The Impact of Gender Role Attitudes on Female Board Participation

(1)

The Impact of Gender Role Attitudes on Female Board Participation

Laura M. van de Werfhorst* Supervisor: Dr. L. Dam

University of Groningen January 2015

Abstract

This study examines the relationship between gender role attitudes and board room gender diversity. To identify the impact of gender role attitudes on the proportion of women in the board of directors, a sample of 37 countries is used for the period 2002-2013. Societies with more traditional gender role attitudes are significantly related to less female board participation. Comparing the impact of women’s and men’s attitudes separately shows that the values of women in particular are important for explaining the representation of women in corporate boards. Women’s gender role attitudes have a direct effect on their own labor market behavior as well as a strong indirect effect on their families’ decisions to work, and therefore have a large effect on female board participation.

Keywords: Gender role attitudes, female board participation, board diversity, female labor force participation

JEL classification: J16, G34, M14

*

(2)

2

I. Introduction

In the past century women have increasingly been able to combine family and career, which is reflected in rising female labor force participation rates. However, women remain to be poorly represented in top management positions and corporate boardrooms. Although gender equality is about giving equal chances and not necessarily about obtaining equal results, the low female participation rates in corporate boards suggest that a glass ceiling still hinders women’s careers. As a first step to break this ceiling, it is important to fully understand the barriers to female board participation. This paper investigates the relationship between gender role attitudes and female board participation and finds that more traditional gender role attitudes have a significant and negative influence on the likelihood of observing female board directors.

Previous studies point out the importance of female board participation by showing that it is positively related to firm performance (Carter et al., 2003; Erhardt et al., 2003). In the short run, announcements of female board appointments have a positive impact on stock returns which indicates that investors on average believe that female directors add value. In the long run, female board participation leads to an improvement of firm value (Campbell and Vera, 2010). Despite these prosperous consequences, female board participation rates can still be considered low when compared to the overall percentage of women in the labor force. Figure 1 presents the development of board room composition for the countries that are part of the sample of this paper over the time period 2002 to 2013. Despite a slow but steady improvement in the past decade, the average female board participation rate still amounted only to 13.1% in 2013 which indicates that women remain highly underrepresented.

Since 1993, multiple countries such as Austria, Denmark and Greece have tried to increase female participation rates by implementing diversity quotas for state-owned companies. In 2003, Norway was the first country to introduce a diversity quota law, which required all listed companies to comply with a female board participation rate of 40%. More and more governments are now also considering gender quotas (Adams and Kirchmaier, 2013). For example, on November 14th, 2012 the European Commission proposed a legislation that sets the objective of 40% for female nonexecutive directors in publicly listed companies, with the exception of small and medium enterprises (European Commission, 2012).

(3)

8.12% 9.04% 7.93% 8.05% 8.76% 9.13% 9.18% 9.16% 9.24% 10.11% 11.21% 13.31% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Women Men

Figure 1: Average board composition for 37 countries over the period 2002-2013

memberships for existing female board members, instead of actually changing the number of female board members (Adams and Kirchmaier, 2013). To fully assess the consequences of the boardroom gender policies, a better understanding of the underlying causes of female board participation is important. Instead of using direct diversity quotas, addressing these underlying determinants might be more effective in increasing the number of female directors. Therefore, this study will further investigate the barriers for female representation in corporate boards. Previous literature finds that determinants exist of firm-specific characteristics, such as company size and type of industry (Grosvold et al., 2007; Harrigan, 1981). Also board-specific characteristics are found to be of significant influence, such as board size (Bianco et al., 2011). Finally, multiple country-specific variables are related to board diversity, such as the participation of women in higher education and the presence of a gender diversity quota law (Burgess and Tharenou, 2002; Sweigart, 2012). The effect of culture, and in particular the role of gender role attitudes, has been relatively unexplored in the literature on female board participation. In light of a growing attention to the importance of culture for economic outcomes (Alesina and Giuliano, 2010), this paper studies the impact of gender role attitudes on female board participation.

(4)

4 surveyed by the World Values Survey (WVS). Examples of these statements are “When jobs are scarce, men should have more right to a job than women” and “On the whole, men make better business executives than women do”. To assess the sensitivity of female board participation to gender role attitudes, the average response to the gender role statements is regressed on the proportion of female board members.

The sample that is used for this research consists of an unbalanced panel of 3,879 listed corporations in 37 countries worldwide: Australia, Brazil, Canada, Chile, China, Czech Republic, Egypt, Finland, France, Germany, Great Britain, Hong Kong, Hungary, Indonesia, Italy, Japan, Kuwait, Malaysia, Mexico, Morocco, Netherlands, New Zealand, Norway, Philippines, Poland, Qatar, Russia, Singapore, South Africa, South Korea, Spain, Sweden, Switzerland, Thailand, Turkey and United States. Data from these countries is obtained for a twelve-year time period from 2002 up to and including 2013, however the data on gender role attitudes is not available for all years.

The results show that in societies where gender role attitudes are more traditional, the probability of having female directors on the board is lower. More traditional societies believe that men are better equipped to perform well in a paid job than women, who should be dedicated to the household. Therefore, it is not surprising that women are not well-represented in their corporate boards. Not only the views on the differences in working capabilities between men and women are relevant in explaining female board participation. The items that consider the economic aspects of working women, such as “If a woman earns more money than her husband, it's almost certain to cause problems”, are of even larger influence. When testing the impact of women’s and men’s attitudes separately, it is found that for most of the gender role issues the responses of women have a significantly larger effect on female board representation. Women with more traditional gender role attitudes are likely to behave accordingly and these women will not ambition a board membership. Moreover, women’s attitudes can have an indirect influence on their families’ labor market behavior. It appears that in order to increase participation rates of women on boards, first there should be a transformation in the mindset of people such that they become receptive for working women. Especially, the gender role attitudes of women themselves are important as these will have the largest effect on labor market decisions.

(5)

5 female board participation. The main contribution of this paper is to improve the understanding of the underlying causes of female board participation. It indicates that implementing direct gender diversity quotas might not be an effective policy if no attention is paid to the underlying gender role attitudes.

The remainder of this paper is structured as follows. The next section will review past literature related to the determinants of female board participation. It mainly addresses papers that discuss gender role attitudes as the underlying cause of female labor force participation. Section 3 describes the econometric model and the methodology that are used to study the relationship between gender role attitudes and female board participation. Moreover, it discusses the relevant control variables. Section 4 describes the variable definitions and sources and provides the descriptive statistics of the dataset. Section 5 presents and discusses the estimation results. Finally, section 6 summarizes the main results and briefly concludes.

II. Literature on Cultural Aspects of Female Labor Force Participation

Although the impact of culture on female board participation has not yet been investigated intensively, the previous literature that discusses cultural aspects of women’s participation in the labor market might be closely related. Smith and Ward (1985) point out the long-term growth in female labor supply in the United States, as well as the more accelerated growth between 1950 and 1980. Subsequently, they examine the underlying causes of this increase in female labor force participation and find that about 60 percent of the growth is attributed to the combined effect of rising male and female wages that took place in this time period. However, half of the effect of wages actually works through declining fertility rates and therefore indirectly affects female labor supply. It appears that demand factors are not able to fully explain female participation in the labor force and Smith and Ward suggest that changes in gender role attitudes and increasing levels of schooling and social security might account for the remaining part.

(6)

6 hand, having a working mother might influence her son’s view on gender roles and shift his preferences for how work should be divided between a husband and wife towards more equality. As a result, he is more likely than other men to marry a working woman. On the other hand, being raised in a family with a working mother might demand more participation in the household from the son during his upbringing. This might make him more productive in the household, which allows his future wife to participate more in the labor market.

Fernández et al. (2004) test the intergenerational channel by making use of the variation across states in mobilization of men for the US army during World War II. States with high mobilization rates witnessed an increase in female labor supply as the women that were left behind needed to make a living for their families. Fernández et al. show that the jumps in female labor supply have a positive effect on the female labor force participation of succeeding generations. Furthermore, the authors find that across US states the ratio of average fertility of working women relative to the fertility of nonworking women is positively correlated with the female participation supply in following generations. Hence, Fernández et al. conclude that an increasing proportion of men being raised by a working mother positively affects women’s propensity to work.

The implications of this research are also tested for the Japanese economy by Kawaguchi and Miyazaki (2009). They find that men who are raised by full-time working mothers are more likely to disagree with statements that support traditional gender roles than men whose mothers did not work full-time. As a consequence,sons of working mothers are more likely to have working wives themselves. Several more papers confirm the influence of labor market behavior of mothers on succeeding generations (Fernández, 2007; Fernández and Fogli, 2005; Fogli and Veldkamp, 2007).

(7)

7 attitudes, stronger family ties societies are significantly related to less female labor force participation.

Farré and Vella (2013) further analyze the transmission of gender role attitudes across generations and their effect on labor market participation in the United States. They use survey data on a group of women who were aged between 39 and 47 years old and lived in the United States in 2004. The researchers construct an index of attitudes towards gender roles by aggregating the women’s responses to statements as “A woman’s place is in the home, not in the office or shop”. First, the women’s indices are significantly related to the corresponding indices that their children expressed beforehand in the 90’s and early 2000’s when they were aged 15 to 22 years old. This relationship holds even after controlling for economic and cultural background. This implies that there is indeed an intergenerational transmission of views regarding the role of women in the family and labor market. Second, Farré and Vella find evidence of a relationship between the respondents’ gender role attitudes and their own economic decisions as well as their families’ decisions to work. If women’s indices reflect a more traditional view on a woman’s role in the labor market, their probability to work decreases significantly. Moreover, the researchers find evidence that women who are married to men with traditional gender role attitudes are less likely to work. Hence, a mother’s view on the role of women in the labor market and family will not only influence the labor market decisions of her daughter, but also has a strong effect on the labor force participation of her daughter-in-law.

(8)

8 Adams and Kirchmaier examine whether increasing female labor force participation is sufficient to increase the proportion of female board members by using a dataset consisting of 9,888 listed companies in 22 countries over the period 2001-2010. The authors show that female board participation across countries is even worse than suggested in most surveys if you focus on the proportion of unique female directors and do not account for multiple directorships. Furthermore, they find that labor force participation indeed has a significant positive impact on the representation of women on boards provided that part-time and unemployed workers are excluded from labor force participation. However, they find that other economical and cultural factors also have a significant role in female board participation. For example, an increase in the level of government services to families, as measured by taxes and social security contributions over income, makes it easier for women to advance their career. Also, cultural norms and country-level policies targeting boardroom gender diversity appear to have a positive effect on the representation of women on boards. It is not clear whether the gender diversity policies can be effective on their own in addressing other causes of female underrepresentation. The multiple determinants of female board participation indicate that even when women are full-time employed, they may still face barriers to be promoted to board membership.

In sum, the literature suggests a significant relationship between gender role attitudes and female labor force participation. A shift from preferences for traditional gender roles towards preferences for more egalitarian division of the responsibilities in work and household is proven to have a positive impact on female labor supply. However, the impact of gender role attitudes on female board participation in particular remains unexplored in the literature. As the determinants of female labor force participation are likely to be similar to the determinants of female board participation, a negative relationship between more traditional gender role attitudes and female board representation can be expected.

III. Methodology

To assess the impact of gender role attitudes on the proportion of women in director positions, female board participation is estimated as

(9)

9 The dependent variable, , is female board participation, which measures the proportion of women that are part of the board. are the main explanatory variables that indicate the gender role attitudes that dominate in a country. Therefore, the coefficient is of particular interest in this study as it measures the sensitivity of female board participation to gender role attitudes. is a vector of firm-specific control variables that exists of board size, return on equity, firm age, market share, and dichotomous variables indicating the presence of blockholder ownership and a gender policy. is a vector of country-specific variables that includes the share of total gross national income, tax revenues, fertility rate, ratio of female to male enrollments in tertiary schooling, female labor force participation and dichotomous variables indicating the presence of a codetermination law and a quota law. Finally, gives a vector of industry dummies that captures the main cross-sectional differences in the sample. As the main variables, the gender role attitudes, are relatively time invariant, it is unlikely that significant results are found by regressing over the years. Instead, most information can be gathered from the cross-sectional variation in the data, therefore no firm-fixed effects are included in the estimation. Not including fixed effects implies that there might be country-specific characteristics, besides the variables that are controlled for, that bias the estimations. Therefore, the results should be carefully interpreted. The dependent variable, the percentage of females on boards, is continuous but constrained between zero and 100 percent. It is expected to be zero for a large part of the population and strictly positive otherwise. Therefore, it is appropriate to estimate the equation through maximum likelihood estimation by using a Tobit analysis (Verbeek, 2008). A Tobit analysis can describe both the probability that the dependent variable is zero, as well as the distribution of the dependent variable given that is it positive. As such, the Tobit analysis allows drawing inference on both the effect of gender role attitudes on the occurrence of having at least one woman on the board, as well as the size of these effects in terms of the percentage of women on the board (Hill, Griffiths and Lim, 2008).

(10)

10 “If a woman earns more money than her husband, it's almost certain to cause problems”. The estimation coefficients provide insight in the sensitivity of female board participation to each of these specific items.

Furthermore, a principal component analysis is performed to transform the seven individual statements into one factor that describes aggregate gender role attitudes. The principal component analysis identifies dominant patterns in the data and reduces the number of dimensions while retaining as much as variation in the dataset as possible. It is a multivariate technique that transforms the (possibly correlated) variables into a new set of uncorrelated variables; the principal components. The latter components are ordered in such a way that the first principal component captures most of the variation that was present in the original variables, the next component retains as much of the remaining variation possible and so on (Joliffe, 2002). These principal components are subsequently aggregated into one factor such that it can be tested whether the traditional gender role attitudes together are important for explaining the presence of female directors. Furthermore, this is a better approach than regressing all individual statements together in one specification as it is likely that the responses to the individual statements are highly correlated. The principal components analysis enables the study of their joint significance, while circumventing problems of multicollinearity.

In addition, the difference in impact between women’s and men’s attitudes is investigated by estimating the separate effects of their responses to the gender role statements. The opinions of men and women are likely to be aligned, because of the generational transmission of gender role attitudes. However, their relative impact on female board participation might differ as the mothers’ attitudes have a direct effect on their own labor market behavior as well as an indirect effect on their entire families’ labor market choices (Farré and Vella, 2013). Therefore, women’s gender role attitudes have a two-way effect on female board participation and might be of larger influence than men’s attitudes.

(11)

11 least one woman on the board. As a result, the estimated coefficients give the effect of the explanatory variables on the likelihood of having at least one woman on the board.

Relevant control variables

Adams and Kirchmaier (2013) do not regress female board participation on individual items from the World Values Survey (WVS), but instead use more general country-level value measures in order to grasp the culture effect. They claim that the advantage of using these scores rather than the individual items is that they account for multiple dimensions of culture and that their country-level coverage is more complete. Furthermore, they suggest that if respondents are unsure about agreeing with the statements on gender role attitudes, their answers might just reflect the labor force participation rates that arise in practice. For example, if respondents do not know how to respond to the statement ‘When jobs are scarce, men should have more right to a job than women’, they may assume that they should simply agree because of the low female labor force participation rates that are present in their country. Hence, Adams and Kirchmaier claim that individual WVS responses could be reflecting labor force participation rates by men and women rather than intrinsic values. However, this study circumvents this potential problem by controlling for female labor force participation rates. In line with the conclusions of Adams and Kirchmaier, it is expected that female labor force participation has a positive impact on the representation of women on boards.

Other control variables are selected by reviewing literature that examines determinants of female board participation. Several firm-specific characteristics are included as controls, the first of which is board size. Larger boards have more seats available to fill and therefore offer more opportunity for diversity (Conyon and Mallin, 1997). Burke (1999) indeed finds that bigger boards consist of more women.

The second firm-specific control variable is return on equity. Ryan and Haslam (2007) find that women are more likely to be appointed as a board member if firms have experienced a weak performance in the months before. The authors argue that the inclusion of women on boards might symbolize a change to firms’ stakeholders and a first step towards a better performance. In addition, women are more likely to possess characteristics that are considered to be especially important in times of bad performance, such as being understanding, intuitive and creative.

(12)

12 Amburgey, 1991; Meyer and Brown, 1977). In accordance with this, Baron et al. (1991) find that older organizations are less likely to have women in management positions.

Fourth, company size is used as a control variable which is measured by market share. Bianco et al. (2011) argue that larger firms are subject to a higher degree of market scrutiny which imposes social expectations for gender diversity. Empirical evidence for this positive relationship between company size and the diversity of boards is given by Grosvold et al. (2007).

Another determinant of female board participation is shareholder concentration. Kang et al. (2007) argue that more dispersed ownership structures imply that shares are held by more diverse shareholder groups, which in turn will increase the demand for a diverse board composition. Indeed, they find that more shareholder concentration results in fewer female board members.

Furthermore, equation (1) is controlled for the presence of a gender policy within the organization. If the company designs and executes a policy of a minimum level of diversity within the firm, its board can be expected to be more gender diverse as well. Similarly, a quota law that is implemented by the government is likely to have a positive impact on the female board participation as these laws in fact oblige companies to conform to a certain level of female representation in the board. An analysis by Sweigart (2012) indeed shows evidence of increases in female board participation as a consequence of the presence of quota laws.

Adams and Kirchmaier (2013) also control for the presence of co-determination laws in the countries, as they argue that women are more likely to care about employee welfare than men and therefore are more likely to be part of the board as employee representatives.

Several more country-specific control variables should be considered in this research, such as gross national income. Increasing gross national income indicates economic development, which might have a positive influence on female board representation (Adams and Kirchmaier, 2013). Furthermore, Adams and Kirchmaier find a positive impact of both tax contributions over income and fertility rates as these are indicators of how family-friendly countries are. In a country with a high degree of family-friendliness, it is easier for women to combine work with their family such that they do not need to drop out of the labor force.

(13)

13 The final aspect that should be considered in the economic model of female board participation is the influence of the industry in which a company operates. Harrigan (1981) concludes that women are especially well-represented in the corporate boards of the process-, technologies-, services- and financial industry. Also, Hyland and Marcellino (2002) find evidence of an unequal participation of women in top functions across industries. Their results indicate that women are most likely to possess board positions in companies that operate in the financial-, service-, real estate or insurance sector. To take this influence of industries into account, industry dummies will be included in the estimations.

IV. Data

To identify what determines female board participation, this research makes use of a panel dataset that consists of 3,879 listed corporations for the time period of 2002 up to and including 2013. The firms are based in 37 different countries worldwide, in an attempt to keep the number of cross-sections as large as possible. Table I in the Appendix presents an overview of the number of firms included for each country. It shows that the main share of the firms is based in Anglo Saxon countries, as 52% of the firms are originating from the United States, United Kingdom, Canada and Australia. US corporations in particular form a large part of the dataset (26.4%).

Table 1 provides definitions of the dependent-, explanatory- and control variables. The dependent variable of this study is female board participation. Information on the proportion of female executives on the board is obtained from the Asset4 database. This data can be easily transformed into a dichotomous variable indicating the presence of at least one woman in the board room.

(14)

Table 1. Definition and source of the variables

Definition Source

Percentage of females The number of female board members divided by the total number of executives of which the board comprises.

Asset4 Presence of females Dummy variable that equals 1 when the firm has at least one female board member

and 0 otherwise.

Asset4

Gender role attitudes:

Jobs scarce The proportion of people within a country that have agreed with the following

proposition: “When jobs are scarce, men should have more right to a job than women”.

World Values Survey Being housewife is fulfilling The proportion of people within a country that have (strongly) agreed with the

following proposition: “Being a housewife is just as fulfilling as working for pay”.

World Values Survey University more important for

boy

The proportion of people within a country that have (strongly) agreed with the following proposition: “A university education is more important for a boy than for a girl”.

World Values Survey Child suffers with working

mother

The proportion of people within a country that have (strongly) agreed with the following proposition: “When a mother works for pay, the children suffer”.

World Values Survey Job not best way for women

to be independent

The proportion of people within a country that have disagreed with the following

proposition: “Having a job is the best way for a woman to be an independent person”. World Values Survey Problem if women have more

income than men

The proportion of people within a country that have agreed with the following proposition: “If a woman earns more money than her husband, it's almost certain to cause problems”.

World Values Survey Men make better business

executives

The proportion of people within a country that have (strongly) agreed with the following proposition: “On the whole, men make better business executives than women do”.

World Values Survey

Firm-specific control variables:

Board size The number of executives that are part of the board. Asset4 Return on equity Measure of the profitability of the corporation. It gives the net income (minus the

bottom line and preferred dividend requirement) as a percentage of the average of last year’s and current year’s common equity.

Worldscope Firm age The amount of years that passed since the year of incorporation till 2014. Therefore it

is not time variant.

Worldscope Market share Measure of the size of the corporation. The market value is the share price multiplied

by the number of ordinary shares in issue which is updated whenever new tranches of stock are issued or after a capital change, expressed in billions of US Dollar. Each firm’s market value is then divided by the total of market values of all firms included in this dataset for that particular year.

Datastream

Blockholder ownership Dummy variable that equals 1 if the corporation is owned by a reference shareholder who has the majority of the voting rights, veto power or golden share.

Asset4 Company has gender policy Dummy variable that equals 1 if the corporation has a policy regarding the gender

diversity of its board and 0 otherwise.

Asset4

Country-specific control variables:

Share gross national income Measure of economic development. Gross National Income (GNI) is the sum of value added by all resident producers plus any product taxes (less subsidies) not included in the valuation of output plus net receipts of primary income from abroad. The GNI is divided by the midyear population and expressed in current thousands of US Dollar. Each country’s GNI is divided then by the total GNI of all country’s added for that particular year.

World Bank

Tax revenues (% of GDP) The amount of compulsory transfers to the central government for public purposes as a percentage of the Gross Domestic Product (GDP)

World Bank Fertility rate The average number of children that would be born to a woman if she were to live to

the end of her childbearing years and bear children in accordance with current age-specific fertility rates.

World Bank Female/male tertiary school

enrollment

Measure of women in higher education. It gives the percentage of women to men enrolled at tertiary level in public and private schools.

World Bank Female labor force

participation

The proportion of the female population that are aged 15 and older that is economically active.

World Bank Codetermination law Dummy variable that equals 1 if the country has any type of codetermination laws that

involve board-level employee representation and 0 otherwise.

Adams & Kirchmaier (2013)

Quota law Dummy variable that equals 1 if the country has any type of mandatory quota for listed companies targeting the diversity on boards and 0 otherwise.

(15)

15 characteristics of the interviewee, so that it is possible to compare the results of women’s opinions with that of men’s. If the variable sex is reported as ’unknown’ or ‘missing’, the observations are omitted from the research.

The firm-specific control variables consist of the board size, a dummy for blockholder ownership and a dummy that represents whether the company has a gender diversity policy, which are obtained from the Asset4 database. Other controls are return on equity and the firm age, as provided by Worldscope, and market value, for which the Datastream database is used. Moreover, gross national income, tax revenues, fertility rate, ratio of women to men in tertiary school enrollments and female labor force participation are part of the country-specific control variables and are obtained from the World Bank Organization. Table 2 presents the descriptive statistics of the main variables. The descriptive statistics of the gender role attitudes of men and women separately can be found in Table II in the Appendix.

Table 2. Descriptive statistics of variables for the 3879 firms worldwide (2002-2013)

Mean Stdev Min Max Obs

Percentage of females 9.765 10.453 0.000 66.670 26,718

Presence of females 0.586 0.493 0 1 26,718

Gender role attitudes:

Jobs Scarce 0.180 0.135 0.021 0.880 6,316

Being housewife is fulfilling 0.699 0.141 0.179 0.920 6,316

University more important for boy 0.127 0.088 0.011 0.462 6,316

Child suffers with working mother 0.266 0.112 0.150 0.791 2,468

Job not best way for women to be independent 0.490 0.072 0.254 0.584 2,468 Problem if women have more income than men 0.135 0.070 0.047 0.466 2,468 Men make better business executives 0.220 0.126 0.080 0.850 6,094

Firm-specific control variables:

Board size 10.517 3.759 1 45 26,694

Return on equity 12.714 19.049 -98.790 98.700 40,747

Firm age 38.482 31.141 2 187 25,548

Market share 0.0003 0.0007 1.43e-09 0.019 41,396

Blockholder ownership 0.140 0.347 0 1 26,842

Company has gender policy 0.149 0.357 0 1 26,842

Share gross national income 0.041 0.019 0.001 0.103 46,449

Tax revenues (% of GDP) 15.505 6.538 8.214 32.374 38,605

Fertility rate 1.797 0.410 0.901 3.709 42,669

Female/Male tertiary school enrollment 120.938 25.710 62.082 676.162 35,194 Female labor force participation 54.143 7.075 18.300 69.000 42,669

Codetermination law 0.062 0.241 0 1 46,548

Quota law 0.064 0.245 0 1 46,548

(16)

16 The summary statistics show that the average proportion of women on the board over the entire sample is approximately 9.8%. Also, for approximately 59% of the observations there is at least one woman participating on the board. Overall, the average agreement with the traditional gender role statements is not that high, except for statements that can be considered to be less sensitive such as “Being housewife is fulfilling” and “Job not best way for women to be independent”. Furthermore, the attitudes of men and women appear to be quite aligned.

To prevent extreme values of the variables from biasing the results, outliers are removed from the dataset. As a consequence, market values of above 400 billion US dollars are omitted and the return on equity values are cut off if below -99% and above 99%. In addition, since the tax revenues as a percentage of GDP are extremely low in Kuwait, its observations are omitted from this research.

Table 3 presents the pair wise correlation matrix for all variables. The results indicate that some of the correlations between the statements on gender role attitudes are high (e.g. the correlations between ‘Men make better business executives’ and ‘Jobs scarce’ and ‘University more important for boy’ are 0.90 and 0.92 respectively). Therefore, regressing these statements together in one specification will likely result in a multicollinearity problem. As described in the previous section, a factor analysis could test whether gender role attitudes as a whole are important for female board participation while being able to circumvent multicollinearity issues.

The correlations indicate that more traditional views on gender roles (so higher proportional agreements with the statements) are associated with lower female board participation rates. So the negative correlations are in accordance with the expectations that followed from the literature, although some correlations appear to be small. Moreover, the signs of the correlations between the control variables are all in line with the hypotheses except for the corporations’ return on equity and age.

(17)

17

Table 3. Pair wise correlation matrix

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [1] Percentage of females 1

[2] Presence of females 0.79 1

[3] Jobs scarce -0.40 -0.43 1

[4] Being housewife -0.06 -0.02 -0.20 1

[5] University for boy -0.36 -0.38 0.86 -0.05 1

(18)

18

V. Results

The Tobit regression estimations of the gender role attitude variables on the percentage of women that are part of the corporate board are presented in Table 4. In column (1) to (7) the seven attitude variables are in turn added to the specification. The final column reports the significance of the factor that aggregates all seven items in explaining female board participation. The results show that each of the individual gender role attitude variables are significant and all, except for one, are negatively related to female board participation. So, the results are largely in accordance with the hypotheses. More specifically, the proportional agreement with the following WVS statements has a negative impact on the proportion of female directors: “When jobs are scarce, men should have more right to a job than women”, “Being a housewife is just as fulfilling as working for pay”, “A university education is more important for a boy than for a girl”, and “On the whole, men make better business executives than women do”. These results indicate that the probability of having female directors is lower in societies where attitudes on the working capabilities of men and women are more traditional, in the sense that people feel that men should be working in a paid job while women do the housework. The items that consider the economic aspects of working women have a similar negative effect. The proportional agreement with “Having a job is not the best way for a woman to be an independent person” and “If a woman earns more money than her husband, it's almost certain to cause problems” is negatively related to female board participation. Furthermore, the factor that aggregates all seven individual items has a significant negative influence. This proofs that the gender role attitudes together constitute an important determinant for female board participation. If a society has a traditional view on gender role attitudes, it is believed that men are better equipped to perform well in a paid job than women and therefore it is not surprising that women are not well-represented in corporate boards.

(19)

19

Table 4. The effect of gender role attitudes on female board participation

Table continued on next page

Dependent variable: percentage of females of total board members

(1) (2) (3) (4) (5) (6) (7) (8)

Jobs scarce -68.510*** (11.463) Being housewife is fulfilling -16.449*** (5.534) University more important for boy -76.244*** (13.275) Child suffers with working mother 51.077*** (16.536) Job not best way

for women to be independent -48.058* (26.586) Problem if women have more income than men -166.734*** (43.005) Men make better business executives -40.925*** (8.139) Gender-role attitudes factor -7.183*** (2.262)

Board size 0.662*** 0.704*** 0.726*** 0.866*** 0.867*** 0.837*** 0.674*** 0.813*** (0.114) (0.115) (0.115) (0.161) (0.162) (0.161) (0.117) (0.162) Return on equity 0.007 0.007 0.009 0.011 0.011 0.011 0.005 0.013 (0.016) (0.016) (0.016) (0.021) (0.021) (0.021) (0.016) (0.021) Firm age 0.041*** 0.040*** 0.040*** 0.061*** 0.056*** 0.058*** 0.041*** 0.057*** (0.009) (0.009) (0.009) (0.012) (0.012) (0.012) (0.009) (0.012) Market share 1021.771*** 984.471*** 1042.807*** 1170.722** 1196.904*** 1205.265*** 1021.494*** 1230.149*** (288.456) (289.779) (288.900) (458.862) (459.976) (457.746) (289.526) (459.240) Blockholder ownership -0.843 -1.193 -0.858 -2.181 -2.265 -1.656 -1.011 -1.796 (1.260) (1.264) (1.260) (1.482) (1.498) (1.490) (1.287) (1.497) Company has gender policy 0.624 1.095* 0.529 0.542 0.638 0.572 0.656 0.602 (0.628) (0.626) (0.633) (0.733) (0.735) (0.731) (0.631) (0.733)

(20)

20

Table 4. Continued

This table presents the estimation results of Tobit regressions with inclusion of industry dummies which control for the main firm-specific characteristics that might bias the results. The dependent variable for all specifications in this table is the number of female directors as a percentage of the total number of board members. The explanatory variables consist of the proportion of people that agreed with the corresponding statements concerning gender role attitudes as well as both firm-specific and country-specific control variables. Variables definitions, measurement units and sources are as shown in Table 1. The dataset consists of 3879 firms that are based in Australia, Brazil, Canada, Chile, China, Czech Republic, Egypt, Finland, France, Germany, Great Britain, Hong Kong, Hungary, Indonesia, Italy, Japan, Kuwait, Malaysia, Mexico, Morocco, Netherlands, New Zealand, Norway, Philippines, Poland, Qatar, Russia, Singapore, South Africa, South Korea, Spain, Sweden, Switzerland, Thailand, Turkey and United States over the time period 2002-2013. Availability of data varies over the years and countries. Standard errors are shown in parentheses. Asterisks indicate significance at the 0.01 (***), 0.05 (**) and 0.1(*) levels.

participation. However, all effects are quite large as the smallest coefficient for “Being a housewife is fulfilling” still implies that a 10% increase in agreement with this statement, ceteris paribus, decreases the expected female board participation by 1.6%.

More surprisingly, the results show that the proportional agreement with “When a mother works for pay, the children suffer” has a significant and positive effect on the probability of female board representation. An increase of 10% in people that support this statement, ceteris paribus, increases the expected female participation rate by approximately 5.1%. There are two potential mechanisms that can explain this result. First, women who believe that their families suffers from their absence might only be willing to leave the household if they expect to get a top function. Since these women anticipate that their children are affected if they work, they only find it worthwhile to invest their time in the labor market if they can get a high-paid job. Therefore, only highly motivated women with good potential remain active in the labor force such that their competences can be easily recognized and they are more likely to be appointed as a board member. Second, the statement might also have an effect on female board participation via the quality of childcare. If a society largely feels that children suffer when mothers leave the household in order to work, it might invest more in childcare services to reduce the harm. As the overall quality of childcare services Codetermination law -0.339 -1.258 0.631 7.618 10.577 13.271** 1.069 12.194* (3.484) (3.635) (3.461) (6.532) (6.520) (6.536) (3.467) (6.541) Quota law 6.067* 5.171 5.514* 10.405 -4.364 -1.805 5.880* 0.109 (3.362) (3.396) (3.346) (7.562) (8.233) (7.225) (3.326) (7.203) Constant -55.537*** -109.826*** -75.145*** -128.964*** -93.644*** -2.193 -90.086*** 10.016 (14.588) (10.684) (12.709) (19.619) (14.816) (27.731) (11.642) (35.536) Industry

dummies Yes Yes Yes Yes Yes Yes Yes Yes

Nr. of

observations 2043 2043 2043 1161 1161 1161 2001 1161

(21)

21 increases, more women might feel comfortable to take on a job and leave their children at daycare. As a consequence, more women are available for board functions and therefore female board participation increases.

With respect to the estimations for the control variables, Table 4 indicates that most signs are in accordance to the findings of previous literature. However, the positive impact of company age is unexpected and in contrast to the results of other researchers (Baron et al., 1991). It appears that older organizations are more likely to include women in their boards. Older firms might be better known by the public and therefore receive more market scrutiny. This can impose social expectations for female representation on these firms.

Other surprising results are the negative and significant coefficients for GNI (% of total GNI of all countries). It appears that the larger the share in total GNI, the less women participate on the boards. In wealthier countries with higher incomes per inhabitant, there might be less need for families to have a dual income. Families can live of one income, such that there is no urge for women to build a career for themselves.

The final unexpected finding for the control variables is the significant negative impact of tax revenues (% of GDP) on female board participation. As Adams and Kirchmaier (2013) suggest that higher tax revenues imply a higher degree of family-friendliness, it was expected to be of positive influence on female board representation. However, the negative effect might be caused by the fact that a higher taxation of labor discourages work. Countries with high tax revenues as a percentage of GDP are likely charging high income taxes. Since high payroll taxes lower the payoff of additional work, it is less attractive for women to work if their family can also make a living of their husbands income. Therefore, in countries with high tax revenues women might be less inclined to proceed their careers and female board participation is less likely.

Men and women’s separate influences

(22)

22 To assess the statistical significance of the differences in their effects, the difference between the female and male proportion of agreement is used as an explanatory variable as well. The results for these difference variables are reported in Table IV of the Appendix. A significant coefficient estimation for the difference variables indicates that there is a significant difference between the impact of women’s and men’s gender role attitudes on female board participation.

The results in column (1) of Table 5 and Table 6 show that agreement with “Jobs scarce” by both men and women has a negative effect on female board participation. Although the difference in impact does not seem large, the coefficient is significantly higher for men. Traditional views among male inhabitants about the right on a job in case of a scarcity have a stronger negative effect on female board participation than the views among females. In most businesses men are mainly responsible for the appointment of the board of directors, as top executive functions are dominated by men (the so-called ‘old boys network’). This might be the reason that men’s opinions about women’s right on a job are more important for the resulting board composition.

Columns (2) of both tables support the previous findings of a negative relationship between the proportion of men and women that agrees with “Being housewife is fulfilling” and the percentage of female board directors. However, the size of their impact is significantly different. In particular, the attitudes of women towards being a housewife are of a larger impact on female board participation than those of men. If women themselves are satisfied with staying at home and doing the housework, they are not inclined to build a career and are unlikely to become a board member.

Similarly, column (3) confirms that the proportion of both men and women that agrees with “University more important for boy” is negatively related with the proportion of women in board positions. Again, women’s attitudes are of a significantly larger impact. If women themselves believe that they will profit less from university than men, it is not likely that they will enroll themselves in higher education. As a consequence, women might be less represented in top positions such as board memberships.

(23)

23

Table 5. The effect of gender role attitudes (expressed by women only) on female board participation

(1) (2) (3) (4) (5) (6) (7) (8)

Female gender role attitudes:

Jobs scarce -62.216*** (12.102) Being housewife is fulfilling -20.161*** (5.395) University more important for boy -93.525*** (18.731) Child suffers with working mother 33.118** (15.661) Job not best way

for women to be independent -61.218** (28.120) Problem if women have more income than men -187.230*** (53.302) Men make better

business executives -41.693*** (9.106) Female gender role attitudes factor -8.866*** (2.890)

Board size 0.635*** 0.732*** 0.701*** 0.853*** 0.881*** 0.856*** 0.668*** 0.820*** (0.114) (0.116) (0.115) (0.161) (0.162) (0.161) (0.117) (0.162) Return on equity 0.004 0.008 0.005 0.011 0.011 0.011 0.003 0.013 (0.016) (0.016) (0.016) (0.021) (0.021) (0.021) (0.016) (0.021) Firm age 0.041*** 0.040*** 0.039*** 0.060*** 0.056*** 0.058*** 0.041*** 0.056*** (0.009) (0.009) (0.009) (0.012) (0.012) (0.012) (0.009) (0.012) Market share 997.310*** 988.122*** 1023.668*** 1181.045** 1189.824*** 1191.107*** 1011.745*** 1233.748*** (289.127) (289.294) (289.306) (459.801) (459.535) (457.936) (289.774) (459.350) Blockholder ownership -0.817 -1.298 -0.894 -2.113 -2.284 -1.718 -1.045 -1.930 (1.262) (1.263) (1.263) (1.486) (1.494) (1.488) (1.289) (1.497) Company has gender policy 0.891 1.028 0.783 0.592 0.633 0.595 0.765 0.601 (0.624) (0.625) (0.629) (0.735) (0.734) (0.732) (0.630) (0.734)

(24)

24

This table presents the estimation results of Tobit regressions with inclusion of industry dummies which control for the main firm-specific characteristics that might bias the results. The dependent variable for all specifications in this table is the number of female directors as a percentage of the total number of board members. The explanatory variables consist of the proportion of women that agreed with the corresponding statements concerning gender role attitudes as well as both firm-specific and country-specific control variables. Variables definitions, measurement units and sources are as shown in Table 1. The dataset consists of 3879 firms that are based in Australia, Brazil, Canada, Chile, China, Czech Republic, Egypt, Finland, France, Germany, Great Britain, Hong Kong, Hungary, Indonesia, Italy, Japan, Kuwait, Malaysia, Mexico, Morocco, Netherlands, New Zealand, Norway, Philippines, Poland, Qatar, Russia, Singapore, South Africa, South Korea, Spain, Sweden, Switzerland, Thailand, Turkey and United States over the time period 2002-2013. Availability of data varies over the years and countries. Standard errors are shown in parentheses. Asterisks indicate significance at the 0.01 (***), 0.05 (**) and 0.1(*) levels.

via the quality of child care, is still possible as men are highly represented in the political arena and therefore can have a large effect on child care services.

Whereas table 4 presented a significant impact of the aggregate attitudes towards “Job not best way for women to be independent” on female board participation, the separate regressions for female and male attitudes show that only female responses are of significant influence. Column (5) of Table 5 indicates that if the proportion of women that agrees with this statement increases, the likelihood of female board representation decreases. On the other hand, no significant relationship is found for the values of men. So, the attitudes of men concerning this issue are less relevant. If women believe that they do not require a job to be independent, they are more likely to resign from the labor force. As such, the likelihood of a board membership also decreases.

Subsequently, the findings in columns (6) indicate that both male and female values with respect to “Problem if women have more income than men” are of negative influence on female board participation. The effect of women’s opinions is again significantly larger than that of men. If women believe it will be a problem if they earn more than their husband, it is likely that they will try to prevent this by not striving for the highest functions and settle in Codetermination law -0.531 -2.005 0.330 8.604 9.743 19.202*** 1.410 14.852** (3.520) (3.625) (3.481) (6.532) (6.502) (7.002) (3.476) (6.709) Quota law 7.017** 4.266 6.131* 7.548 -4.595 -8.190 6.231* -2.800 (3.402) (3.410) (3.362) (7.487) (7.953) (7.793) (3.341) (7.352) Constant -64.008*** -107.014*** -69.979*** -119.476*** -95.904*** -9.625 -90.917*** 36.494 (14.589) (10.694) (14.037) (19.649) (14.814) (27.982) (11.763) (44.411) Industry

Nr. of

observations 2043 2043 2043 1161 1161 1161 2001 1161

(25)

25

Table 6. The effect of gender role attitudes (expressed by men only) on female board participation

(1) (2) (3) (4) (5) (6) (7) (8)

Male gender role attitudes:

Jobs scarce -62.984*** (10.258) Being housewife is fulfilling -9.453* (5.328) University more important for boy -56.876*** (9.896) Child suffers with working mother 70.091*** (17.074) Job not best way

for women to be independent -26.141 (22.355) Problem if women have more income than men -119.834*** (32.355) Men make better

business executives

-34.027*** (7.005) Male gender role

attitudes factor

-5.713*** (1.829)

Board size 0.688*** 0.677*** 0.733*** 0.893*** 0.863*** 0.833*** 0.661*** 0.824*** (0.114) (0.115) (0.115) (0.161) (0.162) (0.161) (0.117) (0.162) Return on equity 0.009 0.005 0.010 0.009 0.011 0.012 0.005 0.013 (0.016) (0.016) (0.016) (0.020) (0.021) (0.021) (0.016) (0.021) Firm age 0.041*** 0.040*** 0.040*** 0.062*** 0.057*** 0.058*** 0.042*** 0.057*** (0.009) (0.009) (0.009) (0.012) (0.012) (0.012) (0.009) (0.012) Market share 1034.036*** 974.073*** 1040.939*** 1154.836** 1195.582*** 1209.649*** 1029.869*** 1221.171*** (288.241) (290.283) (288.974) (456.994) (460.000) (457.818) (289.601) (458.863) Blockholder ownership -0.895 -1.065 -0.851 -2.248 -2.163 -1.654 -1.045 -1.746 (1.259) (1.266) (1.260) (1.477) (1.498) (1.491) (1.287) (1.495) Company has gender policy 0.486 1.215* 0.474 0.477 0.647 0.573 0.705 0.604 (0.630) (0.627) (0.635) (0.731) (0.735) (0.732) (0.631) (0.733)

(26)

26

This table presents the estimation results of Tobit regressions with inclusion of industry dummies which control for the main firm-specific characteristics that might bias the results. The dependent variable for all specifications in this table is the number of female directors as a percentage of the total number of board members. The explanatory variables consist of the proportion of men that agreed with the corresponding statements concerning gender role attitudes as well as both firm-specific and country-specific control variables. Variables definitions, measurement units and sources are as shown in Table 1. The dataset consists of 3879 firms that are based in Australia, Brazil, Canada, Chile, China, Czech Republic, Egypt, Finland, France, Germany, Great Britain, Hong Kong, Hungary, Indonesia, Italy, Japan, Kuwait, Malaysia, Mexico, Morocco, Netherlands, New Zealand, Norway, Philippines, Poland, Qatar, Russia, Singapore, South Africa, South Korea, Spain, Sweden, Switzerland, Thailand, Turkey and United States over the time period 2002-2013. Availability of data varies over the years and countries. Standard errors are shown in parentheses. Asterisks indicate significance at the 0.01 (***), 0.05 (**) and 0.1(*) levels.

simpler jobs or focus on the household. This way, women’s attitudes towards the economic aspects of working might be a barrier for female board participation.

Column (7) shows that the proportion of both men and women that agrees with “Men make better business executives” is negatively related to female board participation. However, Table VI in the Appendix indicates a significant difference between the impact of female and male attitudes. Again, women’s values concerning this statement have a larger influence on the likelihood of having female board members. If women themselves believe they are less compatible with being a business executive than men, they are likely to feel the same about similar kind of top functions. Therefore, these women are unlikely to apply for board memberships and are not well-represented on boards.

Finally, column (8) of table 5 reports that the factor that aggregates women’s responses to all seven gender role statements is significantly and negatively related to female board participation. Similarly, the men’s gender role attitude factor has a negative impact on female board membership. Again, the impact of women is significantly larger than that of men.

In general, the results show that women’s attitudes are most important in explaining female board participation. The opinions of women concerning five of the gender role issues have a larger effect than the views of men. Also, the influence of the female gender role factor Codetermination law 1.243 0.285 2.052 3.296 6.597 5.073 -2.329 4.516 (3.454) (3.599) (3.459) (6.516) (6.572) (6.508) (3.966) (6.531) Quota law 4.232 5.773* 4.264 16.901** 2.158 7.186 7.821** 7.321 (3.363) (3.398) (3.366) (7.507) (8.268) (7.098) (3.626) (7.124) Constant -58.001*** -113.359*** -83.410*** -134.540*** -93.604*** -14.720 -93.887*** -15.730 (14.170) (10.657) (11.961) (19.098) (14.801) (25.970) (11.406) (28.925) Industry

Nr. of

observations 2043 2043 2043 1161 1161 1161 1998 1161

(27)

27 is larger which indicates that overall women’s gender role attitudes are of larger impact on female board participation than men’s. If women themselves have more traditional gender role attitudes, they are likely to behave in accordance with these patterns. This implies that women with more traditional views will be less likely striving for an ambitious working career and therefore are also unlikely to become a board director. Women’s gender role attitudes have a direct effect on their own behavior and effort to obtain top functions. As a consequence, women’s opinions will have a large impact on female board participation.

Furthermore, Farré and Vella (2013) indicate that women’s views have an indirect influence on the decisions to work of their daughters and daughters in law. So besides influencing their own labor market behavior, women’s gender role attitudes are also able to impact their families’ behavior. This two-way effect of female gender role attitudes might explain their large effect on female board participation.

Robustness check

To check the robustness of the findings, the model in equation (1) is estimated by alternative estimation techniques besides Tobit. The results of Ordinary Least Squares regressions are presented in Table V in the Appendix. The table shows that the signs and significance levels for the OLS estimations are similar to that of the Tobit analysis. The only main deviation is the insignificant coefficient for the factor that aggregates the gender role attitudes. Subsequently, Table VI of the Appendix depicts the results of Logit regression estimations in which the gender role attitudes are regressed on a dichotomous variable that indicates the presence of at least one female director on the board. These Logit estimations are again similar to the results in Table 4, including the significant negative impact of the gender role attitudes factor. As the previous findings are largely robust to the different estimation methods, the conclusions are independent of assumptions about the distribution of the data.

VI. Conclusions

(28)

28 the period 2002 up to and including 2013. The results show that gender role attitudes are indeed highly relevant in explaining the extent of female board members. The more traditional a society’s gender role attitudes are, which implies that it is mainly believed that men are more fit to work in a paid job and women would better stay home as a housewife, the less women are represented in the corporate board of directors. This is in line with past empirical literature that finds that people’s views towards the role of women in the family and labor market significantly influences women’s decisions to work (Farré and Vella, 2013). Furthermore, the results of this paper indicate that in general women’s opinions about gender roles have a larger effect on a country’s female board participation than men’s. Women’s gender role attitudes can directly be carried out in their own labor market choices as well as indirectly influence their families’ decisions to work. Therefore, women’s views in particular have a large impact on the representation of women on boards.

The results suggest that the underlying causes for the relatively low percentage of female board members are deeply rooted in the cultural framework of countries. Therefore, simply implementing a gender diversity quota to which boards should comply might be a less effective approach than dealing with the actual cultural determinants of female board participation, namely gender role attitudes.

A first limitation of this paper is that it uses an unbalanced dataset since the information on gender role attitudes is not available for all years and all countries. Second, the explanatory variables are relatively time invariant such that the results are mainly gathered from the cross-sectional variation in the data. This implies that the relationship between gender role attitudes and female board participation originates from differences among the countries and might not be present over time. The countries included in the dataset are diverse in nature, such that other country-specific characteristics might be the underlying factor of the found relationships. For example, general cultural differences might be the actual cause of variation in female board participation, which would imply that gender role attitudes are only correlated with female board participation. Furthermore, the relationship between gender role attitudes and female board participation might differ among the various types of countries, such that including all countries disperses the true influences.

(29)

29 women whose labor market decisions fit their gender role preferences. Therefore, the actual underlying factor of female board participation might be solely women’s gender role attitudes. Moreover, further research could detect policies that can successfully influence gender role attitudes such that the dominant view regarding the role of women in the family and labor market becomes less traditional. Since this paper finds that traditional gender role attitudes hinder female board participation, policies targeting gender role attitudes could potentially decrease the barriers to women’s career advancements and increase the number of female board directors.

VII. References

Adams, R. B., & Kirchmaier, T. (2013). Making it to the top: From female labor force participation to boardroom gender diversity. ECGI-Finance Working Paper, (347).

Alesina, A., & Giuliano, P. (2010). The power of the family. Journal of Economic Growth, 15(2), 93-125.

Baron, J. N., Mittman, B. S., & Newman, A. E. (1991). Targets of opportunity: Organizational and environmental determinants of gender integration within the California Civil Service, 1979-1985. American Journal of Sociology, 1362-1401.

Bianco, M., Ciavarella, A., & Signoretti, R. (2011). Women on boards in Italy. Rome: Consob.

Burgess, Z., & Tharenou, P. (2002). Women board directors: Characteristics of the few. Journal of Business Ethics, 37(1), 39-49.

Burke, R. (1999). Women on Canadian corporate boards of directors: Getting the numbers right!. Corporate Governance: An International Review, 7(4), 374-378.

Campbell, K., & Vera, A. M. (2010). Female board appointments and firm valuation: short

and long-term effects. Journal of Management & Governance, 14(1), 37-59.

(30)

30 Conyon, M. J., & Mallin, C. (1997). Women in the boardroom: evidence from large UK companies. Corporate Governance: An International Review, 5(3), 112-117.

Erhardt, N. L., Werbel, J. D., & Shrader, C. B. (2003). Board of director’s diversity and firm financial performance. Corporate Governance: An International Review, 11, 102–111.

European Commission. (December, 2013). Database on women and men in decision-making.

European Commission. (2012). Proposal for a directive of the European Parliament and of the Council on improving the gender balance among non-executive directors of companies listed on stock exchanges and related measures.

Farré, L., & Vella, F. (2013). The intergenerational transmission of gender role attitudes and its implications for female labour force participation. Economica,80(318), 219-247.

Farrell, K. A., & Hersch, P. L. (2005). Additions to corporate boards: the effect of gender. Journal of Corporate Finance, 11(1), 85-106.

Fernández, R. (2007). Culture as learning: The evolution of female labor force participation over a century (No. w13373). National Bureau of Economic Research.

Fernandez, R., & Fogli, A. (2005). Culture: An empirical investigation of beliefs, work, and fertility (No. w11268). National Bureau of Economic Research.

Fernández, R., Fogli, A., & Olivetti, C. (2004). Mothers and sons: Preference formation and female labor force dynamics. The Quarterly Journal of Economics, 1249-1299.

Fogli, A. and Veldkamp L. (2007). Nature or Nurture? Learning and Female Labor Force Dynamics. Federal Reserve Bank of Minneapolis, Research Department Staff Report 386.

Goldin, C. and Katz, L.F. (2002). The Power of the Pill: Oral Contraceptives and Women's Career and Marriage Decisions. Journal of Political Economy, 730-770.

Greenwood, J., Seshadri, A., & Yorukoglu, M. (2005). Engines of liberation. The Review of Economic Studies, 72(1), 109-133.