthi et al. (1998) examine if U.S. nationals, relative to Chinese nationals, would have a greater preference for individual-based over team-based performance measures. The second hypothesis examines if U.S. nationals, relative to Chinese nationals, would make decisions favoring their own performance over that of their teams. Awasthi et al. (1998) find that both hypotheses are partially supported, however, the effect was the other way around. Brewer (1998) finds support for the hypothesis that the level of ABC success will be greater in Malysia plants than in U.S. plants. Williams and van Triest (2009) propose that the larger the cultural distance between home and host country, the more likely a focal subsidiary will be decentralized. No support is found for this hypothesis. Drogendijk and Holm (2012) examine if HQ will exert more influence on subsidiary competence development in one of the PD relationships28, compared to others. Seven hypotheses are presented. Four hypotheses are supported and three are not supported. Newman and Nollen (1996) find support for the hypothesis that in masculine

countries, work units with more merit-based rewards practices will be higher performing than work units with less merit-based reward practices. In feminine countries, work units with less merit-based reward practices will be higher performing than work units with more merit-based rewards.

Dimitratos et al. (2011) hypothesize that, first, high power distance of the country in which the focal internationalized firm is located will be negatively related to hierarchical decentralization in the international strategic decision-making processes (SDMPs) of that firm, second, high individualism of the country in which the focal internationalized firm is located will be negatively related to lateral communication in the international SDMP of that firm, and third, high uncertainty avoidance of the country in which the focal internationalized firm is located will be positively related to formalization in the international SDMP of that firm. Support is found for the first and third hypothesis, no support is found for the second hypothesis.

25 Only hypotheses are included in this table, authors who use anything else than hypotheses, are not included

One exception is made for Williams and van Triest (2009) who use propositions. Not included in the table are: Van der Stede (2009), Williams and Seaman (2001), Jansen et al. (2009), Efferin and Hopper (2007), Chenhall (2003), Recht and Wilderom (1998), Douglas et al. (2007)

26

When a study mentions hypotheses like H1a, H1b, this is counted as two hypotheses

27

A study can include multiple cultural hypotheses

28 Drogendijk and Holm (2012) name four PD relationships, based on acceptance of power differences of HQ and

(20)

Brock et al. (2008) hypothesize that high levels of power distance, assertiveness, uncertainty avoidance and individualism, in the home country, will be positively related to expatriate deployment at the subsidiary, and that cultural distance has an asymmetrical effect on expatriate assignment at the subsidiary. Support is found for the positive effect of power distance and assertiveness, on expatriate deployment. Furthermore, support is found for the hypothesis that cultural distance has an asymmetrical effect on expatriate assignment at the subsidiary. No support is found for the positive effect of uncertainty avoidance and individualism, on expatriate deployment.

Segalla et al. (2006) hypothesize that managers from the Latin European region and the Germanic region are less likely than managers from the Anglo region to use incentive compensation for their sales force. Furthermore, Segalla et al. (2006) hypothesize that managers from the Anglo region, France and Italy, and the Germanic region are less likely than Spanish managers to decide on equity for sales force incentive compensation allocation. No support is found that managers from the Latin European region are less likely than managers from the Anglo region to use incentive

compensation for their sales force. All other hypotheses are supported.

Choe (2004) hypothesizes that, first, a greater amount of nonfinancial performance information and advanced cost control information (ACCI) is provided in Korean firms than in Australian firms, second, a greater amount of financial performance information and traditional cost control information (TCCI) is provided in Australian firms than in Korean firms, third, the positive interaction effects of advanced manufacturing technology (AMT) and non-financial performance information on production performance are greater in Korean firms than in Australian firms, fourth, the positive interaction effects of AMT and ACCI on production performance are greater in Korean firms than in Australian firms, fifth, the negative interaction effects of AMT and financial

performance information on production performance are greater in Australian firms than in Korean firms, and sixth, the negative interaction effects of AMT and TCCI on production performance are greater in Australian firms than in Korean firms. Support is found for the third and fourth hypothesis, all other hypotheses are partially supported.

Carmona et al. (2011) find support for the hypothesis that, when making capital budgeting selections, participants belonging to individualistic cultures (U.S.) will favor short-run financially superior projects, whereas participants belonging to collectivist cultures (Spanish) will favor projects aligned with long-run balanced success. Tsui (2001) hypothesizes that the interaction effects of management accounting systems (MAS) and budgetary participation on managerial performance will be different depending on the cultural background of the managers. High levels of budgetary

participation will be associated with a negative relationship between MAS and managerial

performance for Chinese managers but will be associated with a positive relationship for Western managers. Support is found for the hypothesis.

(21)

Collins et al. (2005) hypothesize that, first, being Latin American is positively relate to macho stereotypes (chauvinistic, classic, or aggressive macho), second, chauvinistic- and aggressive machos will be negatively related to a bureaucratic budget procedure attitude and an accepting budgetary culpability attitude, and positively to a risk-seeking budget risk attitude, and third, classic machos will be positively related to a bureaucratic budget procedures attitude, an accepting budget culpability attitude, and a risk-seeking budget risk attitude. Support is found for the first hypothesis. Partial support is found for the second and third hypothesis.

Awasthi et al. (2001) find support for the hypothesis that individuals working under imposed performance evaluation and reward systems are less satisfied with such systems than those working under self-selected systems, and that this difference is greater for U.S. nationals than for their Chinese counterparts. Richards (2000) hypothesizes that U.S. multinationals exercise more control over their U.K., rather than their Southeast Asian, affiliates. No support is found for this hypothesis. Douglas and Wier (2005) hypothesize that, first, participative budgeting will be higher (lower) in companies employing U.S. (Chinese) subjects, second, incentives to create slack will be higher (lower) in companies employing U.S. (Chinese) subjects, third, slack creation behavior will be more (less) frequent among U.S. (Chinese) subjects, and fourth, models of slack creation behavior will vary across different national cultural settings. Support is found for the first and second hypothesis. Partial support is found for the fourth hypothesis and no support is found for the third hypothesis. Finally, Chow et al. (1999) hypothesize 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, and that the similarity between local employees' preferred MCS and the design of the MCS would not differ between foreign and locally- owned firms. Support is found for both hypotheses.

4.4 Perception of culture

As mentioned by Harrison and McKinnon (1999) there is an excessive reliance on the value dimensional conceptualization of culture. Within the value dimensional concept, studies have almost exclusively adopted the work of Hofstede (1980). This section examines how culture is perceived and discusses which cultural concepts are used. Chenhall’s (2003) study is the only study that

incorporates concepts from anthropology. All other studies perceive culture through the value dimensional concept. Of these studies, 19 use the value dimensions of culture by Hofstede (1980). The results are summarized in table 7.

(22)

Table 7 - Cultural concepts used

(Cultural) concept used29 Frequency

Hofstede 19

GLOBE 2

Schwartz 1

Kogut & Singh 1

Trompenaars 1

Ronen & Shenkar 1

Concept of ‘Machism’ 1

Lara-Cantu’s sex role inventory 1

Young’s concept of ‘Face’ 1

Forsyth’s ethical values 1

No specific concept used 2

Total 3130

Relating to other concepts than Hofstede’s value dimensions, Drogendijk and Holm (2012) use the GLOBE measures, Brock et al. (2008) use Schwartz’s (1994) value dimensions as well as the GLOBE measures, Williams and van Triest (2009) use the cultural distance concept of Kogut and Singh (1988), Awasthi, et al. (1998) use Trompenaars’s (1994) cultural distances, Segalla et al. (2006) use Ronen and Shenkar’s cultural representation of Europe (1985), Collins et al. (2005) use the concept of ‘Machism’ and Douglas and Wier use the concept of ‘Face’ by Young (1985) and the ethical values of Forsyth (1980)31. Richards (2000) sums cultural studies and acknowledges that cultural differences exist, without using a specific cultural concept. Carmona et al. (2011) focus on the general concept of culture, not mentioning any cultural method. Efferin and Hopper (2007) mention Hofstede, however, they do not actually use Hofstede’s dimensions in their research, Drogendijk and Holm (2012) use the GLOBE measures and emphasize cultural positions. Finally, Collins et al. (2005) apply Lara-Cantu’s sex role inventory and the concept of ‘Machism’. The remaining studies all use Hofstede’s value

dimensions.

29 Some studies use multiple cultural concepts, so the frequency exceeds the total of 24 articles 30

Carmona et al. (2011) are excluded from the table because they use the individualism-collectivism dimension, derived from multiple cultural concepts

31 One might argue whether Forsyth’s ethical values are a cultural concept. However, to provide an overview of

(23)

4.5 Reasons for using the value dimensions of Hofstede

Section 4.4 shows that Hofstede’s (1980) value dimensions are widely applied. Therefore, this section discusses what the reasons were for using the value dimensions of Hofstede (1980). Although the arguments are spread, the main reason for using Hofstede is because it is extensively, almost exclusively used in prior research regarding culture. Other reasons for using Hofstede are: the chosen dimensions of Hofstede best suit the research (e.g. Awasthi et al., 1998), easy comparison of

countries, or easy comparison with prior studies (e.g. Chow et al., 1999) or, because Hofstede differentiates between national- and organizational culture (Recht and Wilderom, 1998). The results are summarized in table 8.

Table 8 - Arguments for using Hofstede

Arguments for using Hofstede32 Frequency Extensively, almost exclusively used 9

Chosen dimension(s) best suits the research 6 Easy comparison of countries, or with prior

studies

2

Differentiates between national- and organizational culture

1

No reason mentioned 1

Total 19

4.6 Criticism mentioned on the value dimensions of Hofstede

Sections 4.4 and 4.5 show that Hofstede’s (1980) value dimensions are widely applied, and that it is mainly applied because it is almost exclusively adopted in previous literature. However, as noted in the introduction of this paper, Hofstede’s (1980) value dimensional concept is not without limitations. This section provides more insight if authors are mentioning criticism on Hofstede’s (1980) value dimensions. Of the nineteen studies that apply the cultural value dimensions of Hofstede, seven of them are mentioning criticism on the value dimensions of Hofstede and twelve are not mentioning any criticism at all. Results are summed in table 9.

From his literature review, Chenhall (2003) derived that Hofstede (1980) is criticized on: assuming that different cultural values have the same intensity within a culture, assuming that cultures do not change over time, and presenting a restricted view of culture. Furthermore, Chenhall (2003) argues that the cultural values do not include subtle notions of culture, involving myths and rituals, and do not consider language and narrative.

32

(24)

Table 9 - Criticism mentioned on Hofstede

Criticism mentioned on Hofstede’s value dimensions?33

Frequency

Yes 7

No 12

Total 19

Dossi and Patelli (2008) acknowledge that there is criticism that applies to Hofstede, but argue that Hofstede’s value dimensions are still commonly used in both international business literature as well as management accounting research. Williams and Van Triest (2009) acknowledge that Hofstede is frequently criticized, but that his dimensions have been used and validated

repeatedly. There are two studies who do mention criticism on Hofstede, but only use Hofstede: to acknowledge that (cultural) differences between countries exist (Jansen et al., 2009) or because it provides greater generalizability and comparability of results with prior studies (Norman and Wier, 2005).

Two studies mention criticism on Hofstede and do not, or only partially, rely on Hofstede in their research. Drogendijk and Holm (2012) mention criticism on Hofstede and therefore, do not use Hofstede in their analysis. Instead, Drogendijk and Holm (2012) rely on the GLOBE measures. Collins et al. (2005) also mention criticism on Hofstede and, therefore, they apply Lara-Cantu’s sex role inventory and the concept of ‘Machism’. However, Collins et al. (2005) do use Hofstede as a starting point.

4.7 Research method used

To increase our understanding of the methodological aspects of the papers, this section takes a closer look at which method was used for data gathering. I find that mail surveys and

questionnaires are the most common used methods to gather data. All methods used for gathering data are summed in table 10.

33

(25)

Table 10 – Methods used for data gathering

Data gathering methods Frequency34

Mail survey35 9 Questionnaires 7 Laboratory experiment 3 Interviews 2 Documents 1 Observation 1 Literature review 1 Case study 1

Uses previous theory to explain transferability 1

Data gathering from systems 1

Follow up phone calls 1

Visiting seminars 1

Total 29

Nine studies relied on a mail survey and seven relied on questionnaires. Van der Stede (2003); Dossi, Patelli (2008); Dimitratos et al. (2011); Brock et al. (2008); Choe (2004); Collins et al. (2005); Douglas et al. (2007); Chow et al. (1999); Douglas and Wier (2005) all apply a mail survey. Studies that use a questionnaire are Williams and Seaman (2001); Jansen et al. (2009); Brewer (1998); Williams and van Triest (2009); Drogendijk and Holm (2012); Segalla et al. (2006); Tsui (2001); Douglas and Wier (2005).

Chenhall (2003) is the only author to use a literature review. Efferin and Hopper (2007) are the only authors who use a case study in their research. Furthermore, three laboratory studies were applied (Awasthi et al., 1998; Awasthi et al., 2001; Carmona et al., 2011). Other methods which were used are: interviews, studying documents, direct observation, making use of previous theory to explain transferability, data gathering from systems, follow up phone calls and visiting seminars.

34

An author can use multiple methods to gather data

35 Because multiple studies explicitly indicated they used a mail survey, the table distinguishes between mail

(26)

5 Discussion and Conclusion

The research goal of this paper is to provide a review of developments in MCS literature, thereby analyzing studies on their theoretical and methodological strengths and weaknesses. In order to reach this goal, the sample was scanned on different characteristics, these are:

incorporation of factors other than NC, the effect of these variables on aspects of MCS, support found for cultural-MCS hypotheses, perception of culture, and methods used for data gathering. The perception of culture is divided in three parts: cultural concept used, reasons for using Hofstede’s value dimensions and criticism mentioned on Hofstede’s value dimensions. A literature review was conducted, including studies from culture-MCS literature, from 1996 till 2014. This chapter serves to provide new insights that are derived from critically evaluating the findings, and shows which practical lessons can be learned.

Starting with the variables other than NC, that were examined, 17 times factors are explicitly examined and 24 times variables were not explicitly examined, but were mentioned in the future research part. This leads us to notice that it seems that authors are aware of the importance of other variables in culture-MCS research. Looking at the effect of these variables, the results are mixed. Half of the studies find that other variables have an effect on aspects of MCS, whereas the other half do not find an effect of these variables on aspects of MCS. One might question why this is the case. There can be many reasons for finding no support for the effect of other variables, for example, the research design, sample selection, sample size, measurement of variables, the presence of

moderating factors. Looking specifically at the studies in my sample, they argue that finding no support is related to other moderating factors and research design. However, there are also studies not mentioning why no support is found. More insight into the effect of other variables is clearly needed. Looking at the factors mentioned in the future research part, organizational culture is most often mentioned. It seems that organizational culture is getting increasing prominence in the literature, Kattman (2014) argued that in today’s global environment organizational culture is more influential than national culture. Henri (2006) finds that organizational culture has a direct effect on PMSs and Morsing and Oswald (2009) argue that corporate culture should be addressed when implementing MCS.

To increase understanding of NC as influential factor on aspects of MCS this paper sought to address the support found for cultural hypotheses. Of the total of 46 investigated cultural-MCS hypotheses, 28 are supported, nine are partially supported and another nine are not supported. It is interesting to see that the majority of cultural hypotheses are supported. This confirms that NC has an effect of aspects of MCS. 18 hypotheses are not- or only partially supported. One might argue why this is the case. As mentioned in the previous paragraph, this can relate to research design, sample selection, sample size, measurement of variables or the presence of moderating factors. In this case however, where no support is found, this might be due to the cultural concept that is used in the research design. Furthermore, when Hofstede’s value dimensions are applied, incorporating only some of his dimensions, might moderate the findings. Although some value dimensions are used in the theoretical section of the study, all value dimensions are of course present in the empirics (Harrison and McKinnon, 1999). Only five of 19 studies using Hofstede’s value dimensions, are incorporating five dimensions . Which means that 14 studies only use some value dimensions, or maybe even one value dimension in their research. Since the excessive reliance on Hofstede, this might mitigate the findings regarding national culture as influential factor of MCS.

(27)

Relating to the perception of culture, it turns out that it is almost exclusively perceived from the value dimensional conceptualization. Only Chenhall’s (2003) literature review includes articles that rely on other concepts, drawn from anthropology. All other studies rely on the value

dimensional conceptualization of culture. The excessive reliance on the value dimensional conceptualization of culture means that MCS research to date still ignores other perspectives on culture, for instance, from the anthropology, sociology and history literatures (Harrison and

McKinnon, 1999). Capturing culture in values means that these values are time-bonded. The values are established at a specific point in time. This might be problematic, as Taras et al. (2011) argue that some aspects of culture are growing more different. So relying on value dimensions established years ago, might not be representative of the current state. Therefore, research should incorporate other perspectives on culture.

Within the value dimensional conceptualization of culture, I find that studies almost exclusively apply the value dimensions of Hofstede (1980). The main reason for using Hofstede’s value dimensions is that Hofstede’s (1980) dimensions are almost exclusively used in prior research. However, since its introduction, criticism on the value dimensions of culture by Hofstede (1980) is growing. Baskerville (2003) and Chenhall (2003) mention that Hofstede assumes that different values have the same intensity within a culture. Hofstede (2001) stressed that cultural change would be extremely slow. However, multiple studies seem to think otherwise. Chenhall (2003) argues that countries’ cultures maybe changing faster due to globalization. Taras et al. (2012) provides evidence that suggests that cultures change much faster than expected by Hofstede (2001). Inglehart and Baker’s (2000) World Value Survey found evidence of a massive cultural change in period of several years. So to provide more recent data, Hofstede (2001) presented an updated, edition of “Cultures Consequences: comparing values, behaviours, institutions and organisations across nations”. It seems that this second edition is not without limitations either (Baskerville-Morley, 2005). Other criticism on Hofstede is that his value dimensions present an restricted view of culture, and that it ignores more subtle notions of culture like myths, rituals, language and narrative (Chenhall,2003). Taras et al. (2012) state that Hofstede’s research design raises questions about the generalizability of his findings. Finally, McSweeney (2002) criticizes Hofstede on using only small numbers of

respondents in some countries, on spreading surveys across respondents who were all employed by IBM and that the data to construct national cultural comparisons, were largely limited to responses from marketing-plus-sales employees. One might argue whether we should use other concepts in measuring culture.

There are a number of other concepts present in the extant literature. I found that several other concepts were used, for instance, the GLOBE measures by House et al. (2004), Schwartz’s (1994) value dimensions and Trompenaars’s (1993) cultural distances. House et al. (2004) measured practices and values that exist at the level of industry, organization and society. House et al. (2004) replicate and extend Hofstede’s (1980) study to test hypotheses relevant to relationships among societal-level variables, organizational practices, and leader attributes and behavior. Schwartz (1994) offers criticism on Hofstede’s (1980) study and presents an alternative conceptual and operational approach for deriving cultural dimension of values. Schwartz’s (1994) study acts as a check on Hofstede’s (1980) study, but also extends Hofstede’s (1980) study by including countries in the sample that Hofstede (1980) first did not include. Trompenaars (1985) came up with a questionnaire which was largely based on cultural and personal pattern variables or value dilemmas identified by

(28)

Parsons and Shils (1951). It was designed to tap the values of employees by asking respondents how they would act in a series of briefly described imaginary situations (Smith and Trompenaars, 1996). So should we focus more on these concepts, instead of Hofstede? It seems that these other concepts do not overcome the criticism that applies to Hofstede. The GLOBE measures offer some

improvements, like providing more recent data, however, the GLOBE measures do not solve all the problems that apply to Hofstede (Drogendijk and Holm, 2012). Drogendijk and Slangen (2006) argue that Schwartz’s framework is superior to Hofstede’s because it is based on a conceptualization of values and it was developed with systematic sampling, measurement and analysis techniques (Brett and Okumura, 1998). However, Drogendijk and Slangen (2005) conclude that it is premature to consider Schwartz’s framework to be superior of that of Hofstede (1980) because they find that the explanatory power of both methods is equal. Hofstede (1996) criticizes Trompenaars (1993) on confusing conceptual categories with dimensions and on using research samples which were small and poorly matched.

Thus, other cultural concepts are not without criticism. The criticism relating to the change of culture trough time, probably applies to all cultural concepts which try to capture value dimensions of culture. In my review I found that from nineteen studies that rely on Hofstede, seven of them are aware of criticism that applies the concept, but still use Hofstede in their research. I think it is plausible that, based on this paragraph, the explanation for this is that there simply are not ‘better’ concepts present in the current literature, that overcome the criticism on Hofstede. Therefore, it seems that studies are falling back on Hofstede’s measures.

Looking at the method used to gather data, ten studies applied methods other than the mail survey and questionnaire. The mail survey questionnaire might has led to lower costs in gathering data, the reliance on the survey questionnaire has limited our understanding of the culture phenomena (Harrison and McKinnon, 1999). It seems that other methods are clearly warranted. Looking at my sample, for example, Awasthi et al. (1998), Awasthi et al. (2001) and Carmona et al. (2011) all apply a laboratory experiment. As this chapter has argued finding no support for cultural, as well as non-cultural hypotheses, can be related to research design. Therefore, research should focus on using other methods and other research designs in order to validate findings.

(29)

6 Limitations and future research

6.1 Limitations

As with all studies, this study is not clear of limitations. First, and foremost, are the limitations applying to the sample. The sample comprised mainly articles from top, or very good journals, however, also some other non-top and non-very good journals were used. Furthermore, the sample selection is a limitation. By using only a few search terms and two different databases, the search for articles was limited. One might argue whether using more or different databases or search terms, might lead to more articles being found. Therefore, it goes without saying that the total sample of 24 articles is not exhaustive. Much more articles might be present, and including those articles might lead to different findings. In order to overcome this limitation, studies that were found in the scanning process, were also scanned on their references, if they incorporate other useful studies. However, this still forms a limitation.

Second, the relation that is examined throughout this paper is the cultural-/non-cultural- MCS relation. However, I made no distinction made between MCS design, MCS usage, MCS

performance and so on. This might mitigate the validity of my findings. The reason why no distinction is made is because this would probably lead to a too small sample, which in turn also affects the validity of the findings. I do acknowledge that only investigating the relation of NC on MCS design might be beneficial to the strength of the findings. Table 2, in chapter 4, showed which MCS characteristics were studied. Examining this table leads me to conclude that almost all studies examine the effect of cultural- and non-cultural variables, on different aspects of MCS. One might argue that this limits the validity of my findings.

Third, no distinction is made between ‘level 1’ and ‘level 2’ cultural variables. Kirkman et al. (2006) explain the difference as follows, cultures is classified as ‘level 1’ variable, when it has a direct influence on MCS and culture is classified as ‘level 2’ variable where culture is a moderating factor. Carmona et al. (2011); Tsui (2001); and Williams and Seaman (2001) examine national culture as moderating factor. Incorporating these studies in the sample, presents a bias in the current study.

Fourth, relating to the support found for cultural- and non-cultural-hypotheses, a hypothesis is often stated in relation with a positive or negative influence on a dependent variable. Whereas in section 4.2 and 4.3, no support is found for positive effects of cultural or non-cultural variables, this does not mean that there is no effect at all. It might be the case that these variables have a negative effect, which is not examined in these studies.

6.2 Future research

This study shows that culture is commonly viewed through the value dimensional lens. It seems that the majority of directions for future research, suggested by Harrison and McKinnon (1999), still apply today. This section will provide some additional directions for future research, however, there is a close relationship between future research mentioned by Harrison and McKinnon’s (1999) study.

(30)

Organizational culture was most often mentioned as possible influential factor of aspects of MCS. However, the variable is only two times explicitly examined. Furthermore, because of the increasing prominence of organizational culture, more research should explicitly examine

organizational culture as influential factor of aspects of MCS. Cultural hypotheses that are examined in my sample, offer mixed results. This might relate to excessive use of the value dimensional

concept of culture. Therefore, future research should focus more on other perceptions of culture, for instance, from anthropology, sociology and history. The fact that within the value dimensional concept of culture mixed results are found for hypotheses, suggests that we still have little knowledge on the effect of cultural as well as non-cultural variables on MCS. Therefore, future research should provide more research examining these variables, affecting aspects of MCS. The mail survey and questionnaires are still widely applied in MCS research to date. Future research should focus on using other methods and other research designs in order increase understanding of the wider perspective of culture. Other suggestions for future research are discussed in chapter 5.

Referenties

GERELATEERDE DOCUMENTEN

This supervisory controller contains a state tracker which tracks the state, a control decision maker which sends appropriate actions back to the plant and a supervisor which

All the devices and systems that are used by managers to ensure the consistency of the decisions and behaviour of employees with the organization’s strategies and

This study aims to shine a light on the extent of which Harrison & McKinnon’s (1999) criticism and suggestions, about research on the influence of national

The paper by Perego & Hartmann (2009) elaborates on organizations which have ‘’adopted environmental management systems to control the environmental impact of their products and

The first and the most important difference between working in the Netherlands and in Russia as perceived by the Russian expatriates was the absence of

It is important to check whether these interactive controls are experienced as a negative influence on employees intrinsic motivation Mallin & Bolman Pullings (2009) , or

A specific field of research in this topic is the influence of national culture on management control systems (MCS) and accounting practices.. This field of research has tried to

Daarnaast ontbreken in het huidige rapportagemodel onderlinge verbanden tussen kritieke prestatie-indicatoren en wordt naar aanleiding van dit onderzoek geadviseerd dit