1
University of Amsterdam
The characteristics of board diversity as drivers for higher firm performance.
Amsterdam,
July 2
th2014
Supervisor:
Erik Dirksen MSc
2
ndsupervisor:
Dr. Wendelien van Eerde
Written by:
2 Abstract
Previous studies that address board diversity within the executive board on firm performance came to different outcomes. In this paper, board diversity is measured by 4 main characteristics that together form the independent variable. The influence of the board diversity is compared with indicators of the firm performance. Therefore, the research question is tested: ‘Does more board diversity, by the following characteristics: Gender,
multiculturalism, functional background of the CEO and the Dominance of the CEO, lead to higher firm performance?’. The dataset constructed to test this RQ, consisted of 60 US based
firms active in 5 different sectors. The timespan that was used is from 2011 to 2013 in order to ensure that the results in this paper are up to date, and therefore applicable for practice. The findings for the strategic decision on board diversity however did not contain convincing evidence to support that more board diversity in the executive board leads to higher firm performance. The results show on the other hand that some assumptions that are often presented as important implications for managers, are not based on empirical evidence.
Keywords: Board diversity, gender, multiculturalism, functional background, dominant leader.
3 Table of contents
CH1: INTRODUCTION ... ERROR! BOOKMARK NOT DEFINED.
CH2: LITERATURE REVIEW ... 9
THE RELATIONSHIP BETWEEN BOARD DIVERSITY AND FIRM PERFORMANCE ... 9
LINK BETWEEN GENDER DIVERSITY AND FIRM PERFORMANCE. ... 14
LINK BETWEEN NATIONAL BACKGROUND AND FIRM PERFORMANCE... 18
LINK BETWEEN FUNCTIONAL BACKGROUND AND FIRM PERFORMANCE ... 20
LINK BETWEEN DOMINANT LEADER AND FIRM PERFORMANCE ... 22
CH3:METHODOLOGY ... 24 RESEARCH DESIGN ... 24 DATA SET ... 26 DATA COLLECTING ... 27 DEPENDENT VARIABLE ... 27 INDEPENDENT VARIABLES ... 299 CONTROL VARIABLES ... 31 MISSING DATA ... 32 CH4: RESULTS ... 333
THE DIFFERENT HYPOTHESIS MEANS ... 35
NORMALITY ANALYSIS ... 4039
CORRELATIONS ... 40
4
OVERVIEW OF THE REGRESSION MODELS ... 43
HYPOTHESES TESTING ... 44 CONTROL TEST ... 44 CH5: DISCUSSION ... 49 LIMITATIONS ... 53 FURTHER RESEARCH ... 54 CH6: CONCLUSION ... 555 LIST OF REFERENCES ... 56 APPENDIX ... 63
5 Chapter 1: Introduction
In the last decades companies have become more international, and are seeking for new markets and opportunities. The globalization so called means that firms are more than ever expanding outside the home market into foreign territories. With the ongoing technological advancements the world has become smaller and foreign destinations easier to reach. Contributed by this shift in market seeking new discussion emerged about the functioning of executive boards, and whether these boards should be adapted to the changing environment. This makes that much research has been done about which , the board diversity debate has been widely researched, and many components of which diversity characteristics need to be implemented to make boards more efficient, and create higher firm performance (Hambrick and Mason, 1984).
The relationship between the firm performance and the board diversity is researched by many different angles, however empirical studies have found contradicting outcomes on the existence of a direct relationship. In the past much empirical research was done to investigate the direct link between the diversity in the board and the performance of the board. The testing of this direct link between the board diversity and performance is complex and there is not a commonly accepted direct link (Kochan et al, 2003). Research exists that shows a positive link between both variables. Diverse teams can outperform teams that are more homogeneous, because they have more different skills and experiences that is contributed to the decision making process (Cox, Lobel,& Mcleod, 1991). However, there are also researches that came to the opposite conclusion that more homogeneous teams outperform more diverse teams, because the lack of communication and stereotypes that can disrupt the decision making process. (O’Reilly, Caldwell, & Barnett, 1989) ( Pelled, 1996). Nowadays in research if diverse teams are more successful on performance, we speak about the conditions
6 that are needed to make a diverse team successful. Stereotypes have to be overcome for a diverse team to outperform an homogeneous team, however, in most of the cases the stereotypes are making it harder to create a learning environment (Ely, Robin, et all, 2012). Other research argues that a more diverse team can outperform in decision making and social cohesion.
Nevertheless, several theories designed frameworks to optimize the relationship between board diversity and firm performance. This is based on the fact that the relationship is present, and these theories are widely used in both the academic world as well as in the business world. The literature in support of board diversity is arguing that more resources are present for the decision making process. The Empirical research data for the relationship between the diversity in boards (both board of directors and top management team) and the firm performance is very limited. Board diversity is mostly investigated by either the demographic or the cognitive characteristics, while a combination is rare (Milliken and Martins, 1996). This research will combine the characteristics and create an economic framework by which all different aspects are included, while there is no empirical evidence that suggests that all the elements together create more resources, and as a consequence a higher firm performance. Among these theories about the top management team (TMT) diversity are the agency theory and the resource based theory. The agency theory looks at how the rights and interests of the shareholders can be protected, and to find the best structure to do so (Van de Walt and Ingley, 2003). Close ties between the boards are more likely to be less when both boards are very diverse (Carter et al., 2003). This will enhance the performance of the executive board (Van de Walt and Ingley, 2003). Another theory that argues that more diverse boards should perform better is the resource dependence view. This view sees the board as a pool of resources, therefore, a more diverse board should come to more diverse resources (Van der Walt and Ingley, 2003). However it is still unclear when these theories are effective, under
7 which circumstances board diversity is needed and when it does not have a significant contribution towards firm performance. Therefore, in this research the markets and branches where board diversity is effective are investigated, to construct a framework where a strategic decision upon board diversity can be made. A framework is constructed in order to analyse the characteristics of board diversity that could have an influence on the firm performance. The data for this characteristics will be taken from business sources, so that the framework can be used in the business world, and a clear overview is given in which cases board diversity is effective to boost the firm performance.
Board diversity can constrain several characteristics that create more diversity or homogeneity among the board members. While the executive board can influence directly the internal aspects of the firm performance, the outside factors that influence firm performance are harder to link to board diversity. So the four characteristics that are widely used to create more board diversity in the TMT will be discussed in relationship with the firm performance. The relationship between characteristics of board diversity and firm performance are all separately tested many times before, with different outcomes. However the value of the characteristics within the board diversity is never tested before. Furthermore, the relationships between the characteristics and the explanation of these significances with firm performance is not tested in a broader perspective. The goal of the research is to see which board diversity characteristic has the biggest influence on firm performance, when there is an influence. Also a more specific framework for which companies this kind of board diversity can be applied is investigated. Therefore existing meta-analyses about the four characteristics are used so the framework can support the theories to be implemented more effectively. Where the grounds of this research are solely based on the economic implications of implementing more board diversity and the strategic choices based on these grounds. Therefore the research question and the main hypothesis of this thesis are:
8
RQ: Does more board diversity, by the following characteristics: Gender, Multiculturalism, Functional background of the CEO and the Dominance of the CEO, lead to higher firm performance?
9 Chapter 2: Literature Review
In this chapter the literature background of the relationships being investigated will be discussed, to base the expectations on together with defining the hypotheses to test the research question. First the relationship of the combined independent variable, and the dependent variable will be discussed. Afterwards the existing theory of the several indicators that together create the independent variable, will be compared with the dependent variable.
The relationship between Board diversity and Firm performance
The term board diversity is used in many different studies and indicates that there is a distinct difference between the members, that creates a dynamic group composition (Harrison and Klein, 2007). The diverse members all have different resources to contribute to the decision making process, and to create an atmosphere where these resources can be used to the fullest. So therefore, more diversity creates that more different views and resources can be used and should enhance the decision making process (Brammer et al, 2009). There are many different ways to create more board diversity, more gender diversity and national diversity are among the most studied ones. However also the role of the CEO in the composition of the board can’t be ignored, while as the leader he has a great influence on the dynamics of the board. This makes that the functional background and the dominancy of the CEO has a large effect on the focus of the board of executives, and so on the diversity of the board. In the globalising and fast moving world of today the reputation of a firm is essential, and enhances the pressure to create more diversity in the boards (Kramer et al, 2006). However in which way the characteristics of board diversity boost performance, by using more different resources, or by a higher perceived reputation is still very unclear, and not investigated extensively. Another
10 aspect is that most of the literature on board diversity is using an American sample, however the boards researched in these samples are mostly the board of directors. The Top Management team is less investigated, even when this board is in charge of the daily operations. While the corporate decisions are of great importance for the future of the company, the board of executives is chosen as the sample for this research. Monitoring the company is essential, however the daily operations managers create the board culture, and are responsible for the strategy, with the CEO as the biggest influencer.
The debate on board diversity can be split into two different views, one is looking at the economic benefits of the diversity debate, the other view looks at the ethical issue of having more different people at powerful positions (Brammer, 2007). The ethical argument for more board diversity is that everyone should have equal chances and be judged on their performance rather than on other grounds, like gender, nationality. Moreover it argues that the people in power rather choose someone from the same social cultural status regardless of the capabilities of the other candidate (Brammer, 2007). The idea that the (business) world should not be ruled by a small group of people from the same cultural and ethnical background, is seen by me as valid, this will not be the background of this research. When the economic reasons show that more diversity is beneficial, or even when it shows it is not harmful on performance, this will be a strong trigger to implement more diversity in boards. It will show that everyone from different backgrounds is suitable to lead, and bring new resources to the board room. So the focus on this investigation will be solely on economic grounds, and see by which strategic decisions on diversity the highest performance can be realised. The economic effect of not choosing the most suited person for an executive board member, can have negative effect on the firm performance (Burke, 1997; Carver, 2002). Essential potential resources in the decision making process can be missed with this approach what can cause a financial decline (burke, 2000). By creating a more demographic, gender and functional
11 background diverse team, a more dynamic and resourceful group can be constructed that can use a more panoramic view on the issue at hand. Also the network of the executive team can be enhanced by implementing people from another background (Pfeffer and Salancik, 1978). Therefore, the strategic decision on the composition of the board of executives can be as important as the strategy of the firm in general. To create some clarity in the wide web of research about board diversity, this investigation will create a framework, on what economic grounds this strategic decision can be made.
Board diversity as a resource or a threat for communication, the opinions are not in-line with each other. A board is effective when it creates strategic resources for the firms decision making process, and business challenges (Boyd, 1990). As the Upper echelons theory says that the executive team has a great effect on the strategic choices and therefore on the firm performance (Hambrick and Mason, 1984). The human capital is based on the different background and experiences of the members of the executive board, containing the specialist background, communication skills, and the personal network. More diverse management teams, can have more extensive knowledge in a broader area what can help to communicate more precise with the external stakeholders (Hillman et al., 2000). Board diversity is also important for the stakeholders to show that the top management team is taking every aspect of their interest into account (Gong, 2006). However, these teams still underperform in doing business with the local clients and fundraisers. Where fundraisers and clients are more in the same social and cultural class as the more homogeneous boards, they can outperform the diverse teams. (Siciliano, 1996) All together the link between board diversity and performance is a link that is influenced by many constraints and conditions.
Previous studies have shown that functional diversity can create more creativity in teams, to come to more innovative solutions, with a wider overview of the problem at hand (Joshi & Roh, 2009). This diversity provides a more extensive look at all constrains that effect
12 decisions, to get more strategic options. The diverse knowledge pool will help to focus on every aspect of the organisation, and the communication with the department the CEO is familiar with can be more fluent, because the executive member has extensive knowledge of how this department works (McLeod, Lobel & Cox, 1996, Hambrick and Mason, 1984). The external communication can be more specific when the CEO has to provide more detailed information. Furthermore, the specialised functional background make it more likely that the CEO more easily create more close tights with the external party.Furthermore, the expert has more authority in the perception of the audience, when experience in this field is present. These community influentials have to personal connections with the community, which creates a wider knowledge about the environment the firm is operating in (Hillman et al., 2000).
The diversity of the executive board can also influence the relationship with the board of directors, and the shareholders of the firm. The agency relationship between the top management of the firm and the owners of the firm is highly important for the success of the company, with clear communication as the key element. While the executive board can influence directly the internal aspects of the firm performance, the outside factors that influence firm performance are harder to link to board diversity (Hambrick and Mason, 1984). In order for the owners to keep control over the firm and reduce risks, strategic decision about the executive board have to be taken, therefore a board of directors is implemented to sustain the interests of the shareholders and other stakeholders. Whether executive board diversity helps the interest is still a highly debated topic and there are different ways to look at this relationship. The Agency theory focusses on the relationship between the executive board and the shareholders, to see how the rights and interests of the shareholders can be protected, and to find the best structure to do so (Van de Walt and Ingley, 2003). The balance between the different parties has to be found to take care of all interests. In order to create an objective
13 monitor system the executive and non-executive board should be different in diversity so both boards look with different views at the problem. Also the close ties between the boards are more likely to be less when both boards are very diverse (Carter et al., 2003). This should enhance the interests of the shareholders and diminish the self-interests of the executive board. The Agency theory therefore expects that the overall performance of the board therefore rises (Van de Walt and Ingley, 2003) While the agency theory is often used to monitor the board of directors, a positive relation between board diversity and board performance is not found with the agency theory (Jensen, 1993) Even when a positive relation is expected with the agency theory, empirical evidence has not been found (Carter et al., 2000). Another theory that talks about the relationship between board diversity and board performance is the resource dependence view (Van de Walt and Ingley, 2003). This view argues that the board of directors is a pool of resources and every member brings an unique skill (Hilman et al., 2000). When more different valuable resources are in the board of directors this should make the board more effective. When there are board members with different cultural or social backgrounds, they view a situation differently, what creates valuable information (Van der Walt and Ingley, 2003). So therefore in theory there should be a positive relationship between board diversity and performance. However, the diversity benefit is only present when the communication in the more diverse board is of the same level as the homogeneous board (Carter et al,. 2010). Consequently, this view expects that a positive relationship exists, however this is not supported by empirical research.
The Empirical research data for the relationship between the diversity in boards (both board of directors and top management team) and the firm performance is very limited. Board diversity is mostly investigated by either the demographic or the cognitive characteristics, while a combination is rare (Milliken and Martins, 1996). This research will combine the characteristics and create an economic framework by which all different aspects are included.
14 At the moment there is no empirical evidence that suggests that all the elements together create more resources, and as a consequence a higher firm performance. This paper will construct a framework to combine these characteristics and to see which are relevant.
15 Link between gender diversity and firm performance.
One of the elements to create board diversity is by making a greater mix between genders in the executive boards. So to create more resources in the executive boards the number of females should be increase, while in most firms more males are present in the top management functions. Therefore, if the resource theory is correct, more women in the management team making corporate decisions will have a positive effect on the firm performance. The gender diversity creates a wider spectrum of perspectives, which should help to come to a wider range of corporate strategic decision possibilities. Also the communication skills of females and males tend to be different, while women are involving other voices more easily, instead of trying to solve the problem themselves (Eagly et al., 2003) This other leadership style creates a more open culture where more opinions are taken into consideration. This does create that not only the more diverse views by the difference in gender, also the amount of people involved in the decision making process becomes larger (Rudman and Glick, 2001). Another important factor is that this bigger group, has more chance to understand the environment together with the wishes of the stakeholders. However, the involvement of many persons in the decision making process can cause that this takes much longer, what has a negative effect on firm performance. Moreover, the clarity of the decision has to be preserved, otherwise the implementation together with the willingness to change can make that problems arise. The support for the CEO can be larger within the firm when the members in the top management functions are demographically more like the leader to speed up the process. This however can also be seen as a negative influence while more gender diversity, should create more secure decisions (Westphal & Zajac., 1995).
The employment of more women can have a positive effect, because the females that reach the top of business are more often higher educated than their male colleagues (Hillman et al., 2002). This would suggest that the females that reach the executive board are more
16 specialised in their field of work. Furthermore, the focus of females is in general less solely on the short term financial benefits but they also are more involved with the wishes of all stakeholders involved. Community influentials and specialists are more likely to be appointed and involved in the corporate decision making process when women are in charge (Hillman et al., 2002). Males are in general more focussed on competition, which can create a more efficient working environment, however, the atmosphere can be negatively influenced by the competition and decrease effective communication.
The effectiveness of hiring a female does not immediately have to affect the board diversity advantages. When there are more women in top functions of the firm, this effect seems to increase, while several females have a greater effect on the board culture. The presence of a female can be seen by the others as a necessity for the outside world to reflect positively on the reputation of the company. This may influence the authority of the female in the group and giving the individual a lower status, what makes it harder to express a different sound in the executive board (Brewer and Kramer, 1985). So when the gender diversity is greater the chance that these voices are heard is higher, with as a consequence that more different views are used as a resource, to enhance decision making and boost firm performance (Konrad et al., 2008). On the other side, it can also be that the majority of the top management positions is filled with females, and that more diversity could be created by adding more males to the executive group. However this is in the sample being researched in this paper, the Fortune 500 not even once the case. Even more when 1 out of 4 executives is female, this is already seen as extremely high on gender diversity (Catalyst census, 2013)
Also the presence of females in the firm and displacing this face, is helping the reputation of the firm. In a research it was proven that companies with more women in the board of directors were more often in the “World’s most Ethical companies” list (Bernardi et al., 2009). While this is not about the executive board, it shows a tendency in public perception
17 about gender diversity, whether this is based on ethical or economic reasons. Also it appears that companies indirectly advertise with gender diversity, while it would show that the firm does focus on all groups in society, and is socially responsible. This image seems to create higher firm reputation, what consequently should result in a rise in firm performance as well (Bernardi et al., 2002).
In general the existing literature seems to be in favour of more gender diversity, because it gives a wider range of resources to the executive board. Empirical studies on the other hand are reaching different conclusions, on the one hand there are studies that argue that there is a positive effect of gender board diversity on firm performance (Carter et all., 2003, Campbell & Minguez-Vera, 2008). However, there are also studies that did not find any relationship between the variables (Farrell & Hersch, 2005, Rose, 2007) and studies that found a negative effect of adding more females (Smith & Verner, 2006, Adams & Ferreira, 2008). So whether implementing more females in the executive boards boosts the firm performance is very unclear. Nevertheless, contributed by the perception of the changing environment where firm reputation together with an extensive knowledge pool is becoming more important, the expectation is that more females in the top management functions will help a company to boost firm performance. May be it will not be for the short term profit or return of assets, it probably will be seen in the firm value, while having more different resources in the decision making process the risk can be reduced. For these reasons the following hypothesis is created.
H1: “The presence of more females in the executive boards, creates a higher firm
18 Link between national background and firm performance.
Together with the gender diversity, nationality diversity is often taken as an important element to create more board diversity. Also this variable is widely debated in literature with different views on whether it is an effective tool to boost firm performance.
Nationality diversity is mostly perceived as a possible way to create more resources for corporate decisions, to get to higher firm performance (McLeod, Lobel & Cox, 1996, Hambrick and Mason, 1984). More nationalities create heterogeneous teams that have a higher potential to gather more extensive knowledge. Furthermore, identifying a certain problem may be easier with different cultural expertise, especially in an international context (Bantel & Jackson,1989). The presence of an international executive team is seen as human capital for multinational firms, while the knowledge pool to create a strategy for foreign markets is wider (Athanassiou & Nigh, 2000). In order to understand the different circumstances and the institutions in overseas countries are triggers for the employment of international people (Bartlett & Ghoshal, 1989). With a growing amount of nationalities present in top management positions in firms, by the increase of other countries the firm is present in, the board performance will be enhanced by the wider understanding of the preferences of the clients (Erhardt et all., 2003).
However, there are not only positive effects consisting from more diverse national teams, as the “double edged sword” term of Milliken and Martins (1996) describe. The other side is the communication difficulties that can arise which is seen as the reason why more diversity can decrease the efficiency of the decision making process. Due to social assumptions and miss-communication on differences in norms, values, motives the group cohesion can be affected, having a negative effect on the functioning of the top management team (Shaw, 1981). This can cause cultural barriers in the communication, making the heterogeneity within the group a
19 burden, while homogeneous teams understand each other better and faster, creating a more open social atmosphere, making the group interaction more fluent together with ensuring that positive recognition creates easier support for new ideas. Also the cognitive bias that will establish a communicational barrier is prevented, so that the team can be focussed on the problem at hand (Smith et al., 1994).
The academic world is clearly still divided on whether more national diversity will increase firm performance. Research can be found that found a positive relationship between both variables (Cox, 1991; Watson et all., 1993; Townsend et all., 1998), as expected there is also research that argues the opposite, that there is a negative effect (Shaw, 1981; Shenkar & Zeira, 1992) and another investigation found no relationship (Zahra and Stanton, 1988). Several explanations can cause these conflicting results. First the different samples that were used in other countries, moreover, the different indicators for the variable could have created the divergence.
International operating firms have a disadvantage over local firms, where they have less understanding of the foreign economic culture and the institutions. By hiring the people that have this knowledge the multinational firm can overcome this problem (Chang, 1995). Where the national diversity has a greater effect on companies that are present in more different countries, a control variable will be created to overcome this reliability problem. So the implication of nationality diversity is only necessary for truly international firms to understand the foreign markets, or whether it creates a bigger knowledge pool that boosts also the firm performance of more home oriented companies. With having more resources present while taking corporate decisions the expectation is that this creates more effective and innovative solutions. This reflects in the following hypothesis.
H2: The presence of more national diverse backgrounds in the executive boards, creates a
20 Link between functional background and firm performance
The CEO is the main influencer in the top management team culture, where he can stimulate the rest to be open minded. Therefore, the CEO is the composer to create that the group is diverse, and can search in a wide range of knowledge and resources. The expertise of the leader can determine which resources are used, and how the solutions are judged (Lawrence & Lorsch, 1967). Therefore, having a CEO with a more diverse background should create a more diverse group cohesion, while he has a little expertise in every field. The influence of the functional background on the diversity in the mindset of the executive board is mostly split into two different background groups, the CEO generalist and the CEO market specialist (Lawrence & Lorsch, 1967).
The CEO generalist is a leader that has a broad range of experience in different fields, therefore the functional knowledge is wider, however less in-dept. This means that every subject at hand has the interest of the CEO and that he has some experience with the problem, what can enhance the communication with the other executives by shaping a shared understanding, making the functional generalist to be the bonding and central player (De Brabander & Thiers, 1984). Furthermore, the generalist wants to gather specialist around him, while he does not have that specific knowledge. As well as the knowledge sharing, this will be easier for the CEO generalist, because it is not his unique factor to receive the authority. The focus of leadership is more focussed on communication instead of specialist knowledge, what can create a more open culture (Balkundi & Harrison, 2006). Better connecting the resources in the team can enhance the decision making process to boost the firm performance (Arendt et all., 2005). The generalist is also less likely to attract persons only from a specific field, while with the functional specialist this is more often the case. Another problem that can occur is that the specialist creates a functional homogenous group what limits the different
21 resources that can be used in the decision making process, and has negative effect on the firm performance (Schneider et al., 1995).
On the other side there are also advantages of a CEO market specialist that has extensive knowledge about a particular area of the firm. Especially in a highly innovative and dynamic environment where speed is essential an expert in this field as leader can be beneficial (Lawrence & Lorsch, 1967). This is important when the focus of the company is oriented on a specific market/task, so the interests of the consumer can be implemented immediately. In a fast moving industry, like the IT branch, having a specialist market CEO could increase the pace in which the products reach the market (Cooper and Kleinschmidt, 1994). Also the competitive advantage can become a central focus point by appointing a leader with extensive knowledge of this subject.
To complete the expectation, the literature argues that the CEO functional background has an influence on the board diversity. While the focus of the board diversity is on creating more resources for the decision making process and thereby increasing the firm performance, the CEO generalist is more likely to create such a more diverse executive team (Arendt et all., 2005). In analysing the results the interesting distinction could be made between the It-technical firms and the other firms, to see if there is a difference for highly innovative and dynamic industries (Cooper and Kleinschmidt, 1994). This comes to the following hypothesis:
H3: The presence of a more diverse functional background of the CEO, creates more board
22 Link between the dominant leader and firm performance
After studying the influence of the focus, also the dominance of the CEO has to be discussed. When the leader has more power, the chance is there that more decisions are taken personally, without the involvement of more resources. Consequently the focus can get a dominant effect on the decision making process to make it less creative and innovative (Brabander & Thiers ,1984). So the dominance will be discussed by power that the CEO has within the firm, and therefore creates less executive diversity, while that diminishes the overall influence of the leader. The power is measured by if the CEO is the owner/founder, has a place in both the boards (Top management team and the board of directors) or is only the front man of the executive team. The CEO that is also the founder is more dominant while he has proven during the start and with the success of the company to be capable of leading in this sector. Together with ownership in the firm this is perceived as being the most powerful leader a MNE can have (Adams et al., 2005).
The dominance of the CEO can lead to several problems that can make a board of executives less effective. Especially the founder is more likely to be a specialist in the field the company is operating in, and as discussed in the previous chapter this can lead to a focus on a specific area of interest. With low involvement of other types of people in top management functions, what makes the differences in resources for the decision making process smaller (Schneider et al., 1995). The sources of why a CEO has more power in the organisation are based on specific skills, knowledge or leverage power (ownership or founder), the more dominant the leader is, the harder it gets to change the corporate culture (Finkelstein, 1992). An open culture is needed for the other team members to express new ideas to increase the ability of a firm to be innovative and creative. When more equality in power is present in the TMT, the communication and sharing of knowledge among the team members, happens to be more fluent (De Brabander and Thiers, 1984). So with a more dominant leader the board diversity,
23 is more likely to be small, and even when there is on paper a diverse group, they may have less influence on the decision making process.
Another problem that can occur is that the agency relationship between the share/stakeholders and the top management team can be disturbed. The strong CEO has more independence which makes it harder for the environment and the board of directors to monitor the daily operations within the firm (Van de Walt and Ingley, 2003). In order to create an objective monitor system the executive and non-executive board should be different in diversity so both boards look with different views at the problem. Also the close ties between the boards are more likely to be less when both boards are very diverse ( Carter et al., 2003).
However there are also scenario’s where a dominant leader is outperforming the less powerful CEO’s. In highly uncertain and dynamic market fast decisions can be needed, while moving fast is essential, a strong CEO can help speeding up the decision making process together with the implementation of the ideas. Also when the company is facing difficulties and changes have to be made, this attribute can give an advantage. The communication is done by one person, consequently it is more clear, as well as the pace in which re-organizations can be implemented (Cooper and Kleinschmidt, 1994).
In the end when the environment is not very dynamic or uncertain, the expectation is that a less powerful CEO will have a more diverse board. One, to ensure that the interest of the stake/shareholders are being fulfilled, because of the close monitoring. Secondly, when there is a dominant CEO the rest of the members are reluctant to share knowledge and freely communicate (Edmondson et al., 2003). This creates less resources for the decision making process and consequently can affect the efficiency of the firm. This evolves in the hypothesis:
H4: The presence of a less powerful CEO, creates higher board diversity and therefore more firm performance.
24 Chapter 3: Methodology
While much research has been done on board diversity, the influence of this board diversity on the firm performance is still a highly debated topic, with many different outcomes. Therefore, a clear framework, showing if and when which kind of board diversity is contributing to higher firm performance. This framework based on empirical evidence can be a guide for the business world when to implement board diversity. In order to create this framework, a quantitative research is used to base the results on objective measurements. This concludes to the following research question:
RQ: Does more board diversity, by the following characteristics: Gender, Multiculturalism, Functional background of the CEO and the Dominance of the CEO, lead to higher firm performance?
Research design
In order to test the RQ, a conceptual model has to be constructed where the relationships between the variables are listed and the relationships that interact with each other. First the dependent variable: Board performance, this variable will be measured by two main indicators, the ROA and the Tobin’s q. The influence of the independent variable: Board diversity, on the DV, will be researched, by the combination of the different characteristics. However, also the influence of the different characteristics separately will be taken into account, to specify which board diversity has a positive influence on firm performance. The independent variables: the four characteristics will be measured in a scale of 3 different possibilities of diversity levels, so that the optimal amount of board diversity can be measured. The independent variables will be measured by categorising the percentage of presence of the board diversity into three different groups. Then the firm performance of the
25 different groups will be researched to see whether the board diversity has an influence on the firm performance. To get a detailed insight in when board diversity is beneficial also moderators are installed into the research model, the type of business is chosen as the moderator. The moderator can explain if the influence of an IDV on the DV can be explained by another reason, most likely the performance of the whole branch where is accidently less or more board diversity. This results in the following conceptual model:
Board diversity
(4 characteristics of board diversity)
(gender, multiculturalism, functional background, dominant leader.)
Board performance
(Control variables: different branches, USA oriented.)
The relationships between characteristics of board diversity and firm performance are all separately tested many times before, with different outcomes. However to which extent the different characteristics of board diversity influence firm performance is never tested before. Furthermore, the relationships between the characteristics and the explanation of these significances with firm performance is not tested in a broader perspective. The goal of the research is to see which board diversity characteristic has the biggest influence on firm performance, when there is an influence. Also a more specific framework for which companies this kind of board diversity is important. In the end of this paper a framework will be constructed with the specific sectors where board diversity is more important, and which board diversity characteristic should be implied. With the goal to make an economic based framework what would create that companies have an incentive to implement more diverse executive boards.
26 Data set
As the sample for the research 60 companies from the Fortune 500 were chosen, from 5 different branches. Only the 12 largest companies in every sector are included in this research, the largest companies are chosen on the basis of the indication of Forbes. After picking the sample, the companies will be structured in a specific group based on the branch the firm is operating in, so the sample can be compared. American based companies are chosen in order to make sure the data set is great enough, and the data is accessible in English. The chosen sample is a data set of 60 companies, to create a large sample as a solid research base for the framework, also that the companies can be divided into smaller sub-categories, on the other hand the sample is not too large to collect all the necessary data. In total 6000 indicator data numbers will be collected to construct a solid base to test the different hypotheses. Because of that every company has many different executives of which the background has to be checked, which makes the sample of 60 companies, into a data collection set of 840 persons for every year.
This to explain if the significant relationship when present, between firm performance and board diversity, is not caused by another constraint. The data will be collected for the period from 2011 until 2014, this scope is chosen because the time frame is of such a size that the differences in the presence of board diversity can be evaluated. On the other hand the scope is still manageable to test the relationship between the variables. Where many different indicators are needed in order to investigate the relationship, the scope of the amount of firms had to be limited, to keep the data set manageable and able to precisely collect the relevant data. Furthermore, the data is up to date, and therefore the framework is useful for companies at the presence, and further research can add extra years to be able to follow the trends, and keep the framework relevant. The dataset will contain the characteristics of the executive board members instead of the characteristics of the board of directors. This because the
27 executive board members are in charge of the strategic decision making process, and are the ones that are working on the future of the firm daily. Investigating which leaders are in charge of the corporate strategy will be the challenge, and therefore the information of the characteristics is based on the information the company itself gives on the members of the executive board.
Data collecting
The data that will be gathered to investigate the board diversity will be obtained by Riskmetrics, where the data of the S&P 500 firms can be found. The database Wharton Research Data Service (WRDS) is used to gather the data for the dependent variable, to measure the performance. For the independent variables, and to gather the data for the several board diversity characteristics the annual reports of the companies is checked, together with the official websites to evaluate the data. For every independent variable for the three years being measured, the data will be obtained from the annual reports or the company website. Afterwards the companies data will be inserted in one of the three groups possible. Also the LinkedIn pages of the executives will be used to determine the national background and the previous work history to obtain the data needed to categorise the whole company by the TMT’s.
Dependent variable
There are a few different views on how to measure the dependent variable “firm performance” (Choe, C., G. Tian & X. Yin 2009). First there is the return on assets, what is a theory to calculate board performance. This theory measures how effective the firm is using
28 its total assets, so the (ROA) is a good indicator if the firm is profitable in comparison to all activa (Mclean, et al, 2007). This accounting theory objectively calculates the results of the firm within the time frame. The formula that is used to calculate the ROA:
(Net Income)/(Total assets)
Secondly the Tobin’s q equation will be used (Chung and Pruitt, 1994). The Tobin’s q is a market-based measurement that shows also how the market perceives the performance of a firm. So not only how well the firm performed in the last period, but also what is expected in the near future, so both the market value and the book value is taken into account (Hillman, A.J. & G.D. Keim. 2001). The formula of the Tobin’s q:
(Equity market value + Liabilities book value) / (Equity book value + Liabilities book value)
There are several strengths why these calculations measures are used, first these measurements are mostly implemented in research to show the evaluation of the firm performance. Also both methods are using real world data, with objective book data, that makes the findings of the research more reliable and applicable to the business world. However these theories also create some limitations, firstly the book value and the ROA are looking to the past, while the board diversity influence may has only influence on the future results of the firm, what can cause validity problems (Jacobson, R. 1987). Secondly, controlling the many different constraints that can influence the firm performance, will be a real challenge. Especially the external factors that can change the measurements on the relationships between the variables, will be closely watched (Child, J. 1974). This threat should be controlled by creating groups that are operating in similar sectors.
29 Independent variables
Every independent variable will be analyzed by comparing the performance of the three different groups based on the characteristics values. In which group the company is placed for the characteristics is calculated by several different criteria, that are different for every variable.
The first characteristics Gender, the hypothesis: “The presence of more females in the
executive boards, create a higher firm performance” is tested. Three different groups were
created, on the basis of how many females are in the executive boards. Group 1: Less than 5% females on the board
Group 2: 5-25% females on the board
Group 3: More than 25% females on the board
The percentages that are used to categorize the data is chosen by looking at other studies about gender diversity in boards. The main organ that studies the gender diversity in the United States is the Catalyst USA region organization. The percentages that are used in this research are chosen on the grounds of earlier research by this organization. This created the percentages for the three different groups, while this are the numbers that are most common to test gender diversity (Catalyst Census, 2013). In a big survey this organization came to the conclusion that having less than 5% females in boards can be considered low diversity, and more than 25% as high gender diversity.
The second characteristic is the presence of different national backgrounds within the executive boards. Hypothesis: The presence of more national diverse backgrounds in the
executive boards, create a higher firm performance”. The heterogeneity formula will measure
30 as a cultural background indicator, where both the amount of non-Americans and the amount of different nationalities is taken into account. In the end the average of the two different indicators for the natural diverse background of the executive board will determine in which group the company will be placed.
Group 1: less than 10% different national background Group 2: 10 -25% different national backgrounds
Group 3: More than 25 % different national background.
The third hypothesis of a characteristic that will be tested is: The presence of a more diverse
functional background of the CEO, creates more board diversity and therefore higher firm performance. When a CEO has seen many different aspects of the firm, it is more likely that
the focus is more widespread, and the resources in the executive board can be used more easily, and creates higher board diversity. Finance is seen as being a more specialized background, than the marketing focus (Balkundi & Harrison, 2006). This makes that more board diversity is created, and therefore the order of groups will be the following.
Group 1: Finance Group 2: Marketing
Group 3: Diverse background
As the last hypothesis is: The presence of a less powerful CEO, creates higher board diversity
and therefore more firm performance.. With a powerful CEO, the rest of the executives will
be less involved, with as a result less different views on strategy. According to the resource based view, less different resource leads to lower firm performance. If the leader has more power, the chance that decisions are taken by him personally are more likely, which means that less resources are used in the decisions making process. As a consequence the chosen
31 strategy or decision can be less creative and innovative (Brabander & Thiers ,1984). The dominance will be measured by the power of a CEO within the firm, and therefore creates less executive board diversity.
Group 1: Founder is the CEO Group 2: Powerful leader Group 3: No dominant leader
Control variables
In order to see whether the composition of the board is influencing the firm performance, or that maybe other factors created the success of the firm, the control variables “US oriented” and “Sector active in” will help to explain why there is an effect. Firm size is often chosen as a control variable for investigating board diversity, however in this research only the 12 largest US firms of the 5 largest sectors are included. Therefore, all companies are seen as big firms and therefore the control variable is not included in this paper.
Regression models
To test the relationship between the several independent variables with the dependent variables a model structure is created to measure this connection. First the control variables will be tested in combination with the dependent variable characteristics. Afterwards the influence of the independent variables will be included to see if a relationship exist between the dependent variable and the independent variable. An overview of the models used in this research can be seen in table 7 at page 43.
32 Missing data
In order to create a framework based on the relationship between the dependent variable, and its indicators, together with the dependent variables all data was needed. The smaller size of the scope of the firm mate it possible to collect all the data for the indicators of the independent variables.
For the dependent variable, firm performance, two indicators were created, the indicator Return on assets (ROA) and the indicator the Tobin’s q. To construct these indicators much financial data had to be collected and the missing data had to be analyzed. No missing data was found for calculating the ROA, in neither one of the three years under investigation. There was missing data to calculate the Tobin’s q on the other hand, in the year 2011 one market value of a company was not available, and was deleted, while the data could not be guessed on the basis of the Hotdeck Imputation principle. However this will not be done by the usage of a statistical program, but by making an educated guess, by the personal background of the executive’s history. The missing data for the characteristic, cultural background, is guessed by the University the executive has attended, furthermore, when also this information could not be obtained, the assumption is made that before the year 2013 the composition of the board was the same.
33 Chapter 4: Results
In order to test the hypotheses first the preparation of the data for analysis has to be conducted. All the data shown in them tables, is based on gathered information by own analysis of over 6000 indicator numbers. The data sample exist of 60 companies with a total of 839 executive board members, with 17.5% females in 2011 to 19,35% in 2013. In 2011 15% of the board members had a different nationality than the USA one, in 2013 this number did go up towards 16%, for the functional background of the CEO the statistics did not change over the years, the dominance of the CEO on the other hand decreased. The tables in the following chapter are all derived from the collected data and are personally constructed. In table 1 the structure of the groups can be seen, together with the differences per year. The changes within the groups are not substantial, however this is logical when taking into account that executives have long term contracts, and major sudden changes in the top management team are rare. The largest group is the one with 5 to 25 % females in the executive board, it means that most major US companies have moderate gender diversity. For the nationality diversity the groups are more equally distributed, with the remark that 37% of the companies are operating mainly in the United States, so that would make national diversity less of a necessity. The dominant leader groups 2 and 3 are similar in numbers, only group 1 is significant smaller. This is not a surprise while the data set is constructed of the 12 largest companies of the different industries and that in are mostly mature companies, and therefore less likely to be privately owned or still managed by the founder. For the last characteristic the difference in amounts of groups 1 and 2 is larger than expected.
34 Table 1, nr in the
differ-ent groups. 2011 2012 2013 2011 2012 2013 2011 2012 2013 group 1 group 1 group 1 group 2 group 2 group 2 group 3 group 3 group 3 Nr in group gender 15 15 13 37 36 35 8 9 12 Nr in group nationality 25 24 22 23 24 26 12 12 12
Nr in group Dominant
lead-er 8 7 6 25 28 28 27 25 26
Nr in group Functional
background 20 20 20 7 7 7 33 33 33
When looking at how the different groups are constructed, several noticeable outcomes are detected (Table2). The gender groups are distributed relatively similar over the several sectors, with as expected a higher percentage of females present in top management team functions in the branch consumer goods, than in the energy and heavy industry. When looking at the nationality groups, there is a higher diversity in the technology branch than in the other branches. But this can be explained by the high percentage of sales outside the US, where all the technology firms in this data set are earning more turnover abroad. This also explains the difference compared to the consumer goods and the nationality diversity outcome. Another noticeable difference with the technology sector is for the same reason the financial sector while they were mostly US oriented. The functional background has a surprising outcome, while in the financial sector the CEO mostly has a financial background as expected, on the other side the companies mainly focussing on marketing (consumer goods branch) do also have CEO’s with a financial background. The construction of the groups for the dominant leader, are more in-line with the assumptions. In the financial sector the CEO is more often less powerful, while in the fast moving technology market the CEO is more likely to be a dominant leader (Cooper and Kleinschmidt, 1994).
35
Table 2: Group by
sec-tor --(2011) gender group Group na-tionality Group Functional background Group dominant leader
1 2 3 1 2 3 1 2 3 1 2 3
Financial sector 4 6 2 6 3 3 7 0 5 1 1 10
Technology sector 3 7 2 1 6 5 0 3 9 5 4 3
Consumer goods
sec-tor 0 9 3 7 3 2 6 2 4 1 8 3
Healthcare sector 3 8 1 5 5 2 3 1 8 0 5 7
Energy and Heavy
in-dustry sector 5 7 0 6 6 0 4 1 7 1 7 4
The different hypothesis means
In this section the different means of the groups are discussed, to get a clear overview of the firm performance. Also the fundament of the expectations can be researched, together with the explanations of the outcomes of the statistical research can be analysed. It will gather information on why there is or is not an influence of the IDV’s on the DV. The firm performance is tested by three indicators, Profit, ROA and Tobin’s q, so all the means of the groups are included.
First the profit mean is investigated, in order to get a clear overview about which companies performed better on making profit. For every characteristic the combination of the 3 groups together make the 60 companies investigated. The average of the whole data set on profit in millions is in 2011: 7571, for the year 2012: 7620 and for the final year 2013: 8426. When looking at the profit mean of the different gender groups the companies that score high, and low on gender diversity, score better than the executive boards with medium gender diversity (Table3). All the different groups are gradually increasing the profit earned without any major changes. When it comes to the nationality groups the medium is performing above average, and the not diverse and very diverse are performing below average. This could be explained
36 by the communicational problems as a result of having high nationality and cultural differences (Milliken and Martins, 1996). Also the top management boards have access to a less diverse knowledge pool can score lower on firm performance (Athanassiou & Nigh, 2000). The lower mean of very diverse national teams is a surprise, especially when considering that the profit mean of the companies exporting more than 30% outside of the USA is 9000, and the profit mean of the firms exporting less than 30% is 5000. Furthermore, the steep rise in the diverse group 3 in the year 2012 to 2013 is noticeable, while there was not a shift in companies to this group, so the companies that performed very poorly in the 2012 recovered in the next year. The fact that the technology firms are highly present in group 3 makes this more surprising, while the profit mean of the technology firms in 2012 was 10761 (Table3), what suspects that the executive board members of less successful or smaller technology firms have a more diverse national background. The most effective functional background of the CEO seems to be the marketing specialty. Nevertheless, the sample of marketing background is small, and the data is influenced by the large profit contributor Wal-Mart. At last the dominant leader appears to have a positive effect on the firm performance indicator profit, while it was expected that a less powerful CEO would boost firm performance, by creating more board diversity (Schneider et al., 1995). The average of the profit means for board diversity are for all the years pretty similar. Where there is one exception for group 3 in 2012, however this can be explained as mentioned before by the very low profit mean for nationality diversity. Creating more diversity on executive boards does not seem to have any effect on the firm performance indicator profit.
37 Table 3: Profit mean
in groups (Mil) 2011 2012 2013 2011 2012 2013 2011 2012 2013
group
1 group 1 group 1 group 2 2 group group 2 group 3 group 3 group 3 Profit mean in group
gender 8651 9949 10104 6653 6041 7000 9792 10054 10765
Profit mean in group
nationality 5322 6408 7167 10945 10928 9864 5789 3426 7617
Profit mean in group Functional
back-ground 6310 6467 6786 11925 14634 14495 7412 6831 8132
Profit mean in group
Dominant leader 12206 10719 11128 7627 8100 8164 6146 6214 8084
Average of Profit mean for
Board diversity
indi-cators 8122 8386 8796 9288 9926 9881 7285 6631 8649
Secondly, the Return on assets performance is discussed, by analysing the differences between the groups. The outcomes of the means of ROA give similar results as the means on profit. More gender diversity has a slightly higher performance than the more homogeneous teams. This is also the case when analysing the groups of national diversity and the functional background (Table4). Group 2 of the functional background is mainly pushed by two main contributors: Apple and Wal-Mart. The dominant leader characteristic is the only group where less board diversity seems to have a greater effect on the ROA. This is due to the high percentage of financial companies in group 3, that score very low on ROA (Table,4), while the technology branch was performing high on ROA and was a large contributor for group 1 (Table1,Table2,Table3). On average more board diversity could have a positive effect on the DV indicator ROA.
38 Table 4: Mean of ROA in
different groups in % 2011 2012 2013 2011 2012 2013 2011 2012 2013
group
1 group 1 group 1 group 2 2 group group 2 group 3 group 3 group 3 Mean of ROA in group
gender 6,65 6,65 5,94 8,18 6,72 7,69 6,56 7,25 8,30
Mean of ROA in group
nationality 5,15 5,22 5,02 9,46 8,58 8,34 9,05 6,32 9,89
Mean of ROA in group
Functional Background 5,60 4,78 4,89 11,99 12,02 11,72 7,84 6,89 8,06
Mean of ROA in group
Dominant leader 11,31 8,93 8,94 6,96 7,47 7,74 7,04 5,41 6,75
Average of ROA mean for Board diversity
indi-cators 7,18 6,39 6,20 9,15 8,70 8,87 7,62 6,46 8,25
At last the indicator for the firm performance Tobin’s q is analysed (Table5). By having a look at the overview, the general trend seems to be that moderate board diversity would lead to a higher Tobin’s q. On the other hand in table 5 the influence of the financial sector on the group means cannot be ignored. Due to the negative image of the financial sector the market value of these companies is relatively low, and therefore this sector is influencing the means for Tobin’s q. With the high presence in the group 3 for Dominant leader, and in group 1 of the functional background, these means would be slightly higher when the financial sector was excluded in the research.
Table 5: Mean of Tobin's Q
in different groups 2011 2012 2013 2011 2012 2013 2011 2012 2013
group
1 group 1 group 1 group 2 group 2 group 2 group 3 group 3 group 3 Mean of Tobin's Q in group
gender 2,35 2,30 2,35 4,17 3,75 4,28 2,87 3,10 3,19
Mean of Tobin's Q in group
nationality 4,18 3,41 3,72 3,18 3,29 3,15 2,90 3,03 4,56
Mean of Tobin's Q in group
Functional background 2,38 2,12 2,64 5,01 5,07 4,77 3,93 3,62 4,01
Mean of Tobin's Q in group
Dominant leader 3,10 2,65 2,59 4,40 3,83 4,36 2,83 2,86 3,11
Average of Tobin’s Q mean for Board diversity
39 After analysing the different means of the several groups, the influence of the characteristics of board diversity on firm performance is likely to be limited. For the return of assets and the Tobin’s q more board diversity could stimulate these indicators of the DV, however the indicator profit seems to be consistent between all groups. Furthermore, most major differences could be explained by the data set and the branch the firm is operating in. Also the control variable “US oriented” seems to be an effect on the firm performance, and could explain therefore the differences in outcomes by group (Table6). The expectation for the hypothesis testing is therefore, in contradiction with most existing literature, that board diversity does not create higher firm performance. Only having more nationality diversity could have a significant influence, however this could be due to the control variable US oriented.
Table 6: Average Means of DV indica-tors by sector
Profit ROA Tobin's q
2011 2012 2013 2011 2012 2013 2011 2012 2013 Financial sector 7336 7529 9881 1,98 1,62 2,63 1,10 1,34 1,79 Technology sector 11314 10761 12480 11,98 9,56 10,30 3,71 3,61 3,56 Consumer goods sector 5500 5657 5867 8,50 8,56 9,21 4,74 5,38 6,15 Healthcare sector 4812 5289 6019 8,52 7,69 8,22 2,78 3,05 3,71
Energy and Heavy
40 Normality analysis
Before further analysis is performed, the normality of the independent variables and the moderating variables are addressed. In appendix table 3 there can be seen that 5% trimmed mean of the independent and the control variables, the 5% highest and lowest scores are not taken into account while calculating the mean. The sample shows that the numbers that are not entered in the normally distributions are small. Where this is obvious for the independent variable group indicators, while there are only 3 options. Also for the original indicators the number of cases not included is small, even while in many cases the value was 0. The Kolmogorov-Smirnov Test on the other side is indicating that the data set is not normally distributed, where the indicators score significant (Appendix 1, Table 2).
The variables “amount of board members” and the “characteristics” are not normally distributed by the analysis of the skewness and kurtosis, after grouping the companies the data is relatively more normally distributed by the indicators kurtosis for the characteristic groups: Gender and Dominant leader, while it is below 4. The skewness of the characteristics group is moderately positive or negative for the dominant leader group year 2013, however because the other years did not have this problem, the choice was made to ignore the small negative skewness.
41 Correlations
Some of the correlations are expected and logical to explain, so does the ROA correlate with both the profit and the Tobin’s q (Appendix 2, Table 1). Another correlation that is easily explainable, US oriented correlate with the national background group. The surprising element is that the control variable “branch” in which the company operates has no correlation with the firm performance indicators, even more that the Tobin’s q and profit are not connected. Also the fact that most independent variables groups do not correlate with the dependent variable indicators. Only the cultural background group has a moderate significant effect on the indicators ROA and Tobin’s q (Miles and Shevlin, 2006).
In order to understand what indicator for the characteristic national diversity creates the significance in relationship with the firm performance, another correlation schedule is constructed to see what triggers this effect. When measuring the independent variables indicators without the groups, the heterogeneity correlates with the dependent variable (Appendix 2, Table 2). Also the correlation of the independent variables gender and heterogeneity/nr of foreigners at the executive boards can be found, when not being grouped. Furthermore, the relationship of the ROA and the focus on the US is significant, what would suggest that the companies that get more than 30% of their sales outside of the US are performing better.
42 Regression analysis
To test the hypotheses several regression methods are used to investigate the influence of the IDV’s on the firm performance. However, before this analysis can be used, first a framework of which methods are suitable for the research is constructed. First multicollinearity is discussed, after an overview of the regression models is given, and at last the hypotheses are tested.
The multicollinearity issue can occur when the independent variables are highly inter-correlated. This can cause a reliability problem for the regression model, while it is difficult to measure the influence of one specific independent variable (Hill et al. 2009). There are two main indicators if multicollinearity is an issue, the first one is the tolerance, and the second one is the Variance Inflation Factor (VIF). When the values of the VIF are above 10 this is a clear indication that am inter-correlation effect can cause problems for the regression models (Neter et al, 1985). Consequently, multicollinearity is not an issue in this data set, while all the VIF outcomes are close to 1. The tolerance also indicate that multicollinearity is not a problem, and therefore the regression models can be performed
43 Overview of the regression models
In table 7 the regression models that are used are shown. As been explained in the methods section the regression models will explain if there is a significant influence of the independent variables on the dependent variables. The step by step regression method is used in order to see whether the control variables are explaining the relationship between the IDV’s and the DV’s.
Table 7: Regression models:
Regression models Dependent variable Independent variable Control variable
Model Control Profit Branch, US oriented
ROA Branch, US oriented
Tobin's q Branch, US oriented
Hypothesis 1
Model 1a Profit Gender group Total Branch, US oriented
Model 1b ROA Gender group Total Branch, US oriented
Model 1c Tobin's q Gender group Total Branch, US oriented
Hypothesis 2
Model 2a Profit National diversity group total Branch, US oriented
Model 2b ROA National diversity group total Branch, US oriented
Model 2c Tobin's q National diversity group total Branch, US oriented
Hypothesis 3
Model 3a Profit Functional background group Total Branch, US oriented
Model 3b ROA Functional background group Total Branch, US oriented
Model 3c Tobin's q Functional background group Total Branch, US oriented
Hypothesis 4
Model 4a Profit Dominant leader group Total Branch, US oriented
Model 4b ROA Dominant leader group Total Branch, US oriented
