The effect of national culture on the relation between gender diversity
on boards and firm performance.
MSc IB&M Thesis
Author: Maarten Kat Student number: S3335356 Email: m.j.kat@student.rug.nl Professor: Prof. Dr. Halaszovich, T.
Co-assessor: Dr. Schlägel, C.
Faculty of Economics and Business University of Groningen
Duisenberg Building, Nettelbosje 2, 9747 AE Groningen, The Netherlands P.O. Box 800, 9700 AV Groningen, The Netherlands
http://www.rug.nl/feb
ABSTRACT
This thesis stretches on the importance of gender diversity on boards, as there is a significant positive effect on firm performance. There is contributed, due to significant results, on the perspective of the critical mass theory, as a benchmark of three or more females is needed within the board to capture the full benefits of gender diversity. Furthermore, this thesis contributes to the research field by drawing upon an important factor influencing group compositions, group dynamics, social roles, and gender roles. Namely, national culture. There are significant contrary results to the power distance hypothesis, which may be explained by an earlier mediation effect. Furthermore, the masculinity hypothesis is supported, meaning that masculine societies negatively affect the positive effects of gender diversity. Contributions are made by showing the importance of females in numbers rather than a ratio, and showing the importance of national culture on shaping and affecting the structure, operations, and management of firms and the differences between countries. Supporting theories of national culture characteristics such as low gender egalitarianism and maintenance of distinction between classes and gender roles. This study has been conducted over a 5-year period on 154 firms, representing 20 different countries, and 12 industries.
ACKNOWLEDGEMENT
TABLE OF CONTENTS
INTRODUCTION ... 6
LITERATURE REVIEW ... 9
Gender Diversity and Firm Performance ... 9
Tokenism and Critical Mass Theory ... 11
The Effect of National Culture ... 13
Power distance ... 14
Masculinity-Femininity ... 15
Hypothesis Development ... 17
Conceptual Model ... 20
METHODOLOGY ... 20
Data Collection and Sample ... 20
Measurement and Measures ... 21
Dependent variable ... 21 Independent variable ... 22 Moderators ... 22 Control variables ... 22 RESULTS ... 24 DISCUSSION ... 32 CONCLUSION ... 36 REFERENCES ... 38 APPENDICES ... 47
Appendix A. Countries within regions, and allocated industries. ... 47
Appendix B. Nested models (Stata Corp. LLC, 2013) ... 48
Appendix C. Mediation effect of power distance. ... 50
Appendix D. Marginal effects. ... 51
Appendix E. Board national culture. ... 52
LIST OF FIGURES
Figure 1. Conceptual model research design. ... 20
Figure 2. Average yearly gender ratio. ... 24
Figure A1. Example multilevel model. ... 48
Figure A2. Means’ distributions based on hierarchical structure. ... 48
Figure A3. Notations of multilevel means. ... 49
Figure A4. Marginal effects Tobins’ Q & power distance. ... 51
Figure A5. Marginal effects Tobins’ Q & masculinity. ... 51
Figure A6. Marginal effects ROA & power distance. ... 51
Figure A7. Marginal effects ROA & masculinity. ... 51
Figure A8. Heteroskedasticity test A. Full sample, Tobins’ Q. ... 55
Figure A9. Heteroskedasticity test B. ≥ 3 females, Tobins’ Q. ... 55
Figure A10. Heteroskedasticity test C. Full sample, ROA. ... 55
Figure A11. Heteroskedasticity test D. ≥ 3 females, ROA. ... 55
LIST OF TABLES Table 1. Descriptive statistics. ... 24
Table 2. Correlation matrix. ... 25
Table 3. Descriptive statistics full and base sample national culture scores. ... 28
Table 4. Results nested models panel data hypothesis 1. ... 28
Table 5. Nested model regression analysis H2-3 for Tobins’ Q. ... 29
Table 6. Nested model regression analysis H2-3 for ROA. ... 30
Table A1. Countries within regions, and allocated industries. ... 47
Table A2. Mediation effect of power distance. ... 50
Table A3. Regression analysis board nationality scores Tobins’ Q. ... 52
Table A4. Regression board nationality scores ROA. ... 53
Table A5. Heteroskedasticity test A. Full sample, Tobins’ Q ... 54
Table A6. Heteroskedasticity test B. ≥ 3 females, Tobins’ Q. ... 54
Table A7. Heteroskedasticity test C. Full sample, ROA. ... 54
INTRODUCTION
Females have taken a rise in the corporate business environment in previous years. Due to increased importance in board gender diversity (Lee, Marshall, Rallis, & Moscardi, 2015; Eisenberg, Sundgren, & Wells, 1998; Adams, & Ferreira, 2009; European Commission, 2012). An increase in performed researches on gender diversity by organizations, such as the OECD (2018), European Commission (2018), and the United Nations (2018), has led towards greater pressures towards firms to integrate females into their organization, and to increase gender diversity. The increased awareness of gender diversity has led to the legislation of appointing females in boards of firms forced by certain countries (Sila, Gonzalez, & Hagendorff, 2016; Adams & Ferreira, 2009). The increased popularity for gender diversity over the previous years represents the importance of this subject.
Different cultural aspects may affect the relationship of gender diversity on firm performance differently. For instance, in Asia women are expected to take responsibility for most of the household and family chores, due to cultural norms, relatively more compared to Western societies (Benson & Yukongdi, 2005), gender-discrimination in India, where females are seen as less than males (Ram, Strohschein, & Gaur, 2014), or the emancipation of females in the Netherlands (Eleveld, 2015). These are all events, that are deeply rooted in cultural norms, values, and beliefs. There is a difference in the ratio of gender diversity among countries. For instance, in the US and UK, females represent, respectively, 16.1% and 15% in boards. On the other hand, in China females represent 8.1%, in Malaysia 7.8%, and in Singapore 6.4% of the boards (Süssmuth-Dyckerhoff, Wang, & Chen, 2012). For that reason, In this research there will be researched whether the relationship between gender diversity and firm performance is influenced by different national cultures. As aforementioned, institutional external influences do moderate the relationship, and cultural aspects cannot be neglected in this relationship (Low, et al., 2015). The author of this study proposes to analyse the effects of the dimension of power distance and masculinity on the relationship of gender diversity and firm performance, due to the fact these are related to group compositions, group dynamics, and group roles (i.e. gender roles and social roles) (Hofstede, 1984).
minority and majority (Bear, Rahman, & Post, 2010). In previous research there, has been found a weak relationship between gender diversity and firm performance (Salokoglu & Demir, 2016), one of the possible reasons could be that, in the majority of their sample, there was only one female representative on the board, which could indicate that the females did not have a substantial impact on the board. Considering the relationship between gender diversity and firm performance, influencing factors could be national culture, and the critical mass theory.
LITERATURE REVIEW Gender Diversity and Firm Performance
The Top Management Team (TMT) of an MNE can be seen as a consultancy group, that manages the firm, and protects and promotes their shareholders, the firm is guided by the TMT, and most important influences firm performance (Hambrick & Mason, 1984; Hambrick, 2007; Finkelstein, Hambrick & Canella, 2009). The functioning of a TMT is highly related to the performance of an organization (Zahra & Pierce, 1989). There is an increase in the popularity of researching board diversity (Terjesen, Sealy, & Singh, 2009) in the International Business (IB) research field. Especially gender diversity has experienced an increase in popularity. There are different perspectives on gender diversity. However, the main perspective on gender diversity is presented by the effective equality (Reguera-Alvarado, de Fuentes, & Laffarga, 2015), fair representation, and an equitable ratio (Sytsma, 2006) between females and males, it is the heterogeneity of females and males (Conyon & He, 2017) in boards. The author defines the following absolute form of gender diversity: the egalitarianism and heterogeneousness in an organization’s board between genders. However, for this study, the representation of three or more females within the board is sufficient, as it follows the critical mass theory. This is the critical mass needed to have a substantial effect on board compositions (Erkut, et al., 2008; Konrad, et al., 2008), and firm performance (Torchia, et al., 2011). Arising from the fact that the minority has a significant influence, and board compositions change drastically (Erkut, et al., 2008; Konrad, et al., 2008). Assuming that from this tipping point (three females or more), gender diversity has a significant influence on firm performance.
There has been argued by scholars that female TMT members bring different perspectives to the table. Convoluted issues are seen through the lens of this fresh perspective, that contribute to problem-solving and strategy formulation by correcting information biases (Dewatripont, et al., 1999; Westphal, & Milton, 2000). In addition to this, there is suggested that females do not serve as mere tokens, and bring disparate perspectives and fresh ideas to the board room, which could add value. However, the greater the diversification between managers, it is more likely that conflict could occur, arising from disagreement and too contrasting perspectives (Adams & Ferreira, 2009). Besides that, there is a difference between risk-aversion between males and females, females tend to take fewer risks in comparison to males. Due to the fact that females are more risk-averse in decision-making (Faccio, Marchica & Mura, 2011; Allini, Rossi, & Hussainey, 2015). One of the benefits for TMTs with increased gender diversity is that these TMTs take on fewer risks and have a greater overall firm performance (Perryman, et al., 2016).
As mentioned before gender diversity, in comparison to homogeneity, in groups enlarges the base of accessible information for decision making (van Knippenberg, et al., 2004; Dahlin, Weingart, & Hinds, 2005). There is an underlying assumption for this reasoning, namely, there is a difference in the experience between males and females that result in different sources and knowledge of information (Dahlin, 2005), as well that males and females tend to have different approaches in searching for information (Erhart, et al., 2003; Watson, et al., 1993). As a consequence, that gender diverse TMTs provides additional, necessary information that potentially improves the quality of decision-making. In addition to improved decision-making, heterogeneous groups tend to have a better competence to evaluate information in comparison to homogenous groups (Dahlin, et al., 2005).
and attendance behaviour of females could have a positive influence on the TMT attendance ratio, resulting in better-developed decisions. There are not only differences in the behaviour between males and females. However, when females enter a board, they bring along connections to different social networks, females have different socialization experiences (Simpson, Carter & D’Souza, 2010), and females tend to have a greater understanding of the marketplace and corresponding stakeholders of the firm (Carter, Simkins & Simpson, 2003).
The different perspectives, fresh ideas, and different knowledge bases (Adams & Ferreira, 2009; Dahlin, 2005; (Dewatripont, Jewitt, & Tirole, 1999) due to intensified gender diversity may lead to increased quality of brainstorm sessions, and fosters innovation (Erhardt, Werbel, & Schrader, 2003; Watson, et al., 1993). Other studies relate gender diversity to firm innovation as well, which is a critical predictor of firm performance (Caves & Ghemawat, 1992; Teece et al., 1997; Zahra & Garvis, 2000). Innovation leads to an increase in overall firm performance, because innovation fosters the generation of new ideas, processes, products, and systems within the organization. This can differentiate the firm from its competitors, and expand its operations (Hitt, Hoskisson, & Kim 1997). Furthermore, innovation can prosper capabilities that are firm-specific leading to an improvement of the firm’s performance (Teece, et al., 1997). These capabilities are an ideal combination of the firm’s resources, products, services, and strategies, that can be exploited, and which may create a competitive advantage. Overall, innovation will lead to better firm performance (Hitt, Hoskisson, & Kim 1997). In this case, the gender diversity in the TMT may foster innovation.
Tokenism and Critical Mass Theory
During the previous decade, there has been an increase in the number of arguments that females should be on boards (Cassell, 2000; Conyon, & He, 2017; Huse, 2005). However, females in boards serve as tokens in many countries (Torchia, Calabro & Huse, 2011; Kanter, 1977a; 1977b). Tokenism refers to the symbolic effort made, by corporations, to include a minority group to give an appearance of sexual or racial equality (Kanter, 1977a) within a board. Gender distinctiveness which can occur at low levels of gender diversity is associated with tokenism (Kanter, 1977b). It can be the case that the inclusion of a female in a multinational corporation (MNC) board is only an allegorical attempt of the MNC to show that they are representatives of gender equality.
and the dominant group (Kanter, 1977b). The classification of a token, usually, creates isolation, discomfort, and self-doubt (Kanter, 1977a; 1977b). Tokens can evolve into an out-group, by showing more negative work attitudes (Tsui, Egan, & O’Reilly, 1992), engagement of withdrawal behaviour, and lower communication levels (Liao, Joshi, & Chuang, 2004), in comparison with the in-group representatives. Furthermore, tokens are perceived as negative by the dominant group, which sometimes even involves precipitating derision (Maass & Clark, 1984; Nemeth & Wachtler, 1983), doubted and mistrusted out-group members (Kanter, 1977b). Perceived distinctiveness and tokenism can impede the firm performance (Powell, 1993). In fact, excessive visibility leads to increased pressures to perform and tokens are singled-out. This is because the token is a female, and not based on the individual's performance.
Tokenism is the underlying concept of the critical mass theory (Kanter, 1977a; 1977b). This theory suggests that the influence and interaction of groups depends on the size of a subgroup. There is an increased degree of influence of the subgroup, when their size reaches a certain threshold, also named the critical mass. However, when the minority group increases in size, and reaches a size that it is no longer perceived as a token. There is a qualitative change in the relation between the minority and the majority of the group (Bear, et al., 2010). The increase of the minority group is enhanced with trust in the group, and the major resource benefits brought to the organization by females are inaugurated to an effective use (Kanter, 1977b).
in the previous paragraph, the tipping point of 3 females or more has to be reached on a board, to fully reap the additional prosperity gender diversity has to offer.
The Effect of National Culture
Culture is an important topic in the IB field. Organizations, MNEs, and their TMTs have to deal with the man-made reality. Organizations are built according to people’s individual or group values and beliefs that reflect the community they live in (Hofstede, 1984). These societies are a configuration of organizations and institutions that are the reflection of their culture and the corresponding preeminent values and beliefs (Hofstede, 1984). Many scholars have given a definition of culture. Culture is a complex concept that consists of the prevalent norms, values, meanings, symbols, beliefs, and practices among a society (Schwartz, 2006). Hofstede (1984) defines culture as "the collective programming of the mind which distinguishes the members of one group or society from those of another”. In addition to this, culture consists of thinking patterns that carry the values and beliefs of a society, which are transferred within a society between the members. Culture identifies what is seen as good and bad in a given society (Hofstede, 1984). Furthermore, individual and group actions, goals, and beliefs are shaped by culture. Institutions and organizations represent the underlying concepts and values within a given society and culture (Schwartz, 2006). Culture explains and influences the behaviour of people in a given society (Roth, 1995). Little attention has been given to the cultural aspects related to the relationship between gender diversity and firm performance, but these factors cannot be neglected (Low, et al., 2015).
(Hofstede, 1984). As PD captures important hierarchical-related aspects, it seems appropriate that PD is selected for this study, as there are still hierarchical differences between males and females in the world. Collectivism seems another important dimension that captures group aspects. However, this is strongly related to PD (Beugelsdijk, Kostova, & Roth, 2017). As societies with high collectivism also have a high score on PD (Euwema, Wendt, & Emmerik, 2007), therefore, collectivism has been left out of this study. In order to capture a broader scope of national culture besides the values and beliefs of PD, the second dimension implemented in this study will be Masculinity-Femininity (MAS), as it defines the different roles between males and females within a society or community. As well as, the differences in values between countries that have different scores on the MAS dimension (Hofstede, 1984; Jones & Teegen, 2001). Besides that, it does capture factors related to group processes (Elron, 1988).
Power distance
The main issue that societies address is human inequality, this is conducted differently in every society. Power distance refers to the degree related to power inequality between an individual that is less powerful and an individual that is more powerful, and in which both individuals belong the same society, or social system. There are major differences in power distance across various institutions, these underlying differences are exhibited in communities, families, but most importantly on the work floor and in organizations (Hofstede, 2001).
and between the ‘powerless’ and ‘powerful’ members of society are relationships based on latent harmony (Hofstede, 2001). While on the other hand, in high power distance countries or societies, there should be inequality as this provides everyone their rightful place and individuals are protected by their rightful place, there are prerogatives for powerful individuals, and the relationship between ‘powerless’ and ‘powerful’ is based on latent conflict (Hofstede, 2001).
Societal hierarchical structures are described by gender egalitarianism and power distance, which complement each other. The underlying inequality in hierarchical societal groups, such as females and males, young and old, are captured by both of the variables. This is especially dominated in Chinese Confucian societies (Farh, et al., 1997; House, et al., 2004). There is a negative correlation between power distance and gender egalitarianism (Gudykunst & Lee, 2003). Meaning that in high power distant societies there is low egalitarianism, and vice versa. There is a greater difference between genders in Asia than in Western societies (House, et al., 2004). In India, there is still a presence of gender-discrimination towards females (Ram, et al., 2014), and are trivialized in these countries. In addition, in Thailand, people tend to have a greater interpersonal and professional distance between each other, the formalism maintains the authorial relationships. The normative expectations are even greater for females than for males in these specific countries, due to distinct gender roles and low egalitarianism (Gupta, Surie, Javidan, & Chokar, 2002; Hofstede, 2001). While in countries such as the US (low power distance), there is a greater fit and legitimacy for females in the workplace (Hoobler, Wayne, & Lemmon, 2009). In cultures that are characterized by high scores of power distance, people seek maintenance, and potentially increase, power differences between society members as a form to gain satisfaction. This is because the members of these societies have a higher social consciousness, therefore, have higher motivational levels to conform to their own social class, or are willing to reconcile to higher classes, to which they aspire. While in lower power distant cultures, there is a much lower focus on the differences between classes and social aspirations (Roth, 1995).
Masculinity-Femininity
2001). The difference between cultures on this dimension is not based on the absolute difference between sexes, males and females, fosters complexities in any given society. These complexities are based on this nature-given fact, whom all societies have to cope with, which is addressed in their own specific way. This has affected and shaped the many different societal institutions: “the sex-role system is at the core of our cultural norms” (Chetwynd & Hartnett, 1978). The absolute differences between males and females are the same for all societies1. The differences between cultures is based on the wide margin that is left by these differences. The margin has led to the different designation of values and behaviour to males and females in societal institutions, that is considered as applicable and suitable for that specific gender. The different behaviours and values recognized across cultures, is what distinguishes cultures from one another (Hofstede, 2001).
The antithetical dimension, masculinity-femininity, shows different values and practices at both ends of the continuum. In a masculine culture, there is a clear distinction between social gender roles. In these societies, the male is seen as though, assertive, and has a focus on material success, and the female in this society has to be tender, modest, and should have a greater concentration on quality of life. On the other hand, in feminine societies, there is an overlap of these values between men and female, and both genders can have the ambition to seek satisfaction in values that are supposed for the other gender (according to high masculine societies) (Hofstede, 2001).
A masculine society strives for achievement, assertiveness, and heroism. In these societies, social differentiation between sexes is preferred to be maximal. In which males have assertive roles, and females the caring roles within a society. While in feminine countries this differentiation should be minimal (Hofstede, 1984; Jones & Teegen, 2001), meaning that the differences between males and females are hardly different (Hofstede, 2001). Characteristics of these societies are modesty, relationship, caring for the weak (Hofstede, 1984; Jones & Teegen, 2001), and society members show empathy to others regardless of the group. In these ‘feminine’ societies the males take care of all values, they are tender and maintain relationships, but do not neglect the performance of an organization (Hofstede, 2001).
Masculinity is an important dimension that captures group processes (Elron, 1988). There is believed that rewards come from performance in masculine societies, and from need in feminine societies. There is a difference in the dissemination of rewards between masculine
1 The only real absolute differences that can be given the label "masculine" or "feminine" are related to the ability
and feminine societies, respectively, rewards are favoured in the form of equity (i.e. pay based on performance and merit), or rewards based on mutual solidarity and equality (Hofstede, 2001). Superman would be a role model, in for instance, the US, while in the Netherlands, there would be a preference to provide sympathy for the underdog. Even though, the US (masculine) is associated with “performance society”, this does not necessarily mean that firms from masculine societies outperform firms from feminine societies (Hofstede, 1984).
In feminine countries making group work more rewarding is a primary purpose. This has to be achieved to allow for semi-autonomous groups and cooperation between the group members is fostered (Hofstede, 1984). As well, increased group cooperation will enhance the relationship between group members. In for instance Japan (masculine), work roles for different sexes are set and predetermined (Hofstede, 1984). Besides that, in Venezuela females are suppressed and males are superior (Díaz, 2004).
Hypothesis Development
Gender diversity has many positive effects on firm performance. As this contributes to better decision-making in boards arising from gender diversity (van Knippenberg, et al., 2004; Dahlin, Weingart, & Hinds, 2005), increased knowledge bases and sources (Dahlin, 2005), fresh ideas and perspectives (Dewatripont, Jewitt, & Tirole, 1999; Westphal, & Milton, 2000), differences in behaviour that positively affect the whole group and its performance (Singh & Vinnicombe, 2004), and a more risk-averse decision-making (Faccio, Marchica & Mura, 2011; Allini, Rossi, & Hussainey, 2015; Erhardt, et al., 2003; Watson, et al., 1993). In comparison, homogeneous groups might fail to achieve the same decision-making efficiency, and innovation levels (Dahlin, et al., 2005). The benchmark of three females has to be realized in order to have substantial impact on the overall board and the organization's performance (Erkut, et al., 2008; Konrad, et al., 2008). In this case, females will not be seen as tokens anymore (Kanter, 1977b), and lose the negative consequences of tokens and minority groups. The negative effects of tokens on performance (Powell, 1993) will diminish. As this diminishes the negative pressures on the minority group, and the minority group gains a substantial amount of influence in the group (Bear, et al., 2010).
Hypothesis 1: there is a positive relation between the critical mass of gender diversity on a board and the firm performance.
The power distance dimension addresses hierarchical-related aspects of a culture, which are determined by several aspects, that are to some extent related to differences in gender (Hofstede, 2001). In high power distance cultures, there is a greater distinction between classes and members of society (Hofstede 1984; Inkeles 1960; Lenski 1966), which are preferred to be maintained or even to be increased (Hofstede, 2001). While in cultures with low power distance, there is less fixation on diversification (Hofstede 1984; Inkeles 1960; Lenski 1966), and inequality between social members should be as minimal as possible (Hofstede, 2001). Hierarchical structures in a given society are explained by power distance and gender egalitarianism, and complemented by each other (House, et al., 2004). PD has a high negative relation to gender egalitarianism within a given society (Gudykunst & Lee, 2003). As there are greater inequalities and distinct gender roles in high power distant cultures (House, et al., 2004). Which can be seen in the fact that, females are more suppressed, or even discriminated. in these cultures (Hofstede, 2001; Gupta, et al., 2002; Ram, et al., 2015).
There is expected that the influence of female board representatives is lower in high PD countries, than in low power distant countries. As in the former, females are more suppressed (Hofstede, 2001; Gupta, et al., 2002; Ram, et al., 2015), it is harder for them to gain a sufficient ‘voice’ (Erkut, et al., 2008; Konrad, et al., 2008), and these will possibly be concealed. Therefore, decision-making, and innovation efficiencies will be harder to achieve. Meaning that, the benefits of gender diversity in boards will not come to their maximal potential, such as greater accessibility to information sources (Dahlin, et al., 2005), and females might have less influence on the attendance behaviour of males (Adams & Ferreira, 2009). Females may possibly be more suppressed by the characteristics of power distant countries. In low power distance countries, there is a greater equality and acceptance of social groups and females, and there is a greater fit and legitimacy for females in the workplace (Hoobler, et al., 2009). For that reason, the input, and efforts, of females in boards is more accepted in low power distance countries, leading towards a better and more risk-averse decision-making (Faccio, et al., 2011; Allini, et al., 2015).
The masculinity dimension examines the distinction between gender roles among societies (Hofstede, 1984; 2001), as well as group processes (Elron, 1988). Masculine cultures are seen as “performance societies” (Hofstede, 1984), that labels males and females with specific roles and values (Hofstede, 1984; Jones & Teegen, 2001). In feminine societies, gender equality is favoured (Hofstede, 2001), and group work is more rewarded (Hofstede, 1984). Besides that, the benefits of gender diversity in TMTs might come more to their potential, and quality of decision-making will increase.
Furthermore, relationships and caring for others is valued in feminine societies (Hofstede, 1984; Jones & Teegen, 2001), and society members show empathy to others regardless of the group (Hofstede, 2001). Meaning that minorities in groups (Kanter, 1977a; 1977b) will be more appreciated and taken care of. Because, people in feminine cultures foster gender egalitarianism (Hofstede, 2001), and minority groups will be less suppressed, females will be more accepted. This will enhance the advantages provided by gender diversity in boards. Female’s ‘voices’ will have a greater impact, in feminine societies compared to masculine societies. Therefore, decision-making will be more efficient, leading to increased innovation (Erhardt, et al., 2003; Watson, et al., 1993), and accepted fresh perspectives (Dewatripont, et al, 1999; Westphal, & Milton, 2000), all in favour to firm performance (Hitt, Hoskisson, & Kim 1997).
In the case that the critical mass would be achieved, females gain a considerable impact on the board’s practices, and decision making. There is expected that they have a greater impact in feminine societies. In the case of masculine societies, females would still be less accepted in masculine societies. As there are more distinct gender roles, and gender equality is less favoured (Hofstede, 2001). Therefore, the higher acceptance of minority groups in feminine countries enhances the group processes and benefits of gender diversity more. Meaning that masculine cultures have a negative influence on the relationship between gender diversity and firm performance.
METHODOLOGY Data Collection and Sample
For the population for the research, there has been made a selection of various countries. Based on the S&P 500 firms (Bureau van Dijk, 2017) countries have been selected. As there was expected that, the countries with the most representatives would have the most available data on firms. This resulted in 14 countries, which were complemented by 6 more in order to get a population of 20 countries with varying power distance and masculinity scores. The sample represents firms from countries such as China, the United States, and Norway. As the Norwegian government passed the Norwegian quota law in 2005, which stated that at least 40% of each gender had to be represented within the board of organizations (Economist, 2018). Due to this, the population will be represented by firms that have boards of at least 3 females. Expected is that the critical mass will be achieved in more sample firms.
The data used can be classified as secondary data, since a third party has collected and stored the data in an accessible database (Blumberg, Cooper, & Schindler, 2014). Furthermore, there are different perspectives on what should be an appropriate minimum amount of the sample span. Thomas (2004) argues that 200 units are the minimum amount for an appropriate analysis, while some argue for a smaller sample size of 100 units (Bailey, 1994). Problems might occur when the sample size used is insufficient. A sample of at least 100 units is needed for an appropriate research. The data collected is from the data bank Orbis, which has financial aspects of over 275 million firms (Bureau van Dijk, 2017). A time period of 5 years is chosen starting in 2012 and ending in 2016. Because, this is the most recent available data. A
multi-Gender diversity (≥ 3 females:
critical mass) Firm Performance
Power Distance Masculinity
- -
+
Figure 1. Conceptual model research design.
H2 H3
H1
year approach can correct for the outliers of a one-year-firm observation. This contributes to gain the most accurate perspective on the firm.
For the sample collection firms needed some requirements on data that should be available in the database. Firms had to be publicly listed as from these firms the most data would be available. Secondly, the ROA, operating revenue, liabilities and Tobin’s q should be available as these is needed to measure some of the variables. Thirdly, data on the directors had to be available, in terms of age, time on board, and nationality. This resulted in a total of 16,048 firms of the corresponding countries. Secondly, the top 10 largest firms from all countries, based on operating revenue have been collected, resulting in 200 firms. There were some missing financial values from the Orbis data base, which resulting in a sample of 184 firms.
Unfortunately, Orbis data on board member characteristics is lacking, or only the last available year can be obtained from this data bank. For that reason, a second data bank, BoardEx, is used in order to complement the missing values from Orbis. BoardEx provides information of 1.7 million companies and individuals over multiple years (BoardEx, 2019), which helped filling the missing values of board age, board tenure, and board member nationalities. However, there were still some missing values. Therefore, primary data was added to the overall data set, by analysing annual reports of some firms, and missing board member nationalities were added to the data set. The final sample consists of 154 firms from 20 countries, being active in different industries (Appendix A; table A1). A multi-industry research has been selected to avoid the limitation of generalizability, that single-industry researches suffer (Simpson & Kohers, 2002). The multi-industry setting will contribute to the validity of this research.
Measurement and Measures
Dependent variable
assets of the firm. This ratio shows how well the assets of the firm are utilized to generate profit. A more efficient firm should have a higher ROA ratio (Campbell & Minguez-Vera, 2007). The better the performance, mostly refers to a better competitive advantage (March & Sutton, 1997).
Independent variable
The independent variable is gender diversity. For this study, the number of females that are representatives within the board has been chosen, because of the critical mass theory. The number of females should be “at least three” to make a significant impact on the firm performance of an organization (Kanter, 1977b). For this measure, a dummy will be used "1" indicating ≥ 3 females, and “0” otherwise. Next to that, a ratio measure will be used the percentage of females on the board. In addition to this, the Blau Index will be used which measures the variety and balance of diversity (Stirling, 1998). The index is used to measure heterogeneity in the board by calculating the percentage of each gender in the board, and the total number (n) of board members. The index provides a number in the range from 0 to 0.5. In which 0 means a homogeneous board of one gender (in this study males), and 0.5 means a heterogeneous board in which the number of females and males are equal (Campbell & Minguez-Vera, 2007).
Moderators
Two Hofstede dimensions will be used: power distance and masculinity. For both of the variables, the country score will be used (Hofstede, 1984). Even though, Hofstede’s data is captured 45 years ago, the data is still relevant today to analyse cultural impacts. The chosen moderators are from the first 4 dimensions, later Hofstede added 2 more dimensions (Maseland & van Hoorn, 2017). Furthermore, there will be made a distinction between low and high countries based on the median of power distance (50) and masculinity (56).
Control variables
GDP per capita. Different economic developments lead to income inequalities. Income inequalities might be greater during high phases of economic development (Kuznets, 1955). Higher GDP per capita is associated with greater firm-specific performance (Chun, Kim, Morck & Yeung, 2008). Collected from the world factbook (CIA, 2019).
attitudes within countries may affect gender diversity in boards. Retrieved from The World Bank (WorldBank, 2019).
Firm size. This is measured by the total assets the firm possesses (Marimuthu & Kolandaisamy, 2009). Firm performance can be influenced by the number of resources a firm has. Larger firms have a larger resource pool, which positively affects firm performance, in terms of financial outcomes (Krishnan & Park, 2005; Campbell & Mínguez-Vera, 2008; Vedd & Yassinski, 2015).
Firm age. Measured in the total number of years the firm is actively in business. Young firms might have a greater failure risk, because they might have a lack of valuable resources (Thornhill & Amit, 2003). While, older firms might fail because they cannot adjust to the changing environment and demands (Thornhill& Amit, 2003), or might be less innovative, which impacts decision-making (Huergo & Jaumandreu, 2004).
Board size. Research within the gender diversity field commonly controls for this variable (Campbell & Mínguez-Vera, 2008; Conyon & He, 2017). This is measured in the total number of members within the board. There are contrasting findings on board size. First, larger boards have access to more knowledge (Pearce & Zahra, 1992), but decision-making could be less effective (Lipton & Lorsch, 1992). Both, affecting performance.
Board tenure. Measured in the average years of the directors within the team. Heterogeneity of tenure might lead to greater impower balance, cohesiveness, and lower interaction among members, resulting in poorer outcomes and decision-making (Wiersema & Bird, 1993).
Board nationality score. Measured in the average of the sum of board members national power distance or masculinity score.
RESULTS
The summary statistics are shown in table 1. The 5-yearly panel data of 154 firms led to a total of 770 observations. The mean of Tobins’ Q (Table 1) reveals that that the majority of the firms in the sample are undervalued. A low Q (0-1) indicates that the costs for asset replacement are greater than their market stock price, meaning that the firm is undervalued. A high Q (>1) means that the firm is overvalued (Brown & Caylor, 2006). The mean of the ROA (μ = 4.131) indicates that the majority of the firms was successful in their financial performance. However, the negative minimum value indicates that there were firms with financial losses within the sample. The average female ratio on boards was 14.3%. In figure 2, the average gender ratios are shown. There was a substantial increase over the 5-years, which Table 1. Descriptive statistics.
Variable Obs. Mean Std. Dev. Min Max
Tobins' q 770 0.765 1.030 0 8.99 ROA 770 4.131 6.454 -1.42 48.63 Gender ratio 770 0.143 0.126 0 0,55 PD score 770 52.071 19.147 18 93 MAS score 770 53.246 21.719 5 95 Total assets 770 66,000,000 83,000,000 570,000 440,000,000 Firm age 770 58.734 45.970 0 362 Board tenure 770 5.664 3.442 0.40 20.93 Board age 770 63.835 4.998 47.70 77.25 Board size 770 12.187 4.345 4 28 EPAOS 770 61.839 43.005 1 139 GDP 770 37999.73 16171.74 5783 72100 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 2012 2013 2014 2015 2016
Average Gender ratio
Table 2. Correlation matrix.
may be the result of imposed gender quotas in certain countries (Economist, 2018), or the global acknowledgement of the importance and benefits of gender diversity. The median of power distance (50) should acknowledge a well-balanced distribution of low and high scoring countries. However, the mean (μ = 52) shows that there were slightly higher scoring countries. The masculinity values (μ = 53.2; median = 56) imply that the sample is dominated by higher scoring countries.
Multicollinearity
Correlation coefficients that are below 0.3 can be treated as trivial (Krishnaswamy, Sivakumar & Mathirajan, 2006). Different benchmarks are given for high correlations, such as above 0.7-0.8 (Lang, Secic & Lang, 2006) or even above 0.5 (Deborah, 2016) should be treated as high correlation. For this study, the benchmark of above 0.6 is treated as high collinearity, and variables should be treated with care. As OLS would be hard to conducts with high correlations. The correlation matrix (Table 2) shows that there is a high correlation between Tobins’ Q and ROA, which is not surprising as both capture financial performance. These two performance measures will not be included in the same model. The correlation coefficients of gender diversity indicate that the variable is positively associated with Tobins’ Q and ROA, as predicted by the hypothesis. However, have no direct significant correlation. Furthermore, EPAOS and GDP per capita both have a high correlation with gender ratio and power distance (PD score). Different regressions were ran with the inclusion of these variables (VIF = 2.00), and with the exclusion of these variables (VIF = 1.42), and multicollinearity between independent variables was measured. In the former analysis independent variable VIF values were close to 4. While in the latter al independent VIF values were below 2.
Larger values of VIF indicate that the predictor has a greater variance of the regression weight (Norisus, 1990). Different indicators for VIF are given, greater than 10 is indicative for multicollinearity (Stevens, 2002), greater than 5 (Rogerson, 2001), or even above 4 (Pan & Jackson, 2008). To avoid multicollinearity, that may affect models, the VIF value of 4 is taken and the variables EPOAS and GDP were excluded from the study.
Heteroskedasticity
second on Tobins’ Q with the sample three or more females, third on ROA and the full sample, and finally on ROA and the sample three or more females. For all the four tests the null Hypothesis had to be rejected (p<0.000). To correct for heteroskedasticity, robust standard errors were used (Joecks, Pull & Vetter, 2013). Heteroskedasticity can be a threat for the internal validity of the study (Stock & Watson, 2003), by using robust standard errors the internal validity of this study will increase, and there will be corrected for possible misinterpretations of the data.
Analysis
The marginal effects of the interaction terms (gender ratio x power distance & gender ratio x masculinity) have been calculated, the figures are shown in Appendix D. Marginal effects are calculated by taking the sample mean, the mean minus one standard deviation, and the mean plus one standard deviation. The marginal effects for power distance and Tobins’ Q
Table 3. Descriptive statistics full and base sample national culture scores.
VARIABLE N Mean S.D. Min .25 Mdn .75 Max
PD score (≥3 F) 205 39.02 11.93 18.00 31.00 35.00 40.00 80.00
PD score 770 52.07 19.15 18.00 36.00 50.00 69.00 93.00
MAS score (≥3 F) 205 43.89 25.46 5.00 14.00 61.00 66.00 70.00
MAS score 770 53.35 21.72 5.00 42.00 56.00 66.00 95.00
Table 4. Results nested models panel data hypothesis 1.
MODELS (1a) (2a) (3a) (1b) (2b) (3b)
VARIABLES Tobins' Q Tobins' Q Tobins' Q ROA ROA ROA
Gender ratio 0.958*** 0.417 4.239*** 4.620** 3.097 19.74***
(0.226) (0.299) (0.541) (2.320) (4.007) (5.428)
Total Assets -1.90e-09** -1.49e-09**
-3.55e-09* 8.63e-10
6.21e-09*** -2.00e-08 (8.26e-10) (6.09e-10) (2.04e-09) (4.79e-09) (2.11e-09) (1.51e-08) Firm age 4.37e-05 -0.000483 -0.000551 -0.00111 -0.00465 -0.00684
(0.000764) (0.000810) (0.00101) (0.00251) (0.00403) (0.00427)
Board tenure 0.0126 0.00213 0.0746 0.0537 -0.00502 0.420*
(0.0182) (0.0130) (0.0487) (0.0600) (0.0483) (0.248)
Board age -0.0122 -0.0118 -0.00949 -0.119* -0.138 -0.0184
(0.0183) (0.0242) (0.0210) (0.0649) (0.147) (0.217)
Board size -0.00334 0.00330 -0.00853 -0.00984 -4.15e-05 0.149
(0.0227) (0.0198) (0.0172) (0.125) (0.163) (0.0908) Industry dummy -0.197 -0.346* 0.307 -1.202 -2.120** 1.671 (0.221) (0.203) (0.188) (1.072) (0.949) (1.198) Constant 1.643 1.725 0.150 11.74*** 13.87* -2.845 (1.036) (1.447) (1.592) (2.919) (7.678) (17.60) Observations 770 565 205 770 565 205 Number of groups 4 4 3 4 4 3 Wald Chi-squared 42.38 38.86 27.01 17.74 22.85 17.87 Prob > Chi-squared 0.0000 0.0001 0.0003 0.0132 0.0018 0.0126 Adjusted R-squared 0.6774 0.0546 0.0570 0.50285 0.0366 0.0564
Table 5. Nested model regression analysis H2-3 for Tobins’ Q.
MODELS (3a) (4a) (5a) (6a) (7a) (8a)
VARIABLES Tobins' Q Tobins' Q Tobins' Q Tobins' Q Tobins' Q Tobins' Q
Gender ratio 4.239*** 3.717*** -1.419 3.981*** 8.562*** 0.107 (0.541) (0.636) (3.826) (0.653) (0.524) (2.054) PD score -0.0291** -0.0623*** -0.0850*** (0.0142) (0.0221) (0.0192) Gender ratio x PD score 0.131 0.233*** (0.104) (0.0515) MAS score -0.00996*** 0.0255*** 0.0388*** (0.00291) (0.00322) (0.00672) Gender ratio x MAS score -0.115*** -0.141*** (0.00873) (0.00534) Total Assets -3.55e-09* -3.85e-09* -4.14e-09** -3.40e-09* -3.79e-09* -4.60e-09** (2.04e-09) (2.15e-09) (1.95e-09) (2.00e-09) (2.11e-09) (2.15e-09) Firm age -0.000551 -0.00106 -0.00113 -0.000705 -0.000964 -0.00146* (0.00101) (0.00101) (0.000861) (0.000990) (0.00110) (0.000853) Board tenure 0.0746 0.0749 0.0839* 0.0752 0.0670 0.0808** (0.0487) (0.0487) (0.0438) (0.0475) (0.0418) (0.0390) Board age -0.00949 -0.0159 -0.0157 -0.00910 -0.00444 -0.00704 (0.0210) (0.0195) (0.0183) (0.0179) (0.0200) (0.0178) Board size -0.00853 -0.00371 0.00520 -0.00677 -0.00748 0.00946 (0.0172) (0.0163) (0.00980) (0.0216) (0.0238) (0.0128) Industry dummy 0.307 0.312 0.260 0.292 0.360* 0.288 (0.188) (0.196) (0.221) (0.189) (0.208) (0.226) Constant 0.150 1.933* 3.098 0.630 -1.117 1.829 (1.592) (1.068) (2.098) (1.340) (1.550) (1.743) Observations 205 205 205 205 205 205 Number of groups 3 3 3 3 3 3 Wald Chi-squared 27.01 32.17 34.69 28.05 37.59 45.96 Prob > Chi-squared 0.0003 0.0001 0.0001 0.0005 0.000 0.0000 Adjusted R-squared 0.0570 0.1120 0.1485 0.0676 0.0540 0.1304
Table 6. Nested model regression analysis H2-3 for ROA.
MODELS (3b) (4b) (5b) (6b) (7b) (8b)
VARIABLES ROA ROA ROA ROA ROA ROA
Gender ratio 19.74*** 18.11*** 17.22 18.03*** 36.06*** 20.98** (5.428) (3.969) (14.86) (4.727) (4.391) (10.56) PD score -0.0785 -0.0843 -0.157 (0.139) (0.0834) (0.108) Gender ratio x PD score 0.0226 0.408** (0.286) (0.161) MAS score -0.0456** 0.0936*** 0.118*** (0.0229) (0.0233) (0.0195) Gender ratio x MAS score -0.448*** -0.491*** (0.0500) (0.0478) Total Assets -2.00e-08 -2.11e-08 -2.11e-08 -1.92e-08 -2.04e-08 -2.21e-08
(1.51e-08) (1.59e-08) (1.54e-08) (1.47e-08) (1.50e-08) (1.62e-08) Firm age -0.00684 -0.00853 -0.00854 -0.00782* -0.00875* -0.00988 (0.00427) (0.00657) (0.00641) (0.00449) (0.00509) (0.00612) Board tenure 0.420* 0.424* 0.426* 0.426* 0.390* 0.417** (0.248) (0.244) (0.228) (0.242) (0.223) (0.208) Board age -0.0184 -0.0416 -0.0417 -0.0180 0.00956 -0.00236 (0.217) (0.210) (0.209) (0.209) (0.206) (0.187) Board size 0.149 0.150** 0.151** 0.155 0.152 0.174** (0.0908) (0.0754) (0.0618) (0.119) (0.121) (0.0853) Industry dummy 1.671 1.717 1.708 1.581 1.826 1.731 (1.198) (1.294) (1.367) (1.147) (1.280) (1.479) Constant -2.845 2.520 2.738 -0.404 -7.852 -1.787 (17.60) (11.47) (13.98) (16.71) (16.13) (12.35) Observations 205 205 205 205 205 205 Number of groups 3 3 3 3 3 3 Wald Chi-squared 17.87 19.23 19.24 18.92 22.63 23.34 Prob > Chi-squared 0.0126 0.0137 0.0232 0.0153 0.0071 0.0158 Adjusted R-squared 0.0564 0.0792 0.0801 0.0634 0.0457 0.0646
show that the higher power distance, it out performs lower scores of power distance. The same holds for the ROA, however the effects are close to each other. However, sample mean of power distance is 39, the mean minus one standard deviation is 27, and the mean plus one standers deviation is 50. Which are all considered as low scores based on the median. Meaning that the low scores have a positive effect, this could still be in line with the hypothesis. If we consider the marginal effects of the interaction term of gender diversity x masculinity. There is visualized that, at low levels of masculinity there are more positive effects on firm performance as gender diversity increases, while the effects on power distance, are declining as gender diversity increases at high levels of masculinity.
The models provide significant support for hypothesis 1. Positive significant effects for Tobins’ Q and ROA for hypothesis 2, which is contrary to the hypothesis, not supporting hypothesis 2. However, there has been found a mediation effect of PD on the number of females, which may signify for biased results. Finally, there is significant support for hypothesis 3, for both measures. The R-squared results of the models and differences between performance measures cannot be neglected. The proportion of variation is more explained for Tobins’ Q (13.04% in model 8), than for ROA, which is significantly lower (6.46% in model 8). Furthermore, since there were two performance measures used in this research, robustness checks to analyse the strength of the results are already integrated in the study itself. As the models with different performance measures show similar results there is corroboration for the support of hypothesis 1 and 3, and rejection of hypothesis 2. There can be concluded that the results are of significant strength.
In the final model total assets and board tenure are significant (P<0.05) and firm age partially significant (P<0.1) for Tobins’ Q. For ROA, board tenure and board size are positive significant (p<0.05) meaning that there are indeed other board characteristics influencing firm performance.
ROA in which national culture is included, the board cultural dimensions and board interaction terms have significant levels (p<0.01), the direct effect of power distance a bit lower (p<0.05). In the model of Tobins’ Q only the interaction term of power distance is significant (p<0.05) and masculinity is partially significant (p<0.1). In the model of ROA there are higher significance levels (p<0.05) for the national cultural scores. Both culture measures, national and board, are significant. When the effect on boards with three or more females is checked. National power distance remains significant. However, the board power distance loses its significance. National masculinity loses its significance as well, while the interaction term of board MAS is significant for Tobins’ Q, and the direct and moderator of board MAS for ROA, are significant.
When the boards with a net difference of zero are excluded from the sample (A5), excluding power distance board scores that are the same as national scores, all variables lose their significance level. In the model that excludes firms with the same masculinity score (A6), results show that in the Tobins’ Q model, (national) power distance, masculinity and their interaction terms are significant (p<0.01), and only board masculinity (p<0.01) and its interaction term (p<0.05) are significant. While in the ROA model, only masculinity variables have significance levels: masculinity (p<0.1); interaction term (p<0.05); and the interaction term with board masculinity (p<0.05).
DISCUSSION
Applying the critical mass theory to domain of MNE boards, the results of this study suggest that the contribution of females on boards to firm performance, is dependable on the size of the minority group. Suggested is that the tipping point of at least three females has to be reached for the minority group to make valuable contributions to the firm performance. Findings show that the critical mass can be taken to practice, and show the importance of the social position the minority group takes in a majority group. Furthermore, it indicates the importance of the number of females rather than a ratio.
performance. These results show that below the tipping point females serve as tokens (Torchia, et al., 2011; Kanter, 1977a; 1977b) within the boards. In this case the 1 or 2 females conform to the unanimous opinion of the majority group (Asch, 1951l; 1955; Torchia, et al., 2011). Confirming that 2 representative females within the board, is not a group size large enough to overcome tokenism, and the discrepancy in behaviour between the tokens and the dominant group remains (Kanter, 1977b).
Once the tipping point of three or more females is reached, the group dynamics change. The minority group is no longer perceived as a token, once the favoured size is reached. The relationship between the minority and majority group transubstantiate a qualitative change (Bear, et al., 2010). The minority group obtains more trust (Kanter, 1977b), and gains a significant amount of influence and ‘voice’ (Erkut, et al., 2008; Konrad, et al., 2008). At this point, the females will have substantial influence on the performance of the firm (hypothesis 1). The finding that three or more females is the critical mass are in line with previous studies (Asch, 1951; 1955; Nemeth, 1986; Bond, 2005; Erkut, et al., 2008; Konrad, et al., 2008; Torchia, et al., 2011). The group experiences a metamorphose from a homogenous to a heterogenous group composition, and will be able to reap the advantages a heterogenous gender group has to offer. There will be contributed to a better (van Knippenberg, et al., 2004; Dahlin, Weingart, & Hinds, 2005) and more risk-averse decision-making (Faccio, Marchica & Mura, 2011; Allini, Rossi, & Hussainey, 2015; Erhardt, et al., 2003; Watson, et al., 1993). Furthermore, there will be an increase in fresh ideas and perspectives (Dewatripont, et al., 1999; Westphal, & Milton, 2000), and knowledge bases and sources (Dahlin, 2005). The differences of behaviour present in heterogenous group positively influence the group and firm as a whole (Singh & Vinnicombe, 2004), contributing to firm performance. These levels of performance may not be reached by homogenous groups (Dahlin, et al., 2005).
Secondly, power distance showed contradicting results to the hypothesis and underlying theory, which is surprisingly. This would imply that in power distant countries, there would be higher gender egalitarianism than in low power distant countries. An explanation for the results could be the sample that was used. As argued that the approximately 86% of the sample consists of ‘low’ (<50) power distant societies (table 6). This may have biased the results. As theorized that in low power distant societies the effects of gender diversity would be greater in comparison to high power distant societies.
distant countries. However, the mediation effect does imply that the characteristics of high PD societies, such as maintaining or even increasing the distinction between social classes and societal members (Hofstede 1984; Inkeles 1960; Lenski 1966), and low gender egalitarianism (Gudykunst & Lee, 2003) are present. These behaviours may be rooted so strongly within culture, and are so dominantly present that females are not able, or do not gain the possibility to cover positions within boards.
Thirdly, applying the effect of the masculinity dimension on gender diversity and firm performance. There has been found that higher masculinity scores have a deterioration effect on the relationship between gender diversity and firm performance, even when the threshold of three or more females is reached. Suggesting that the national context in which a firm is settled has a significant effect on the group composition and social roles within a group.
Three different models (6-8) have been used to analyse the effect of masculinity. Masculinity represent the perspectives and values a given society has towards the emotional and social roles of different genders (Hofstede, 2001). Results show that as masculinity increases there is a negative effect on gender diversity and firm performance. Implying that the characteristics of a masculine society to maintain a clear distinction between the varying gender roles (Hofstede, 2001) are experienced within corporate boards. Feminine cultures experience higher levels of gender equality (Hofstede, 2001) and teamwork has higher reward levels (Hofstede, 1984) implying from the results that females are more accepted in boards of feminine countries.
Drawing further upon the theory of the masculinity dimension, which indicates that relationships and caring for others are important values within feminine societies (Hofstede, 1984; Jones & Teegen, 2001), and members of a society show empathy towards others regardless of the composition of a specific group (Hofstede, 2001). Verified by the results that indeed masculine societies have a more negative effect on females, and thus the minority group (Kanter, 1977a; 1977b). While there will be taken more care of them in feminine countries, in which gender egalitarianism (Hofstede, 2001) is more valued within these feminine societies. Continuing on the theory and results, this means that the benefits of gender diversity on boards will be exploited more within feminine societies compared to masculine societies. Due to the higher gender egalitarianism in feminine countries (Hofstede, 2001), greater acceptance of minority groups, and the greater impact of the ‘voices’ of females (Erkut, et al., 2008; Konrad, et al., 2008). All contributing to a greater decision-making process.
strong behaviours of maintaining differences between classes and members of society in the former (Hofstede 1984; Inkeles 1960; Lenski 1966), and distinct gender roles in the latter (Hofstede, 2001).
Establishing further reasoning upon cultures, the results of board national culture are considered. Different models (A1-A5) (Appendix E) have been used to analyse the effect of board national culture. Unfortunately, there is not considered whether a critical mass of a specific culture has been reached. In model 2, only the board cultures are considered. However, there are identical board culture values as national values. In the full model (A3) board national scores have higher significance levels than national cultural scores. This may imply that board nationality can be a substitute or complementary to national culture. Meaning that in high masculine societies, when a more feminine board culture is in place. These feminine traits, values and beliefs may be transferred through board members and integrated within the culture of the board. This may infer that the ‘narrow’ (board) culture may be more important than ‘wide’ (national) culture, and board culture may serve as a substitute for national culture. However, when the boards are considered, that are significantly different from the national values, there are varying or no significance levels.
Drawing upon the results of the difference between Tobins’ Q and ROA. There is a higher variation in the proportion explained for Tobins’ Q compared to ROA for the same models. Furthermore, in some of the models are higher significance levels for Tobins’ Q compared to ROA for some of the variables, as well as in the models with board nationality scores. The difference in the results between performance measures, may emanate from the nature of both measures. Tobins’ Q is a market-based measure, which shows the market value of a firm (Brown & Caylor, 2006). While, ROA is a financial measure, indicating the profitability of a firm (Campbell & Minguez-Vera, 2007). However, both are derived from a competitive advantage (March & Sutton, 1997).
As females have connections with different social networks than males, different socialization experiences (Simpson, Carter & D’Souza, 2010) and a greater understanding of the marketplace and corresponding stakeholders of the firm (Carter, Simkins & Simpson, 2003). Females possibly have greater positive effects on the market-based performance (Tobins’ Q) of a firm. The attraction of females and their socializing and networking capabilities may have more positive effects on the recognition and legitimacy of a firm. Therefore, enhancing the market-based performance more positively than the financial performance of that specific firm.
CONCLUSION
Secondly, the study contributes to the research field by addressing the importance of national culture of a country. It signifies the importance it has on gender egalitarianism, and social roles. Societies that are characterized by maintaining distinct gender roles, this negatively affects the benefits that arise from gender diversity, and group dynamics, even when the benchmark of three females is reached.
Furthermore, the thesis stretches on the importance of other board diversity characteristics. Due to the fact that, there are found significant effects of board national culture, directly on firm performance, as well as on the relationship between gender diversity and firm performance. This may imply that board national culture may serve as complementary to national culture.
For managerial implications this implies that, if the true objective is to capture the advantages of diversity and avoid tokenism, the benchmark of at least three females on a board should be reached. In addition to this, there has to be considered if gender diversity would be favourable in the specific country the firm is operative, or would it only invalidate the efficiency of the processes and dynamics of the board. Moreover, the importance of heterogeneous groups cannot be neglected. Managers should see the importance of the variation in their boards. The minority group has to be considered and the effects derived and forced on this specific group, as it impacts the performance of the firm. As this thesis stretches on the importance of diverse individuals within a board, there has to be notified that extremely diverse boards may trigger dysfunctionality in the group processes (Scwab, et al., 2016).
as the importance of board national culture is shown, this has to be analysed in further research, and consider the effects when a certain minority culture reaches a benchmark and its effects on the group dynamics. As collectivism and power distance are close related to each other, it may be that collectivism explains more about the effect on this relationship. For future research, the effect of collectivism can be examined. As well as, the other dimensions. However, these may be of less importance, as they explain less about group dynamics and roles.
Furthermore, in this thesis the average of the board nationality values was used for the measurements. In the perspective of the critical mass theory, and to fully understand the impact of board members, for future research there is proposed to consider the critical mass theory on board member nationalities (e.g. three or more Italians within a Dutch corporation), and their effect on the board culture. In addition to this, as the results have shown some significant results related to board nationality. Future researchers should more closely examine the effect of board national culture, therefore the joint effects of national culture and board national culture have to. Be considered. As the importance of national culture and board culture is shown, the overall organizational culture cannot be neglected. There has to be examined whether these are complementary or can serve as a substitution effect. Finally, there is proposed for future research to examine whether national culture has the same effect on foreign firms as they operate in a different setting, and their organizational culture or board culture may have greater impacts rather than the national culture.
REFERENCES
Adams, R. B., & Ferreira, D. 2009. Women in the boardroom and their impact on governance and performance. Journal of Financial Economics, 94(2): 291–309.
Allini, A., Rossi, F., & Hussainey, K. 2016. The board's role in risk disclosure: an exploratory study of Italian listed state-owned enterprises. Public Money & Management, 36(2): 113–120.
and firm value. Financial Review, 38(1): 33–53.
Asch, S. E. 1951. ‘Effects of Group Pressure upon the Modification and Distortion of Judgement’, in H. Guetzkow (Eds.), Groups, Leadership and Men: 177–190. Pittsburgh, PA: Carnegie Press.
Bantel, K. 1993. Strategic Clarity in Banking: Role of Top Management-Team Demography, Psychological Reports, 73(3): 1187–1203.
Bear, S., Rahman, N., & Post, C. 2010. The Impact of Board Diversity and Gender Composition on Corporate Social Responsibility and Firm Reputation. Journal of Business Ethics, 97(2): 207–222
Benson, J., & Yukongdi, V. 2005. Asian women managers: participation, barriers and future prospects. Asia Pacific Business Review, 11: 283–291.
Beugelsdijk, S., Kostova, T., & Roth, K. 2017. An overview of Hofstede-inspired country-level culture research in international business since 2006. Journal of International Business Studies, 48(1): 30-47.
Blumberg, B., Cooper, D. R., & Schindler, P. S. 2014. Business research methods. London: McGraw-Hill Education.
BoardEx. 2019. BoardEx Data. Retrieved from http://corp.boardex.com/data/. Accessed on March 30, 2019.
Bond, R. 2005. Group Size and Conformity. Group Processes and Intergroup Relations, 8(4): 331–354.
Breusch, T. S. & Pagan, A. R. 1978. A simple test for heteroskedasticity and random coefficient variation. Econometrica, 46(1): 1287-1294.
Brown, L. D., & Caylor, M. L. 2006. Corporate governance and firm valuation. Journal of accounting and public policy, 25(4): 409-434.
Bureau van Dijk. (2017). Orbis. Retrieved from https://www.bvdinfo.com/en-us/our-products/companyinformation/international-products/orbis. Accessed on December 12, 2018.
Campbell, K., & Minguez-Vera, A. 2008. Gender Diversity in the Boardroom and Firm Financial Performance. Journal of Business Ethics, 83(3): 435–451.
Carter, D., Simkins, B., & Simpson, W. 2003. Corporate governance, board diversity, Cassell, C. 2000. Managing Diversity in the New Millennium. Personnel Review, 29(3), 268–274
Caves, B. E., & Ghemawat, P. 1992. Identifying Mobility Barriers. Strategic Management Journal, 13(1): 1–12.
Central Intelligence Agency. 2019. GDP per capita. Retrieved from: https://www.cia.gov/library/publications/the-world-factbook/fields/211.html. Accessed on April 1, 2019.
