What is the role of country on the relationship between
Top Management Team diversity and firm performance?
An Upper Echelon perspective.
Groningen, April 2012
Rijpkema A.J
Student number: S1480332 Email: aj_rijpkema@hotmail.com
Master thesis of International Business and Management
Rijksuniversiteit Groningen Faculty of Economics and Business
2
Abstract
According to the Upper Echelon theory of Hambrick and Mason (1984) a diverse Top Management Team (TMT) contributes to firm performance. However within previous Upper Echelon research there is a strong focus on the United States (US), this raises questions about the generalizability and the embeddedness of the Upper Echelon theory in a national or institutional context. This research is contributing to previous Upper Echelon research, by filling this gap and by taking a sample beyond the scope of the US to test for its generalizability. It looks more specific at TMT´s from different countries. Based on a sample consisting out of 80 companies from Germany, Japan, UK and the US. This is done by an argumentation that TMT members are likely to influence strategic decision making or day-to-day management operations. This is unique since not much research has been done outside the US. Further, almost no attention has been paid to nationality diversity within TMT´s in previous Upper Echelon research. The findings of this research show that there is mostly a positive contribution of nationality within a TMT on firm performance. And thus find some support for the importance of having foreigners within a TMT. For diversity on age and gender there seems to be a negative relationship with firm performance and thus no support for the Upper Echelon theory. Besides testing for its generalizability it is the first research that builds up a reasoning why country, by looking at its institutional context, expected to have a moderating influence on the relationship between TMT diversity and firm performance. Overall from the variance predicted within the model after taking the countries institutional context into perspective, it appears that country perhaps could have a role in explaining the relationship between TMT diversity and firm performance. Which could imply as suggested within this research that not only diversity within a TMT, but country because of differences within its institutional context, also influences the strategic decision making, as well as the behavior of managers. However it must be noted that overall because of the non-significance of the findings, there is not enough evidence to accept or generalize the findings.
3
Acknowledgement
Table of contents
Page
Chapter 1: Introduction
7Chapter 2: Background
2.1 Introduction 10
2.1.1 TMT diversity and performance 10
2.2 Empirical results from previous research on the relationship between TMT
diversity and firm performance 12
2.2.1 Positive results 12
2.2.2 Negative results 13
2.3 Measurement gap 14
2.4 Conclusion and hypothesis 16
Chapter 3: The influence of country on the relationship between
TMT diversity and firm performance.
3.1 Introduction 17
3.2 Informal institutions 19
3.2.1 Cultural values 19
3.3 Formal institutions 21
3.3.1 Corporate ownership structures related to the institutional
context of a country 21
3.3.2 Corporate ownership related to strategic decision making
and managerial behavior 22
3.5 Conclusion and hypothesis 25
3.6 Conceptual model 26
Chapter 4: Research, Method and Data
4.1 Method 28
4.2 Sample Description 28
4.3 Main variables 29
4.3.1 The TMT focus 29
4.3.2 Dependent variable: Firm performance 32
4.3.3 Independent variable: TMT diversity 32
4.3.4 Moderator: Country 34
4.3.5 Control variables: Firm size, industry, TMT size and country 34
4.4 Descriptives 37
4.4.1 Companies included in the sample 37
4.4.2 TMT´s 37
4.4.3 TMT´s- diversity differences for the US, UK, Germany and Japan 38
4.5 Distribution of the data 40
Chapter 5: Empirical results
5.1 Basic results 43
5.1.1 Hypothesis 1: There is a positive significant relationship between TMT
diversity and firm performance 43
5.1.2 Hypothesis 2: The relationship between TMT diversity and firm performance is
moderated by country 45
5.2 Robustness Check 48
5.2.1 Robustness check for hypothesis 1: ROE 48
5.2.2 Robustness check for hypothesis 1: all industries 50
5.3 Conclusion hypothesis 1 50
5.4 Robustness checks for hypothesis 2 51
5.4.1 Robustness check for hypothesis 2: with ROE 51
5.4.2 Robustness check for hypothesis 2: all industries and ROA 53 5.4.3 Robustness check for hypothesis 2: all industries and ROE 54
5.5 Conclusion for hypothesis 2 54
6
Chapter 6: Discussion and Conclusion
6.1 Conclusion 57
6.2 Implications 58
6.3 Limitations and future research 59
References
637
Chapter 1: Introduction
“The ultimate objectives of top management teams (TMT) efforts are to create a competitive advantage and ensure strong organizational performance (Marimuthu, 2009: 177). A TMT is argued to be the most important resource for a company, because it influences the strategic decision making and strategic implementation, and therefore argued to have a strong influence on the current performance as well as the development of the firm as a whole (Yang Wen & Chen, 2011). Because a TMT is responsible for strategic decision making it is argued that the characteristics of managers will have an influence on this (Roberson and Park, 2007: 552).
The Upper Echelon theory of Hambrick and Mason (1984) is a dominant theory within this research, and argues that a diverse TMT will have a positive influence on the strategic decision making process, and has a positive contribution to firm performance. An example of one of the companies following this theory is Henkel 1: who is promoting diversity within their management by focusing on gender, nationality and age. However this is a company from the United States (US), and a problem is that a lot of previous empirical investigations researching the relationship between TMT diversity and performance are only focused on the US. According to Nielsen (2010: 310) “this focus on a single country raises concerns about the generalizability of the current results”, and could prevent further extensions of generalizability to other country companies regarding their “diversity” management.
Next to the problems of generalizability another problem according to Nielsen (2010: 305) is that because of a strong focus on US samples in current Upper Echelon research, it is argued that “the embeddedness of the Upper Echelon theory in a national or institutional context is limited”. And it is argued by Nielsen (2010: 305) that this prevents researchers from drawing conclusions on the influence of country level institutional and economic factors, on the relationship between TMT diversity and firm performance. Further evidence indicating that national systems indeed may influence the relationship between TMT diversity and firm performance is presented by Crossland and Hambrick (2007) who provide strong evidence that formal and informal institutions of a country system indeed have an influence on the strategic decision making and “how much a CEO can influence firm performance”.
1
8 It is argued by Nielsen (2010: 310) that this effect also can be extended to the Upper Echelon theory. This expected moderating effect of country on the relationship between TMT diversity and firm performance, which not have been tested within previous research, will be the main focus of this research. The main research question therefore is: What is the effect of country on the relationship between TMT diversity and firm performance?
The following sub questions will be researched in order to answer the main question:
1) How is TMT diversity related to firm performance? What are relevant concepts, theories and relations between TMT diversity and firm performance in current literature.
2) What is the effect of country on the relationship between TMT diversity and firm performance? What are relevant concepts, theories and how can country be related to the relationship between TMT diversity and firm performance.
Because also the Upper Echelon theory itself will be tested within this research, and because the argued lack of generalizability of previous results, this research will also contribute to previous research. This will be done by taking a sample beyond the US, and to test the Upper Echelon theory with a sample consisting out of different countries.
In measuring the relationship between TMT diversity and firm performance, most previous Upper Echelon research focuses on demographics to measure diversity. It is argued that because nowadays many companies are globalizing, or even becoming Multinationals (Heiltjes, Ollie & Glunk: 2003) foreigners within a TMT will bring valuable knowledge, which is argued to improve strategic decision making (Huse, 2007 in Van Veen & Elbertsen, 2008: 391). Nationality diversity is thus expected to have a positive contribution to firm performance. Within current Upper Echelon research almost no attention has been paid to nationality diversity (Heiltjes et al., 2003).
10
Chapter 2: Background
2.1 IntroductionHow is TMT diversity related to firm performance according to previous literature? What are relevant concepts, theories and relations between TMT diversity on firm performance in the literature.
2.1.1 TMT diversity and performance
Research of Roberge & Dick (2010) suggests that a diverse workforce leads to an improved decision making process by increased creativity and innovation. And that heterogeneous groups are more likely to posses task-relevant knowledge, skills, abilities compared to homogenous groups. This increased innovation and creativity can be seen by the implementation of multidisciplinary teams that combine different organizational functions to ensure diversity in knowledge and perspective (Roberge & Dick, 2010: 297).
But what if we look more detailed at the TMT level and diversity, and especially its relationship with firm performance?
This is interesting because it is argued that a TMT is the most important resource for a company, because it influences the strategic decision making and strategic implementation, and therefore argued to have a strong influence on the current performance, and the development of the firm as a whole (Yang Wen & Chen, 2011).
The relationship between TMT diversity and Performance, as described before, mainly can be seen from two perspectives:
11 Park, 2007: 553). This is also argued by Roberson and Park (2007:552) who state that managers are responsible for a company’s strategic decisions, and therefore the characteristics of managers will have an influence on the performance of an organization. Or as Heiltjes et al. (2003: 89) state: “studies embedded in the Upper Echelon perspective argue that organizational choices as product innovation, diversification and acquisition strategies as well as organizational performance are influenced by the views, backgrounds and experiences top management”. The Upper Echelon theory of (Hambrick & Mason, 1984) further states that a TMT with more (demographic) diversity will positively influence the decision making process, and will have a positive contribution to firm performance. This is because according to (Marimuthu, 2008: 434) diversity leads to more creativity, innovation and improves the decision making process. Further arguments for positive results are according to Miller & Triana (2009: 756); diverse TMT´s understand the markets in which the firms operates, and they provide a broader range of knowledge compared to homogenous groups. Further they allow companies to think “outside the box” (Herring, 2009: 220) which can be related to an improved decision making, increased innovation and creativity which is in accordance with the Upper Echelon theory of Hambrick and Mason (1984).
12 work related. An overview of the arguments mentioned above, against and in favor of the Upper Echelon theory are summarized in Table 12:
Table 1: Benefits against and in favor of TMT diversity
Benefits of TMT Diversity Benefits of TMT Homogeneity
Improved decision making process (More) shared values or social integration
More relevant task knowledge, skills Less conflicts (with negative impact) because of misunderstandings
Increased innovation and creativity Faster communication processes Distinct information and skills Common communication routines Beneficial task conflict
(work-related)
2.2 Empirical results from previous research on the relationship between TMT diversity and firm performance.
Within this paragraph previous research on the relationship between TMT diversity and firm performance will be discussed, first positive results are presented and then the negative results are presented, finally a conclusion will be given, and a gap in current research will be presented.
2.2.1 Positive results
Results of Carter, Simkins & Simpson (2003:51) provide evidence that there is a positive relation between TMT diversity and firm performance, and their results showed that there is a significant positive relationship between diversity (measured by racial and gender) on firm performance. This is confirmed by a research of Erhardt, Werbel and Shrader (2003: 102) who also found that diversity (measured by gender) is positively related with firm performance. Research of Herring (2009: 220) also argued that diversity is related to business success, which according to him is because “it allows companies to think outside the box” by bringing excluded groups into the box. This enhances an organizations
2
13 creativity, problem-solving and performance. This higher problem solving capacity of a diverse workgroup, is confirmed by a findings of Watson and Kumar (1993: 599). Research of Bantel (1993, in Marimuthu, 2008) argued that diversity leads to a more competitive organization, an improvement in decision making and has a positive relation with organizational performance which is in accordance with arguments in favor of diversity as presented above. This view is also confirmed by Siciliane (1996, in Marimuthu 2008: 433) who found that diversity measured by cultural diversity, resulted in work related conflicts and improved problem solving capabilities which positively influenced firm performance (Marimuthu 2008: 433). Results of Siciliano (1996: 1313) provide support that diversity measured by occupation leads to better organizational performance. It is argued by Roberson & park (2007: 553) that (racial) diversity leads to access to a broader network, and therefore diversity is argued to have access to more resources and stakeholder groups (and therefore has a positive impact on firm performance). In their empirical results Roberson & park (2007: 563) indeed showed that with a greater balance in racial composition there is a positive relation with financial performance of the company. In a research of Marimuthu (2008) they empirically test the relationship between ethnic diversity in the board of directors and the financial performance of the firm. They found that ethnic diversity enhances a firms financial performance. Next to the findings presented above research of (Bantel & Jackson, 1989; Hambrick, Cho, & Chen, 1996; Smith et al., 1994 in Roberson & Park, 2007: 553) all showed a positive relationship between TMT diversity and organizational performance by looking at different demographics of the TMT members. Which is a further indication that diversity within a TMT is positively related to firm performance (Roberson & Park 2007: 554).
2.2.2 Negative results
14 effect on organizational performance, and will have better results compared to heterogeneous TMT´s. In a research of Certo, Lester, Dalton en Dalton (2006: 829) they find no significant support for this , but they argued that “a direct association between TMT heterogeneity and firm performance are modest and does not hold among all TMT indicators, and argued that the observed effects can probably be related to unmeasured moderators.
From previous findings it can thus be concluded that there are different results on the relationship between TMT diversity and firm performance. Some of the results provided significant support for the Upper Echelon theory and see diversity as a variety, other researches found a negative relationship and thus support for homogeneity within a TMT, and see diversity as a “separation”, or other findings even find no significant support for the relationship between TMT diversity and firm performance.
However a main problem with previous research regarding the relationship between TMT diversity and firm performance, is that it is often conducted within a single country, and almost all research is focused on American Samples (Nielsen, 2010: 310). This focus on single countries according to Nielsen (2010: 310) raises concerns about the generalizability of the results. In order to fill this gap in previous research, and to test its generelizability to other country contexts, a sample of multiple countries will be used to test this Upper Echelon relationship within this research.
2.3 Measurement gap
15 Companies increasingly are globalizing, and often are active within many countries, or are even becoming a Multinational Corporation (MNC) (Heiltjes et al., 2003). Knowledge of foreign markets is argued to become very important. A focus on nationality as a demographic indicator can become increasingly relevant in Upper Echelon research, and according to Heiltjes et al. (2003: 89) the focus on nationality as a measure for demographics, is missing in previous Upper Echelon research. This is indeed confirmed by the previous empirical results that have been described earlier within this paper.
But why is nationality diversity within a TMT for companies that operate in different markets important?3
It is argued by Athanassiou & Nigh (2000: 484) that within a MNC, TMT members create social capital in the markets which they visit and leverage, and to create knowledge of these markets. This knowledge is shared at the TMT level which is used to formulate and implement strategy and to fulfill the goals. Carpenter et al. (2004: 762) found that the amount of international experience within a TMT has a positive impact on performance of multinational firms, however the sample only researched the US. It is further argued by Van Veen & Elbertsen (2008: 396) that a high level of nationality diversity within a board should be an advantage for companies that operate in multiple countries. This is also argued by Carpenter et al. 2004 (in van Veen & Elbertsen, 2008: 396) “that it seems likely that because of a globalizing economy, a global strategic posture should be an advantage and also will lead to a better company performance”. So regarding the Upper Echelon theory, nationality diversity is also argued to be considered as important in relation to firm performance. According to this “knowledge based approach” of the Upper Echelon theory, it is further argued that foreigners within a TMT also should bring valuable knowledge, related to foreign markets, which should improve strategic decision making (Huse, 2007 in Van Veen & Elbertsen 2008: 391), and argued to be positively related to firm performance. This is confirmed by Zahra & Filatochev (2004: 892) who argue that organizational learning occurs if members of TMT´s acquire new knowledge and skills, or as firms acquire new knowledge that is brought in by board members. Therefore it can be expected that a higher nationality diversity within a TMT of a MNC is expected to have a positive relation with company performance.
3
16
2.4 Conclusion and hypothesis
Overall as shown, the link between TMT diversity and company performance (following Upper Echelon perspective), has been studied widely, however mostly within the US. Nationality is argued to have a positive influence on performance, which is a gap in current research. Therefore to test the “classical theory” of the Upper Echelon research, I will first contribute to current Upper Echelon research by taking nationality into account as a demographic indicator for TMT diversity, and also use age and gender as other demographic indicators to test the Upper Echelon theory. And secondly I will contribute to previous research, by taking a scope of several countries to test for its generalizability.
Therefore the following Hypotheses are derived:
1) TMT diversity has a significant positive relationship with firm performance
A) There is a significant positive relationship between gender diversity and firm performance.
B) There is a significant positive relationship between age diversity and firm performance.
17
Chapter 3: The influence of country on the relationship between
TMT diversity and firm performance.
3.1 Introduction
Another gap next to absence of nationality diversity as an indicator, and the limitations to the country scope within current Upper Echelon research, is that extensions of the theory into an international or institutional context are limited. In other words as argued by Nielsen: (2010: 305) “the embeddedness of the Upper Echelon theory in a national or institutional context is limited”. And according to Nielsen (2010: 305) this prevents researchers from drawing conclusions on the influence of country institutional and economic factors on the relationship between TMT diversity and firm performance, as well as for generalizing current results (Nielsen 2010: 310).
As described earlier in this paper and as argued by Carpenter et al. (2004: 772) the Upper Echelon theory focuses on diversity by looking at demographics as an indicator for the behavior of managers. However it is argued by Carpenter et al. (2004) that previous Upper Echelon research is lacking other determinants of their behavior. Evidence indicating that (national) institutions may influence the relationship between TMT diversity and firm performance is presented by Crossland and Hambrick (2007) who provide strong evidence that national institutions have an influence on the strategic decision making within a firm, and how much a CEO can influence performance. Crossland and Hambrick (2007: 770) argue that environmental factors almost every time have been measured by industry, something which is also shown in previous Upper Echelon research. No attention has been paid to macro environmental forces or managerial constraint within different countries (Crossland and Hambrick: 2007,770).
But why is country argued to have an influence on the relationship between TMT diversity and firm performance?
18 According to North (1990, in Crossland & Hambrick, 2007: 771) there can be made a distinction between informal and formal institutions within a country, that both have an impact on the strategy of organizations. Informal institutions are according to North (1990, in Crossland and Hambrick 2007: 771) “unwritten societal fundamentals” that consist out of norms, conventions and values. On the other side the formal institutions, which are political rules and procedures, economic rules or other rules that have an impact on how the organization operates.
Institutions within a country are thus argued to have in influence on how the organization operates, but how can this be related to the behavior of managers?
According to Hicheon, Heechun & Hoskisson (2010: 1142) institutions shape the firms behavior and performance in a society. According to Tonoyan, Strohmeyer, Habib & Perlitz (2010: 804) institutions “influence the individual decision making by signaling which choice is acceptable, based on norms and behavior within a society”. Tonoyan et al. (2010: 804) argue that institutions, have an impact on the behavior of people, as well as the behavior of organizations. And according to Aldrich & Fiol (1994, Tonoyan et al.,2010: 805) “institutions have a big influence on how individuals but also firms will act and compete”. Further evidence is presented by Mccann, Hassard and Morris (2010:349) who argue that national institutions have a strong influence on the behavior of firms in different nations. According to Tipton (2009: 402) “government sets the rules” and it is therefore argued that structures of management can be related to the institutions of a country.
Overall country institutional systems are argued to have an impact on the decision making capabilities, as well as having an influence in shaping the behavior of managers and firms. This will be described in more detail within Section 3.3 by an argumentation based on informal institutions: by looking at national values and within Section 3.4 an argumentation based on formal institutions: by looking at corporate ownership4. Finally within Section 3.5 conclusions will be drawn, and a hypothesis will be presented.
4
19
3.2 Informal institutions
Informal institutions are argued by Whitley (1999 in Schwens, Eiche & Kabst, 2011: 332) to exist because of a cultural background, this cultural background describes patterns regarding behavior, collaboration and identity. Culture is argued to have a strong effect on executive decisions or strategic orientations (Hitt, Dacin, Tyler & Park, 1997: 160). This is confirmed by Davis, Schoorman and Davidson (1997 in Crossland and Hambrick 2007: 771) who argue that because “decisions of executives are made within an environment of social norms, morals and expectations, and because cultural values will differ across countries, this will have lead to different degrees of constraint on the decision making process”.
This influence of the informal institutional environment by looking at cultural values will be discussed within the next paragraph. Previous research used a variety of frameworks, to show that national cultural values are related to workplace behaviors and attitudes (Kirkman, Lowe & Gibson, 2006: 285). However because of limitations to this paper a more detailed argumentation of cultural values will be based on individualism and collectivism, and uncertainty avoidance5
3.2.1 Cultural values
Poortinga (1992, in Bond & Smith, 1996: 209) defines culture as “a set of shared constraints that limit the behavior repertoire available to members of a certain group”, from this it can be expected that the cultural environment in which a TMT operates, has an impact on the behavior of managers within a TMT.
If looking more specific at national culture and its expected influence on the behavior of executives, first by looking at how culture affects conflict resolution. Findings of Tse et al. (1994 in Bond & Smith, 1996: 222) showed differences in conflict avoidance, and how conflicts have been resolved between different cultures. Results of Tinsley (2001 in Tsui, Nifadkar & Yi Ou, 2007: 443) provide evidence that different cultural values can be related to differences within management strategies regarding their conflict management (Tsui et al., 2007: 443).
5
20 National culture not only seems to influence conflict behavior, but also seems to have an influence on team processes, cooperation and the decision making behavior of executives.
It is argued by Crossland and Hambrick (2007:772) that within an individualistic culture, stakeholders are more tolerant, and executives tend to be higher in self-reliance and within the decision making process often more unilaterally. Within a collectivistic culture, stakeholders are argued to be more concerned with decision making, and there will be more consensus based decision making (Hofstede, 2001; Smith, Dugan, and Trompenaars, 1996 in Crossland and Hambrick, 2007: 772). On the other side in more individualistic cultures there tend to be an emphasis of personal goals over group goals. And in more collectivist cultures people tend to subordinate personal goals over the goals of the collective (Davis, Schoorman & Donaldson: 1997: 34), and will maintain harmony (Halim and Chew 2008: 58). This difference between collective and individualistic cultures, is confirmed by a research of Huang and v/d Vliert (2003: 162) that states that: “within an individualistic culture people tend to be self reliant, self motivated and place more value on individual interests, and within individualistic countries there is more emphasis on extrinsic aspects then in collectivistic countries”. Relating this to the behavior of managers within a research of Lawrance (1992, in Melahi, 2001: 47) they argue that within more individualistic cultures managers are characterized by values as “autonomy, ambitious and individual security”, “with a strong focus on achievement and individual initiative”. On the other hand in a collectivist culture, managers find interpersonal relations more important than individual achievements (Melahi, 2001: 47). And there is more tolerance and forgiveness (Blunt and Jones, 1997 in Melahi, 2001: 47).
21 Overall it can thus be argued, looking at previous research, that (national) culture is expected to have an influence on the behavior as well as the strategic decision making of managers.
3.3 Formal institutions
Next to the argumentation of the influence of informal institutions, on the behavior as well as the strategic decision making of managers. Formal institutions by looking at different ownership structures, also argued to have an influence on the behavior, as well as the strategic decision making of managers. To build up this argumentation I will first relate ownership structures to the formal institutional context of a country. Then I will focus on two specific features of a formal institutional context, in order to provide a more detailed description of the expected impact of corporate ownership on the behavior of managers, as well as its influence on the strategic decision making process.
3.3.1 Corporate ownership structures related to the institutional context of a country
It is argued by Roe (1991 in Pederson & Thomson, 1999:44) that ownership structures are influenced by the legal systems of a country. It is argued that the regulation of financial systems, determines whether countries adopt a certain ownership structure (Pederson& Thomson, 1999: 46).
A common model in the US, UK, Canada and Australia is the Anglo-Saxon model which is a market based ownership model (Kwee,Van Den Bosch &Volberda, 2011: 989, Moerland,1995:449) this system is characterized by widely dispersed shared shareholdings (Zysman, 1983 Cooke & Sawa, 1998: 217, Pederson & Thomson, 1999) and an active market for corporate control with well developed financial markets (Moerland, 1995:448).
22 The Latin model can be applied on countries as Italy, Spain and France, and also is characterized by a closely held ownership.
According to Moerland (1995: 450) the network based systems of Rhineland and the Latin model both can be seen as the opposite of the Anglo-Saxon market oriented systems, because of the involvement of banks and a more concentrated ownership. These (formal) institutional differences such as the involvement of banks within Germany and Japan, and tight family and state ownership in countries as France and Italy are argued to influence the behavior of firms (Moerland, 1995: 444).
3.3.2 Corporate ownership related to strategic decision making and managerial behavior
Corporate ownership is thus argued to influence the behavior of firms. But how can this be related to a TMT?
A more detailed description of the expected impact related to differences in corporate ownership on the strategic decision making of managers within a TMT as well as on their behavior, is based on the model of Rhineland and the Anglo-Saxon model. However I am aware of the fact that there are other corporate ownership structures, the focus is used to build up argumentation why corporate ownership as a formal institution is argued to have an influence on the behavior, as well as the strategic decision making of managers. With respect to the limitations of this paper.
23 But how can corporate ownership be related to the behavior of managers themselves?
Bakker, Evers, Hovens, Snelder & Weggeman (2005) related the models of corporate ownership to define the Anglo-Saxon and Rhineland management style6.
First a one-tier management within the Anglo-Saxon system. Within this system the board is responsible for managing as well as supervising. The chairman is a CEO. The second management style is a two-tier board, and is characterized as the “Rhineland thinking”, within this management style the management is consisting out of a supervisory board and an executive board. The decision making power is divided between both boards (Bakker et al., 2005: 76). Within a one-tier board there is much influence of the shareholders on the company’s strategic decisions, with a strong focus on short term profit. The focus on “shareholder value” within a one-tier board, has as a result that financial targets are used to test organizational performance, next to this the financial targets are also linked to bonuses of the managers. From this it is argued that the behavior of managers within an Anglo-American system is more focused on extrinsic motivation, and a strong focus on finance and short term performance (Bakker et al., 2005).
On the other side within a two-tier board, employees, suppliers and financers also have influence on the company’s strategic decisions, which is different from the one-tier board. Within the Rhineland thinking, it is expected that managers are thinking beyond money, and want to meet the expectations of a customer (Bakker et al., 2005). Managers should have knowledge of their products, and are also responsible for implementation of their own plans. This separation of control and decision making within the Rhineland model however is argued to lower the strategic decision making capability of a TMT. This is because of the involvement of banks as well as its concentrated ownership and a more long term focus on performance.
Overall it can be concluded that the differences related to the Anglo-Saxon and the Rhineland model are related to different countries and show differences within corporate ownership systems (argued to be a formal institution), but also reflect cultural differences
6
24 (as an informal institution), as being either more individualistic or more collectivistic. Both institutional differences argued to have an influence on the behavior of managers as well as the strategic decision making itself. An overview of the differences can be seen within Table 2:
Table 2: Overview of the aspects of the Anglo-Saxon and Rhineland model
Aspects Anglo-Saxon Rhineland
Focus Short term profit Continuity and trust
Dominant thinking Financial Industrial
Performance orientation
Next 3 months Continuity
Main aim Money, power,
heroism
Expertise, know-how, proficiency Goal of organization Supporting vehicle Labor community
Manager his tasks MBA Cooperating foreman
Motivation Extrinsic Intrinsic
Model of man used Mechanistic Humanistic
Reward depends on Productivity Function
Labor Cost factor Social component
Employees Input Embodiment of
organization Finance through Stock exchange Banks and family
Between companies Competition Cooperation
Driven by Technology, market Design and science
Bankruptcy Start something new The end
Relies on Power Economic power
Management philosophy
Control Involvement
Risk orientation Control mechanisms Trust
Objective Cost control Performance
enhancement Cultural differences Individualism
High power distance
Collectivism
25
3.5 Conclusion and hypothesis
The Upper Echelon theory states that managers organizational choices and behavior, are reflected by the views and backgrounds as well as the experiences of managers. However from the previous argumentation, it can be expected that country, because of its institutional context, also has an impact on the behavior of managers, as well as on the strategic decision making process itself. Therefore it is expected that not only diversity within a TMT, but also country because of differences within its institutional context, expected to have an influence on the strategic decision making, as well as on the behavior of managers. Therefore it is expected that country will moderate the relationship between TMT diversity and firm performance. To test this the following hypothesis is derived:
26
3.6 Conceptual model
From the first part of the literature review it can be expected that there will be a positive relationship between TMT diversity and firm performance regarding the Upper Echelon theory. From the second part of the literature review it is expected that country, because of its institutional context, has an impact on the strategic decision making capabilities, as well as on the behavior of managers within a TMT. And it is therefore argued that country is expected to moderate the relationship between TMT diversity and firm performance. This is presented within the following conceptual model:
Positive influences
+/-
Figure 1: The relationship between TMT diversity and firm performance and the role of country within this perspective.
1. TMT diversity Independent Demographics Nationality Age Gender 3. Firm Performance Dependent ROE /ROA 2. Country Moderator
27
3.7 Overview of Hypotheses
The following hypotheses are derived from the literature review, and will be tested.
1) TMT diversity has a significant positive relationship with firm performance.
A) There is a significant positive relationship between gender diversity and firm performance.
B) There is a significant positive relationship between age diversity and firm performance.
C) There is a significant positive relationship between nationality diversity and firm performance.
28
Chapter 4: Research, Method and Data
4.1 MethodThe research will be a quantitative research, therefore the sample size needs to be large, and consistency is critical. Quantitative research makes it possible to gather data which can be statistically analyzed (Thomas, 2004), and makes it possible to identify the relations involved in establishing the existence and strengths of links between variables in a study (Thomas, 2004). For the statistical analysis of the sample SPSS will be used, to provide information about the descriptive of the sample, to test correlation among the variables, and to test the hypotheses.
4.2 Sample Description
In my research I will focus on the fortune top 500 global companies, because most of previous research is focused on using the US companies fortune top 500 (Wiersema & Bantel, 1992; Miller & Triana, 2009). I will make a distinction in that perspective, and can contribute and add value to current research.
More specific the sample will consist out of companies from the US, UK, Germany and Japan. This can be related to the argumentation of Chapter 3. The US and the UK can be seen as Anglo-Saxon countries, Germany as Rhineland, and Japan argued to be the opposite of an Anglo-Saxon, with a network based system and because of a strong involvement of banks, as well as a long term focus on performance. Therefore this focus makes it possible to test for the moderating effect of country, as well as testing for the generalizability of the Upper Echelon theory itself.
29 Further because it is important that data about TMT´s from the different companies can be gathered including data as nationality, gender, age and TMT size, as well as for firm performance, and more important that this data is consistent among the companies. The focus on companies from the global fortune 500 companies will ensure that most information will be accessible through annual reports. Data about the TMT composition and TMT size is gathered from the ORBIS7 database, annual reports and company websites. If data could not be find then secondary data from websites as www.zoominfo.com, investing.businesweek.com, boardroominsiders.com and people.forbes.com will be used to find information about the TMT members. The other information including financial data of the companies, as well as industry and company size will be collected from the ORBIS database to ensure data is consistent among the companies.
4.3 Main variables
4.3.1 The TMT focus
Within the Upper Echelon theory it is argued by Hambrick (2007:334) that the focus on the characteristics of the TMT will provide stronger explanations of the organizational outcomes, then a focus on the individual top executive as a CEO alone. This is strengthened by much empirical support (Finkelstein and Hambrick, 1996 in Nielsen, 2010: 305). It is further argued that organizational outcome is a “shared activity” and for interactions, capabilities, and for its influence on strategic behavior it is important to focus on the entire TMT. Therefore the focus within this research will be on the entire TMT. The Upper Echelon reasoning further is based on the assumption that individual effects are equally divided among the TMT (Nielsen, 2010: 305).
If looking at the empirical definition of a TMT which varies a lot among previous research on Upper Echelon. The first researches mainly identified the TMT as executives that also serve on the board of directors (Finkelstein & Hambrick, 1990; Haleblian & Finkelstein, 1993; Norburn, 1989 in Carpenter et al., 2004:743). More recent research defined the TMT a as a much broader “team” which derives from executives holding senior most offices, to all the executives listed. Carpenter et al. (2004: 753) argued that TMT membership often is defined by looking at senior hierarchical level, which is indicated by title or position. However it is argued by Carpenter et al. (2004, 753) that the common logic needs to be:
7
30 “when individuals are at the strategic level within the firm, they are likely to influence the strategic outcome of interest to the Upper Echelon Researcher” .
Because the sample is focusing on different TMT´s from different countries, it is important to take a closer look at which TMT members (or executive members) have an influence on the strategic outcomes of the firm. Therefore a closer look at the different countries has to be taken “who constitutes the TMT?” Who are likely to affect strategic outcomes, and are responsible for day-to-day management operations.
Germany
Germany has as described before within this research, a corporate governance system that can be characterized as a two-tier system: with a clear distinction between the management tasks of the executive board and the supervisory and monitoring function of the supervisory board (Van Veen & Elbertsen, 2008: 388). The management board is responsible for the day-to-day operations. Therefore the focus of a TMT within Germany will consist out of the whole management board (Heiltjes et al., 2003: 93).
Japan
31 Japanese boards can be seen as an operations board, because of the amount of insiders, and because the board members maintain their managerial and operational responsibilities. The focus of a Japanese TMT therefore will be on the executives that also serve on the board.
The US and the UK
As described before within the Anglo-Saxon model, there is a one-tier structure with a distinction between executives and non-executives. Decision management and control have a not such a separation as the Coordinated market economy (as in Germany). However the boards of companies in the US serve as a monitoring tool for the shareholders and are consisting mainly out of outsiders (Crossland and Hambrick, 2007: 775), which is different as argued by Osi (2009) from the board of Japanese companies. Therefore to measure diversity and the focus on the operations board within this research, I will follow the definition which is used in most recent researches (Carpenter et al., 2004: 753). The focus within the UK as well as the US will be on the executive team or “executive officers” as stated in the annual reports of the companies, because they can be seen as the managers who are responsible for the company’s management and the operation of the company. If there are more than 20 members serving as executives in the listings of the company, the sample of a TMT was limited to the level of senior vice president and above, which is commonly used in various researches (Carpenter et al., 2004: 756).
32
4.3.2 Dependent variable
Firm performance
Different measures can be used for firm performance. Such as return on equity (ROE), return on investment (ROI) or return on sales (ROS). However according to previous research of Ho Chao & Kumar (2010), ROS or ROA showed a high correlation and generated similar findings in previous Upper Echelon research. In a research of Yang Wen & Chen (2011) they state that ROA and ROE often are used in previous research to measure financial performance. In research of Boone & Hendriks (2009) ROS on ROA are used, both show almost similar results. However it is also argued that ROS indicates short run performance (Boone & Hendriks, 2009).
To improve reliability of the results, two measures can be used for a robustness check. Firm performance can be collected over several years to improve reliability and not only capture short term performance. A lot of previous research examining TMT characteristics relies on ROA, however Certo et al. (2006: 834) are encouraging future research to also focus on alternative firm performance measures such as ROE or ROI.
Because of limitations of this thesis it was not possible to gather data over several years, performance data was gathered over the year 2010. ROA is chosen because it is commonly used in previous Upper Echelon research, and ROE will be used for a robustness check, to improve reliability of the results.
4.3.3 Independent variables
TMT diversity
According to the Upper Echelons t
performance is a reflection of TMT members values and cognitive abilities, and however in reality this is difficult to measure, most
(Priem et al., 1999 in Yang Wen & Chen, 2011).
The focus in this research will therefore also be on demographics to measure diversity. improve reliability, TMT diversity will consist of di
and also nationality diversity argued to be lacking in previous research described in the literature review.
TMT age diversity: The age of each individual team member within 2010.
TMT gender diversity: Male or female of each TMT member
TMT nationality diversity:
country of most education was chosen, if that was unable to be found then the country which the TMT member had worked most
To measure the demographic divers
The Blau index is widely used to measure heterogeneity, (Allison: 1978, Finkelstein & Hambrick, 1996 in Carpenter,
formula:
where Pi is the percentage of individuals in the i nationality diversity is ranging from 0
For age diversity the coefficient of variation is used, a higher value indicates a greater diversity on age (Carpenter,
ccording to the Upper Echelons theory as described above which state that organizational performance is a reflection of TMT members values and cognitive abilities, and however in reality this is difficult to measure, most empirical research focuses on demographics
, 1999 in Yang Wen & Chen, 2011).
will therefore also be on demographics to measure diversity. TMT diversity will consist of different variables inc
also nationality diversity argued to be lacking in previous research described in the literature review.
The age of each individual team member within 2010. Male or female of each TMT member.
TMT nationality diversity: Country of birth was chosen, if could not be found then the country of most education was chosen, if that was unable to be found then the country which the TMT member had worked most was chosen.
To measure the demographic diversity of each TMT the Blau index will be used.
The Blau index is widely used to measure heterogeneity, (Allison: 1978, Finkelstein & in Carpenter, 2002: 8) this index can be calculated with the following
centage of individuals in the i-th category. The Blau index for nationality diversity is ranging from 0-1 , a higher value indicating a more diverse TMT.
the coefficient of variation is used, a higher value indicates a greater (Carpenter, 2002: 8)
33 heory as described above which state that organizational performance is a reflection of TMT members values and cognitive abilities, and however empirical research focuses on demographics
will therefore also be on demographics to measure diversity. To fferent variables including age, gender, also nationality diversity argued to be lacking in previous research, this also has been
The age of each individual team member within 2010.
Country of birth was chosen, if could not be found then the country of most education was chosen, if that was unable to be found then the country
ity of each TMT the Blau index will be used.
The Blau index is widely used to measure heterogeneity, (Allison: 1978, Finkelstein & this index can be calculated with the following
34
4.3.4 Moderator
Country
To test moderating influence, the companies are selected from the different countries as described within the literature review.
To be able to test this, moderator variables need to be constructed. Because the moderator variable is a categorical variable, dummy variables as well as product variables need to be constructed (Voeten & Bercken, 2004: 66 ). The amount of dummy variables is equal to K-1 where K is the total amount of categories. Because there are companies from 4 different countries, three dummy variables will be constructed: 1 = Germany, 2= Japan, 3= US and 4= UK. The reference category will be the US.
To test the moderating influence of country, also product variables will be constructed. For a better interpretation of the moderating effects, the diversity variables also will be standardized. A total of 9 product variables will be used, for nationality diversity: 3 dummies that represent the moderating effect of country on the relationship between nationality diversity on performance. 3 product variables that represent the moderating effect of country and gender diversity on performance, and at last 3 product variables that represent the interaction effects of country and age diversity on performance. The product variables will be entered at last in the regression model (Voeten & Bercken: 2004, 72).
4.3.5 Control variables
Firm size
35 Firm size is also argued to affect the relationship between executive characteristics and organizational outcomes (Miller, 1991 in Carpenter, 2002: 9), and will be included as a control factor. In the analysis, firm size is calculated as the logarithmic transformation of the total number of employees in the company workforce. This is argued to give identical results as total sales or total assets (Carpenter, 2002: 9), and therefore was chosen.
Industry
To prevent industry effects on firm financial performance, the sample is divided into different industries. The categorization is based on the data from ORBIS, and the NACE8 code which determines the main activities of the company (and the major sector as published in ORBIS). Based on this information the following industries have been categorized.
1. Machinery, equipment, furniture, recycling 2. Food, beverages, tobacco
3. Other services
4. Metals & metal products 5. Gas, Water, Electricity 6. Wholesale & retail trade 7. Transport
8. Construction
9. Chemicals, rubber, plastics, non-metallic products
Because the sample size has as a result, that the categorization of industries leads to small samples of different industries, the whole sample is also divided and categorized into 1. Manufacturing companies and 2. Non- manufacturing companies.
Dummy variables will be entered in the regression, to account for the effects of industry (Carpenter, 2002: 10). The categorization of manufacturing and non-manufacturing companies first will be used to test the hypotheses, and the categorization into 9 industries will be used for a robustness check.
8
36
TMT size
Will be measured using the total number of TMT members. TMT size is important because it is argued that most previous studies when measuring the effects of TMT heterogeneity on firm performance do not control for TMT size.
Carpenter et al. (2004, 771) argue that heterogeneity is positively related to group size, and state that “if a research does not control for TMT size, then this is argued to be unable to state whether a significant result can be attributed to heterogeneity itself, or to the effects of the TMT size”. Next to the effect on firm performance, it is also expected that with a larger TMT the heterogeneity increases. Therefore TMT size needs to be used as a control factor. And TMT size can according to Carpenter (2002: 10) be controlled for by the total number of executives within the TMT.
Country
37
4.4 Descriptives
4.4.1 Companies included in the sample
The total sample is consisting out of 80 companies. Banks and insurance companies were excluded from the sample because of different measurement effects of performance. The whole sample contains information about 730 managers. And is consisting out of 22 German companies, 22 Japanese companies and 22 companies from the United states, the sample of the UK was limited to 14 companies because of excluding banks and insurance companies from the list of the fortune Global top 500. An overview of the companies can be seen in the appendix, table A1.
If looking at the different industries most companies are operating in the “machinery, equipment, furniture and recycling” these constitute 39.2 % of all companies. When dividing the sample by manufacturing and non-manufacturing companies, 77.5% of the companies are manufacturing companies. An overview can be seen in appendix table A2 and A3. Looking at the performance indicators of ROA and ROE. It can be seen that there are large differences among the sample, ranging from low values to (high) positive values. The Mean company size is 11.26 indicating a mean of about 77650 employees.
4.4.2 TMT´s
Looking at the proportion of foreigners across the companies within the sample, within a TMT the mean is 18.74%, this indicates that on average 18.74% of all the TMT members are foreigners. The mean percentage of females within a TMT is 6.67%, and the mean age of a TMT is 54.75 years old, and this is ranging from a mean of 46.67 years old to 66 years old within the TMT.
By looking at Nationality diversity the whole sample has a mean of 0.25 with the lowest TMT having no diversity meaning having no foreigners, and the highest 0.78 indicating that there is much diversity, and having a lot different nationalities within a TMT.
38 Age diversity has a mean of 5.08 ranging from 0.76 indicating that there is not much diversity, to 9.97 meaning there is high diversity (compared to the other companies) on age within a TMT. The mean size of a TMT is 9.125 members, the smallest TMT is consisting out of 3 members and the largest out of 22 members. An overview can be seen in Table 3.
Table 3: Descriptive Statistics of the sample
N Minimum Maximum Mean Std. Deviation
Firm Performance : ROE 77 -14.32 206.10 18.97 29.40
Firm Performance : ROA 77 -3.39 22.27 4.92 4.80
Manufactoring or Non-manufactoring 80 1.00 2.00 1.23 .42
Industry (labelled) 79 1 9 3.97 3.06
Size of TMT (Total) 80 3.00 22.00 9.13 4.21
Company Size (LOG Employees) 78 4.61 13.11 11.26 1.22
Nationality Diversity 79 .00 .78 .25 .26
Age Diversity 80 .76 9.97 5.08 1.98
Gender Diversity 80 .00 .49 .106 .14
Mean age TMT 2010 (Yrs) 79 46.67 66.00 54.75 4.40
Females in TMT (in %) 80 .00 .43 .07 .097
Foreigners in TMT (in %) 79 .00 1.00 .19 .22
Country of Company 80 1.00 4.00 2.35 1.07
Valid N (listwise) 74
4.4.3 TMT´s- diversity differences for the US, UK, Germany and Japan
39 It also can be derived from Table 4 that Japan has a mean of 4.42 and the US with 5.16 indicating more homogeneity of Japan compared to the US. An argument of Hofstede9 that within Japan it is difficult for women to “climb up the corporate ladder”, seems to be reflected within gender diversity, because Japan scored very low (0.009) on gender diversity. Indicating that almost no females can be found within their TMT’s.
Overall it can be concluded based on the descriptives of the data, that the data is reflecting differences among TMT’s from different countries, this is in accordance with findings in previous research. An overview of the country differences within the sample by looking at its mean score on the different diversity measures can be seen within Table 4.
Table 4: TMT Diversity mean statistics per Country
Country of Company N Mean Std.
Deviation Std. Error Mean Germany Nationality Diversity 22 .273 .199 .042 Gender Diversity* 22 .028 .092 .020 Age Diverstity 22 5.830 2.046 .436 Japan Nationality Diversity** 22 .0782 .172 .037 Gender Diversity* 22 .009 .042 .009 Age Diverstity 22 4.429 2.281 .486 United States Nationality Diversity 21 .306 .302 .066 Gender Diversity 22 .205 .143 .030 Age Diverstity 22 5.166 1.335 .285 United Kingdom Nationality Diversity 14 .424 .267 .071 Gender Diversity 14 .228 .138 .037 Age Diverstity 14 4.797 2.021 .540
** Japanese TMT´s (mean) Nationality Diversity significantly different from Germany, US, UK. (appendix: B). * Japanese and German TMT´s , (mean) Gender Diversity significantly different from UK and US (appendix B).
9
40
4.5 Distribution of the data
It is important to determine whether the data is normally distributed because this has consequences for the correlation tests. First Skewness and Kurtosis of the variables were analyzed. A Skewness or Kurtosis of 0 indicates a normal distribution of the data, the further away from 0 the more non-normal the distribution. Most variables show a non-normal distribution (appendix table: A5). However because it is difficult to say how much from zero a normal distribution is accepted, further testing is needed. To test this the Shapiro-Wilk test or the Kolmogoriv-Smirnov test (appendix table: A6) can be used. From the outcomes of this test it can be seen that age diversity is normally distributed because (sig: > 0.05) and also log of employees (sig: > 0.05). The others variables as nationality diversity and gender diversity, as well as the dependent variables as ROE and ROA (which are most important because they are the dependent variables), and the control factors as TMT size, industry, and country of origin are not normally distributed according to the Shapiro-Wilk test as well as the Kolmogorov-Smirnov test. The Q-Q plot (appendix figure A6-A9) also shows this graphically for ROE and ROA, because the dots are not centered along the black line. This has consequences for the correlation tests.
Because most variables are not normally distributed a non-parametric test hast to be taken, the non-parametric test of correlation is the Spearman rho10. This test makes it possible to identify whether there is an association between the different variables and whether this is a positive or negative association. This correlation matrix is presented in Table 5.
10
41
Table 5: Correlation between the variables (Spearman RHO)
1 2 3 4 5 6 7 8 9 10 11 12 1.ROE X 2. ROA 0.833** X 3.Manufactoring -Non manufacturing .012 -.014 X 4. Nationality- Diversity .219 .278* .030 X 5. Gender- Diversity .466** .499** .077 .293** X 6. Age- Diversity -.081 -.057 .055 -.018 -.045 X 7. Log employees .022 -.072 -.147 .041 -.076 -.017 X 8.Germany -.144 -.266* .070 .081 -.357** .218 .157 X 9.Japan -.457** -.414** -.131 -.438** -.438** -.245* .081 -.379** X 10.UK .350** .354** .146 .293** .391** -.044 -.211 -.284* -.284* X 11.Size of TMT .144 .288* -.205 .119 .347** .016 .017 -.471** .005 -.034 X 12. Industry Labelled -.017 .079 .261* .139 .157 .259* -.407** .150 -.348** .167 -.064 X
**Correlation is significant at 0.01 level two tailed * Correlation is significant at 0.05 level two tailed
42 (Carpenter et al., 2004: 770). Further age diversity and log of employees are both significantly related to industry, which could indicate differences within age, and amount of employees between the industries. Gender diversity is also positively related to nationality diversity which could indicate that if nationality diversity increases this also positively affects the gender diversity. And both of these variables showed a significant positive relation with ROA. The different countries using dummy variables showed significant correlations with performance measured by ROE, and except for Germany also on ROA. Japan and Germany having a negative correlation with performance and the UK a positive correlation.
43
Chapter 5: Empirical results
5.1 Basic resultsWithin this chapter the hypotheses will be tested, industry will be included as manufacturing or non-manufacturing, and firm performance is measured by ROA.
5.1.1 Hypothesis 1: There is a positive significant relationship between TMT diversity and firm performance
It is expected that there is a positive significant relationship between TMT diversity and firm performance, and this is measured by age, gender or nationality diversity. Regression analysis was used to test the first hypotheses, 8 models were defined and will be tested. The results are presented within Table 6.
Table 6: Dependent variable ROA, Regression Results
44 Table 6 Continued Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 UK - -,843 (1,549) - -1,062 (1,524) - -1,514 (1,564) - -1,534 (1,572) Adjusted R-square .111 0.276 .002 .294 .102 .302 .162 .162 R-Square .122 .343 0.015 .359 .114 .367 .195 .386 F 10.466 5.137 1.171 5.529 9.527 5.626 5.824** 4.611** Significance Of model: ANOVA .002** .000** 0. 283 .000** 0.003** 0.00** .001** .000**
**Correlation is significant at 0.01 level * Correlation is significant at 0.05 level Unstandardized B reported,
Standard errors are in Parentheses
By looking at the regression results, it can be concluded that in model 1 there is a significant positive relationship between gender diversity and firm performance (sig: 0.01), however this becomes non-significant and also negative after adding the control variables in model 2. Germany and Japan as control factors both showed a significant negative correlation with firm performance.
If looking at model 3 no significant results were found, however age diversity showed a negative relationship. Within model 4 the control variables Japan and Germany both remained negatively significantly related to firm performance, as well as a negative relationship for age diversity with firm performance.
45 Overall age diversity seems to be negatively related to firm performance, and nationality diversity positively related, however after control variables were entered within the models the results become non-significant. For gender diversity without control variables there is a positive significant relationship, however after control factors were entered into the model, this relationship becomes non-significant and negative.
Therefore it can be concluded that because if control factors were entered, no significant (positive) relationships were found, that there is not enough evidence to accept the hypothesis 1A, 1B and 1C. That there is a positive relationship between TMT diversity and firm performance. I will elaborate on, and will discuss the results in more detail after the robustness checks in Paragraph 5.5.
5.1.2 Hypothesis 2: The relationship between TMT diversity and firm performance is moderated by country
The regression results will be related to model 2,4 and 6 in Hypothesis 1 (Table 6) that included the control factors.
Table 7: Dependent variable ROA, Regression results
Model 9 Model 10 Model 11 Constant 6.330 (4.847) 6.180 (5.086) 7.356 (4.795) Gender- Diversity -1.684 (1.096) - - Age-Diversity - -2.350 (1.385) - Nationality-Diversity - - 1.894 (0.795) Product Age-Diversity Germany - 1.738
(1.641)
- Product Age-Diversity Japan - 2.354
(1.579)
46
Table 7 : Continued Model
9 Model 10 Model 11 Product Age-Diversity UK - 1.266 (1.825) - Product Gender-Diversity Germany 2.170
(1.757)
- -
Product Gender-Diversity Japan -.539 (3.246) - - Product Gender-Diversity UK 2.809 (1.655) - - Product Nationality-Diversity Germany - - -1,566 (1,425) Product Nationality-Diversity Japan - - -3.028
(1.557) Product Nationality-Diversity UK - - - 1.459 (1.345) TMT size 0.057 (0.140) 0.018 (0.138) -0.33 (0.142) Company size 0.121 (0.413) 0.105 (0.430) (0.00) (0.412) Manufacturing 1.376 (1.155) 1.332 (1.154) 1.483 (1.145) Germany -5.907** (1.840) -5.311** (1.604) -5.341** (1.564) Japan -8.715** (2.561) -6.480** (1.378) -6.880** (1.630) UK -2.771 (1.915) -1.292 (1.539) -1.132 (1.723) Adjusted R-square 0.283 0.288 0.314 R-Square 0.377 0.382 0.406 F 4.001 4.072 4.439
Significance of model: ANOVA 0.00** 0.00** 0.00**
Significance of F-CHANGE 0.306 0.503 0.244
