Gender Diversity in the Board of Directors and the effect on Firm Financial Performance: Including the Role of Stock Markets

(1)

Gender Diversity in the Board of Directors and the effect on

Firm Financial Performance: Including the Role of Stock

Markets

Student number: s2387344 Name: Siebe Hannema Supervisor: Prof. M. Ararat Co-assessor: Prof. H. Gonenc

Master: International Financial Management University of Groningen

Faculty of Economics and Business Date: 28-06-2018

Abstract

This research empirically analyses whether board gender diversity enhances firm performance. Using balanced panel data from 1529 stock listed companies between 2011 and 2016, I have found that firms with more gender diverse boards have higher firm performance by market (Tobin´s Q) and accounting (return on equity) measures. These results hold true with respect to different estimation models. Furthermore, the interaction effect of stock market development on the relationship between board gender diversity and firm performance has been investigated. I have found that stock market development positively affects the relationship between board gender diversity and firm performance. Consequently, this study suggests that stock listed companies should consider installing gender diverse boards because this seems to positively contribute to firm financial performance.

(2)

1

1. Introduction

(3)

2 minimum percentage of female board members. The recruitment of women in order to meet quota regulations will increase the diversity amongst the board but does not necessarily lead to improved firm performance, as some women are merely appointed to reach the company’s regulation quota and not necessarily for their qualities (Bertrand, Black, Jensen. and Leras-Muney, 2014). Therefore, the fact that certain European companies are required by law to have a percentage of women present on their boards of directors will not always lead to the desired positive effect.

This research focuses on determining if gender diversity has a significant impact on the performance of listed firms. Despite a relatively large quantity of literature examining the relationship between board gender diversity and firm performance, the results are decidedly mixed. For instance, female directors bring different experiences and knowledge to boards of directors, as female directors often originate from different backgrounds (Hillman, Albert, Cannella and Paetzold, 2000).

Contrarily, other researchers, including Kenneth and Dittmar (2009), have found evidence for a negative relationship to exist between board gender diversity and firm performance. The negative relationship could be the result of the existence of more conflicts within gender diverse boards. Maznevski (1994) noticed that homogenous boards cooperate better as they have fewer emotional conflicts, which indicates that gender diverse boards are less effective.

(4)

3 shareholder base and increases political support for shareholder protection. Likewise, La Porta, Lopez, Schleifer and Vishny (2000) reported that countries that embrace shareholder protection have more sizeable and higher capitalised stock markets. Shareholder protection enables shareholders to obtain certain powers or rights that are granted by law. Overall, shareholders find their position improving within companies if stock markets are more developed. As such, the improved position is a result of more shareholder protection and liquidity advantages.

An increase in the liquidity position and shareholder protection leads to more monitoring (Maug, 1998), which is beneficial for the control shareholders can effectively exercise within their firm. By having more powerful shareholders, companies will be pressured to act in accordance with the needs of these shareholders. In recent years, shareholders have put pressure on boards to appoint more women, given the benefits of diversity. In developed markets, boards are expected to be more diligent when appointing women, since they are under more scrutiny by the shareholders.

Because stock market development increases liquidity positions and shareholder protection, I expect the qualifications of women appointed to boards of directors to be higher in developed stock markets. This leads to the following research question:

What is the effect of board gender diversity on firm performance, and how does stock market development affect this relationship?

(5)

4 In order to answer the research question and to test the hypotheses, data have been collected from nine different stock markets based in eight countries during the period 2011–2016. The results suggest that board gender diversity has a significant positive effect on firm performance, implicating that board gender diversity actually increases firm performance. Moreover, the interaction effect of stock market development on the relationship between board gender diversity and firm performance is significant, implying that stock market development directly influences the relationship between board gender diversity and firm performance.

(6)

5

2. Literature Review

This literature review provides an exploration of relevant theories and reviews on previous studies regarding board gender diversity and firm performance. The hypotheses are developed subsequently. In section 2.2, several theories are discussed that predict the relationship between board gender diversity and firm performance. Section 2.3 presents a summary of the relevant literature. Finally, section 2.4 illustrates the effect of stock market development on the relationship between board gender diversity and firm performance.

2.1 Importance of gender diversity in the board of directors

The board of directors is basically considered to be the most important decision-making body in an organization, because the strategic decisions boards make are key to firm’s results (Marimuthu and Kolandaisamy, 2009). Therefore, board composition and decision-making dynamics matter. One of the most important differences between men and women is in their cognitive reasoning. According to Jehn (1995) and Kanter (1978), women differ from men in their information seeking and information evaluating processes. These processes are influenced by the experiences, knowledge and values people have. The differences between men and women in their cognitive frameworks will therefore shape board decisions, decision making processes and ultimately firm outcomes. Consequently, I argue that gender diverse boards do have an effect on firm performance. To find theoretical support for this, the agency theory and the resource dependence theory are the most relevant theories to explain the effects of different board compositions on firm performance (Low, Roberts and Whitling, 2015). There will be elaborated upon these theories in the next subsections. Thereafter, the social psychology theory and human capital theory will be discussed to further examine possible outcomes on the relationship between board gender diversity and firm performance. Finally, the concept of tokenism is discussed.

2.2.1 Agency theory

(7)

6 agency theory is based on the assumption that managers do not always act in accordance with the interest of the principals (shareholders). The board of directors plays a crucial role in managing the relationship between shareholders and managers, because poor governance results in higher agency costs and lower firm performance. Jensen and Meckling (1986) argued that the role of the board of directors is to resolve these agency issues, between managers and shareholders, by using compensation policies or replacing managers who do not create shareholder value. However, to resolve agency problems, the independence of the board of directors is crucial. It is known that women in boards of directors tend to have fresh views on issues that need much attention and can alleviate information biases quickly into a formulated strategy (Singh, Terjersen and Vinnicombe, 2008). Moreover, women tend to perform their roles in a more serious manner, hence increasing monitoring independence and enhancing corporate governance. The enhancement of corporate governance contributes to the reduction of agency problems and gives assurance to shareholders to ultimately receive a positive return on their investment.

2.2.2 Resource dependency theory

(8)

7 argue for more board gender diversity (Hillman, Albert, Cannella and Paetzold, 2000). Furthermore, the board of directors has the responsibility to associate with other external firms, fulfilling its dependencies on the environment (Bear, Rahmen and Post, 2010). Women are thought to understand and acknowledge the market better than men, as they are able to cooperate more easily with other firms, consumers, and employees (Hillman, Albert, Cannella and Paetzold, 2000). In other words, firms that possess gender diverse boards are equipped with different resources (knowledge, legitimacy, and prestige), which enables these firms to better cope with risks and uncertainties (Terjesen, Sealy and Singh, 2009).

2.2.3 Human capital theory

The human capital theory addresses the different skills, experiences and knowledge of board members. Becker and Gary (1975) state that the education, experiences, and skills people possess are used to the benefit of an organization. According to Terjesen, Sealy and Singh, (2009), diverse boards have a unique human capital aspect, because both male and female directors have different competencies. Competencies are generated by different levels of education and experiences. As a result of this, decision-making processes by the board of directors will enhance and improve, in the event that knowledge and skills of directors are more diversified. Consequently, the human capital theory contributes to different firm outcomes, as gender diverse boards generate new and more human capital for firms.

2.2.4 Social psychological theory

(9)

8 that board gender diversity, in relation to heterogeneous boards, brings more perspective to the board of directors, which could lead to innovative and effective ideas and hence improve firm performance. Even so, many researchers have highlighted the importance of the negative relationship between the profits of a firm and gender divers boards (Marinova, Plantenga and Remery, 2010).

2.2.5 Tokenism

Another argument researchers use to explain the relationship between board gender diversity and firm performance is the concept of so called tokenism. Tokenism is a practice that incorporates minority members into groups to prevent criticism, because it gives the appearance that everyone is being treated equally. Currently, some firms elect women to boards of directors because the European Union has issued several directives regarding board gender diversity, which has been implemented in for example: Norway, France and Germany. In that situation, women are being appointed to boards for legal reasons and not necessarily for the qualities they possess. This leads to an environment where women are less accepted than men, because they are seen as outsiders (Bertrand, Black, Jensen. and Leras-Muney, 2014). When women operate as a token within a firm, their job pressure increases as women feel more pressure to perform and their capabilities cannot be fully capitalized. As a result of this, the appointment of female directors will not necessarily be beneficial to a firm. However, the token status of women can improve if the number of women increases within the board of directors (Zimmer, 1988). By appointing multiple women to the board of directors, a critical mass is achieved. This means that women are less likely to be ignored.

2.3 Gender diversity and firm performance

(10)

9 performance. Furthermore, Rose (2007) discovered that there is no solid conclusion on the relationship board gender diversity and firm performance. Carter, Simkins and Simpson (2003) examined a sample of U.S. firms and found a positive relationship between ethnic minority board members and Tobin’s Q. They suggested that monitoring independence is enhanced when boards are more diverse. As diverse boards impose different questions and have other values, this will eventually lead to a broader way of thinking, resulting in independent behaviour and improved corporate governance (Singh, Terjersen and Vinnicombe, 2008). The improvement of monitoring independence results in the reduction of agency costs and creates positive returns for shareholders. The findings of Erhardt, Werbel and Schrader (2003) are in accordance with the human capital theory. They noted that the proportion of women in top management tends to have a positive impact on the firm’s return on investment (ROI) and return on assets (ROA). The results suggest that diversity i n the board of directors tends to increase the level of skills and experiences, hence increasing the effectiveness of the board of directors.

With the same positive outcome, Dezso (2008) conducted a study on female effectiveness in top management and used Tobin’s Q as a performance measure. The outcome highlighted a positive relationship between female participation and firm performance in many organizations, as women tend to be more innovative. Innovation creates a competitive advantage and is thus realized when firms strive for gender-diverse boards and the development of managerial talents (Joecks, Pull and Vetter 2013).

Smith and Verner (2006) discovered that the presence of women in top management in the 2,500 largest Danish firms bears a positive relationship with firm performance. They pointed out that the positive effect of women in management strongly depends on their qualifications or capabilities to manage a firm.

(11)

10 they acknowledge the positive effect, which claims that more gender diversity contributes to the overall performance. Secondly, Campbell and Minquez-Vera (2008) found that the balance between men and women on the board of directors is just as important as the presence of women, because evenly balanced boards cooperate better.

In addition, Sussmuth-Dyckerhoff, Wang and Chen (2012) found evidence that female directors have the ability to strengthen the relationships with other firms. Women can build more reliable relationships than men, which enables women to gain more knowledge about their partners and clients (Bear, Rahmen and Post, 2010).

In contrast, several researchers reported a negative relationship between board gender diversity and firm performance (Bøhren and Strøm, 2007 and Adams and Ferreira, 2009). Adams and Ferreira (2009) examined S&P 500 companies and found that the average effect of gender diversity on firm performance (measured by the return on assets and Tobin’s Q) is negative. They found that gender diversity in the board of directors leads to more efforts regarding corporate governance and monitoring activities, due to the perspective that women are more punctual than men. Furthermore, Adam and Ferreira (2009) argue that a broader way of thinking might not necessarily always lead to more effective monitoring, because it may lead to an increase in conflicts.

Moreover, Bøhren and Strøm (2007) argue that small boards and less gender-diverse boards enhance firm performance, because gender-gender-diverse boards decrease the amount of networking activities. In accordance with the resource dependency theory, important networking resources are thus lost, which negatively affects firm performance.

Finally, some authors argue that board diversity is not correlated in any way to firm performance. Rose (2007) conducted a study using 100 Danish companies in their sample. The results indicated that board composition has no relationship with firm performance.

(12)

11 dominant theories, agency and resource dependency theory, I have the opinion that women bring positive benefits to firms, basically as a result of the fact that they increase levels of monitoring independence and secure more resources. Furthermore, the extensive amount of empirical evidence on the relationship does provide clear support for a positive relationship. As a result, I have devised the following hypothesis:

Hypothesis 1: Board gender diversity increases firm performance.

2.4. Interaction effect of stock market development

The development of stock markets implies wider stock market participation and an increase in the size of the market (Rocha, 2008). Wider stock market participation results in increased political support for investor protection (Pagano, 1993). It is important to note that political goals are aimed at increasing the level of investments and expansion of the stock market. The profitability and growth within stock markets influences economic growth and development, which is a key objective for governments. Therefore, governments will implement more investor protection policies to ensure that the central objective of economic growth is achieved. In this regard, governments and political movements will advocate for more investor protection in order to increase the investment levels. Increased investments mean expansion of employment opportunities, increase of both income per capita and gross domestic product (Murrel, 2001).

(13)

12 protection of shareholders enhances the power of shareholders (Bruno and Ruggiero, 2011).

The size of stock markets is commonly measured using market capitalization, which is mainly the number of listed securities multiplied by the current market price. However, the size of the market is not only quantified by capitalization measures but also by liquidity levels of the market (Bainbridge, 2013). A large-size stock market should be complemented with a high level of liquidity to ensure stock market development. The value of market liquidity is anchored on the ability of investors to develop opportunities through which they can transact easily and cheaply. This implies that investors can buy and sell securities in the market at little or no cost.

The liquidity of stock markets is commonly measured using two indices: the stock value traded as a percentage of GDP and turnover rates. Turnover rates are considered to be the total value of shares as a ratio of market capitalization (Levine and Zervos 1996). Higher turnover rates imply that transactions costs are relatively low and the liquidity ratio would be rather high. If the turnover rates of trading volumes increase, shareholders gain on power. This power is elevated due to the increased ability to buy and sell shares and the increase in the profitability of the transactions. Overall, an increase in shareholder power translates to the ability to influence boards of companies (Rocha, 2008).

In summary, stock market development has different approaches to achieve more shareholder power. Through the liquidity approach, stock market development is aligned with greater liquidity. On the other hand, stock market development increases with increased investor protection. An increase in investor protection increases the power of shareholders and thus their ability to monitor the company’s operations and advocate for better firm performance. This is achieved by being able to vote for major corporate matters and closely monitor the (financial) performance of the firm.

(14)

13 Boards are expected to be more diligent when appointing women, since they are under more scrutiny by shareholders. Consequently, I expect the qualifications of women on the boards to be higher in developed stock markets. This is emphasized in the following research question:

Hypothesis 2: Stock market development has a positive interaction effect on the

(15)

14

3. Data and Methodology

3.1 Data

This research contains balanced panel data from companies that are listed on nine different stock markets, covering a time span of six years (2011-2016). The data regarding stock market development is based upon the findings of the World Federation of Exchanges (WFE). The WFE provides stock market-related information on 64 regulated stock exchanges across the world and categorizes these 64 stock exchanges into three groups1_{. In addition, the WFE includes a time}

span of only six years in their database, which limits the possibility of using a larger time span. Furthermore, the WFE only provides data regarding the ten largest stock markets for the years 2011, 2012 and 2013. As a result, I have incorporated nine2_{stock markets in my sample. More specifically, the stock}

markets are the New York Stock Exchange, the Tokyo Stock Exchange, the Deutsche Börse, the Shanghai Stock Exchange, the Hong Kong Stock Exchange, the NASDAQ Stock Exchange, the Toronto Stock Exchange (TMX), the Swiss Stock Exchange and the London Stock Exchange. The source used for collecting financial data and board characteristics is the Thomson Reuters ASSET4 database. Compustat was used to derive information about different industries, because identifying different industries is needed to provide accurate classifications regarding the company’s business activities. The World Bank lists information regarding the gross domestic product per country. Finally, firms with missing data haven been excluded from the sample size. The final sample contains 1529 stock exchange-listed firms with 9171 observations.

3.2 Dependent variable

According to Hansen and Wernerfeld (1989), firm performance has several definitions, because performance measurements are based on long- or short-term probabilities. Most researchers measure firm performance by focussing on corporate social/environmental performance (CSP) or corporate financial

1_{Table 11 in the appendix provides the global classifications of the different stock markets.} 2_{The Euronext stock exchange is excluded from the sample as it incorporates several countries} (the Netherlands, France, England, Belgium and Portugal), which does not allow me to

(16)

15 performance (CFP) (Orlitzky, Schmidt and Rynes, 2003). Both measurements assess whether a firm has sustained a superior position in comparison with its competitors.

Many researchers in the field of board gender diversity have used financial firm performance as their dependent variable (Carter, D´Souza, Simkins and Simpson, 2010 and Abdullah and Ismail, 2016), because financial performance is a widely acknowledged way to measure firm performance. Consequently, this thesis will use financial performance measurements to assess firm performance. The financial performance measurements are split into two categories.

Firstly, there are market-based or forward looking performance measurements. They refer to the behaviour of a security or asset in the marketplace, reflecting long term expectations on firm value (Thaler, 2004). According to Post and Byron (2015) there are multiple forward looking measurements such as the market to book ratio, Tobin’s Q, stock performance and stakeholder returns. This research uses Tobin’s Q as the market based measurement. Tobin’s Q is a widely applied measure within gender related studies, and serves as a proxy for a firm’s ability to generate shareholder value (Carter, D´Souza, Simkins and Simpson, 2010). Tobin’s Q is computed by dividing market value of equity by the book value of equity (Dsezo, 2012).

(17)

16

3.3 Independent variable

The independent variable used in this research is board gender diversity. The academic literature uses several measures for board gender diversity. Whereby, many scholars, including Adams and Ferreira (2009), use the proportion of female directors on the board of directors as a measurement for board gender diversity, hence the percentage of women on the board of directors is used to calculate board gender diversity in this research.

3.4 Control variables

To control for biased results, four control variables are included in the regression. Based on prior research, the control variables in the regression models are: firm size (Fsize), return on assets (ROA), leverage (Lev) and board size (Bsize).

The effect of firm size on firm performance has been investigated thoroughly in the past (Barney, 1991). Larger firms gain more in their productivity than smaller firms do, which enables them to create scale advantages on their activities. These scale advantages can lead to an increase in the company’s firm performance. To control for the skewed distributions of the results (heteroskedasticity), a natural logarithm of the total assets is used.

Furthermore, a financial performance measure is included in the control variables. The return on assets (ROA) is included as a control variable. The ROA and ROE are accounting-based financial measurements for firm performance, whereas Tobin’s Q is a stock market-based measurement. Because, the variables are measurements for the financial performance of a firm, I expect that if ROA increases or decreases, the score of Tobin’s Q and ROE will change in accordance with the ROA. The return on assets is calculated as the earnings before interest and taxes (EBIT) over the assets of a company.

(18)

17 is in line with the findings of Jensen and Meckling (1976), which indicated that higher leverage is associated with more efficiency, because debt is used in order to increase the growth opportunities of companies. Consequently, the increase in growth will have a positive effect on the firm’s performance.

Finally, I have controlled for the size of the board of directors. Adams and Ferreira, (2009) found empirical evidence that the size of a board has direct effect on firm performance. Larger boards possess more skills and knowledge, which enables companies to gain valuable resources. On the other hand, larger boards create more conflicts and reduce governance abilities. Campbell and Minquez-Vera (2008) used board size as control variable, as larger boards automatically have more women on the board of directors.

Additionally, year-, country-, and industry-dummies are included in this research. The year dummies include six years (2011-2016), and the country dummies include the nine stock exchanges. Furthermore, firms are grouped into industry dummies according to their Standard Industrial Classification (SIC) codes derived from Compustat. The firms are grouped into the following categories: agriculture, construction, manufacturing, transportation, trade, finance, and services.

3.5 Interaction variable

(19)

18

Table 1. Variable description

Variable Description

Tobin’s Q Tobin’s Q is computed by dividing market value of equity by the book value of equity

ROE Return on equity, which is calculated as the net income divided by the book value of equity.

BDG Board gender diversity, calculated as the percentage of women present on the board of directors.

Stkdev Stock market development is based on the market size of a stock market (market capitalization / GDP).

ROA Return on assets of a company, which is measured as earnings before interest and taxes (EBIT) over assets of a company.

Lev Leverage, which calculates the operations from external funding, is computed by dividing long term debt by assets.

Bsize Board size, which is the number of people on the board of a company.

Fsize Firm size, which is the natural logarithm of a company’s assets.

BDG*Stkdev The interaction term between the board gender diversity and stock market development.

Country

dummies Nine distinct dummy variables to measure the influence of different stock markets.

Year

dummies Six distinct dummy variables to measure the influence of different years. Industry

dummies Eight dummy variables regarding the influence of different sectors. SIC codes have been used to distinguish sectors.

3.6 Methodology

3.6.1 Ordinary least squares regression

(20)

19 (1)

Performanceit = b0 + b1 BGDi,t + b2Fsizei,t + b3ROAi,t + b4Levi,t+ b5Bsizei,t + ᶓi,t

(2)

Performanceit = b0 + b1 BGDi,t + b2 Stkdevi,t + b3 BGDi,t * Stkdevi,t + b4Sizei,t + b5ROAi,t

+ b6Levi,t+ b7Bsizei,t + ᶓi,t

The first equation measures the first hypothesis, namely the effect of board gender diversity on the performance of a firm. The second equation includes the interaction effect of stock market development. The Performanceit is measured

through Tobin’s Q and the return on equity of a firm. The independent variable is board gender diversity (BGD), and the control variables are firm size (Fsize), return on assets (ROA), leverage (Lev) and board size (Bsize), and ᶓi,t stands for

the error term. In the equation, i represent the different firms, t represents time,

and the constant component is b0I.

In the second equation, an interaction effect (Stkdev*BGD) is included. The interaction effect examines the effects of stock market development on the relationship between board gender diversity and firm performance. To control for serial correlation and heteroskedasticity, all the regression models include robust standard errors.

3.6.2 Endogeneity

According to Adams and Ferreira (2009), endogeneity can affect the relationship between board gender diversity and firm performance and creates biased results. There are two sources of endogeneity: unobserved heterogeneity and reverse causality.

(21)

20 the dependent variables, because the decisions made in a preceding year by a board of directors most likely will affect performance the next year (Liu, Wei and Xie, 2014).

In this research, the amount of unobserved heterogeneity refers to unobserved board or firm characteristics that affect the relationship between board gender diversity and firm performance. The OLS regression assumes that the relationship between variables is constant across firms and time. Whereby, unobserved characteristics are more likely to generate biased results in the OLS regression. To address this problem, random or fixed effects should be employed in the regression model (Smith, Smith and Verner, 2006). To find out which effect is appropriate to use in the regression model, a Hausman test was employed. The Hausman test compares the coefficient estimates from both random and fixed models and assumes consistency. However, the results of the Hausman test are significant (p<0.00)3_{, which suggests that random effects are}

inconsistent and not appropriate to use. Therefore, I have included time (year dummies) and firm (group and industry dummies) fixed effects in the regression model. In order to determine whether fixed effects influence the results, both the OLS and OLS with fixed effects regressions were performed. The OLS regression model with fixed effects is shown in the third and fourth equation.

(3)

Performanceit = b0 + b1 BGDi,t + b2Fsizei,t + b3ROAi,t + b4Lev i,t+ b5Bsizei,t + λ𝑡 +α𝑖 + ᶓi,t

(4)

Performanceit = b0 + b1 BGDi,t + b2 Stkdevi,t + b3 BGDi,t * Stkdevi,t + b4Sizei,t + b5ROAi,t

+ b6Levi,t+ b7Bsizei,t + λ𝑡 + ᶓi,t

Where λ 𝑡 stands for fixed effects

3.6.3 Multilevel regression

In order to find out whether the results of the OLS regressions are robust, a multilevel analysis has been performed. The multilevel analysis obtains hierarchical data, which implies that variables are clustered in other variables on

(22)

21 a higher level (Hox, 1995). In this research, I assume that the data regarding stock market development is nested, because the values for stock market development only differ across time and within different stock markets. Furthermore, firms in the same stock market possess more of the same characteristics (Atje and Boyan, 1993). Nested data is data for which a variable signifies observations as belonging to a group (Hox, 1995). As a result, a higher level is included (two-level model), which enables observations to become more independent from each other (Hox, 1995). In this multilevel analysis, the first level reflects all the firms listed on stock markets and the relationship between board gender diversity and firm performance. The second level measures the influence of the different stock markets.

The advantage of grouping different stock markets enables researchers to measure the relationship between board gender diversity and firm performance within different stock markets. In other words, the multilevel model determines whether the association between variables differ amongst groups, which in this case would be stock markets. Furthermore, random effects are included in the model, because fixed effects are constant over all groups (Hox, 1995). Random effects can influence the intercept or slope of the regression line. This research uses the random intercept model, which allows the scores of the dependent variable of the regression line to vary amongst different stock markets. The multilevel random intercept model includes all the variables presented in the OLS regression, with Tobin’s Q and the return on equity as the dependent variables. To compare the goodness of fit between the OLS4_{models (linear) and}

multilevel model, a likelihood ratio test has been performed. This test indicated that the multilevel model is superior to the linear OLS model, as it fits the data better.

(23)

22

4. Results

4.1 Descriptive statistics

Table 2 provides the descriptive statistics of the variables collected from the stock markets. The table contains an overview of the mean, minimum, 1st quartile, 3rd quartile, standard deviation, and the number of observations. To correct for outliers, extreme values of some variables have been winsorised from this research, because outliers can be biased and cause problems in the regression analysis. The following variables are corrected for their outliers: Tobin’s Q, ROA, Lev and ROE. All these variables are winsorised at 2.5 % on both sides.

(24)

23

Table 2. Descriptive statistics5

Table 2 presents the descriptive statistics of the firm, board, and stock market specific variables. The sample includes 1529 listed firms with 9171 observations (N). Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in the first table.

Variables Mean Min 1st Quartile 3rd Quartile Max Standard deviation N Tobin’s Q 1.585 0.766 1.036 1.818 4.371 0.802 9171 ROE 0.122 -0.256 0.056 0.185 0.523 0.142 9171 BGD 14.359 0 7.14 22.22 62.5 10.886 9171 ROA 0.057 -0.095 0.015 0.084 0.32 0.074 9171 Bsize 10.992 1 9 13 30 3.243 9171 Fsize 16.383 4.369 15.235 17.359 22.021 1.760 9171 Lev 0.214 0 0.093 0.315 0.569 0.149 9171 STKdev 0.976 .088 0.901 1.246 3.171 0.538 9171 4.2 Correlation matrix

According to de Jong, Bihn Phan and van Ees (2010), highly correlated variables weaken the explanatory power of the regression. Therefore, a correlation matrix is presented in table 3 to test for multicollinearity amongst all the variables. A common threshold value for multicollinearity is 0.80 (de Jong, Bihn Phan and van Ees, 2010). All values of the correlation coefficients are relatively small, which indicates that multicollinearity does not seem to be an issue. Only the correlation coefficient of the return on equity on Tobin’s Q has a coefficient of 0.821. This high correlation is intuitively logical since both proxies are firm performance measurements. Due the fact that Tobin’s Q and the return on equity are tested in different models, multicollinearity does not seem to be a problem in this research.

(25)

24

Table 3. Correlation matrix

Table 3 presents the correlation matrix of all the variables. The abbreviations of the variables are explained in the first table. ***, ** and * coefficients are statistically significant at 1, 5, and 10 %, respectively (based on two-sided testing) b. Observations (N) = 9171

1 2 3 4 5 6 7 8 1. Tobin’s Q 1 2. ROE 0.821*** 1 3. BGD 0.226*** 0.273*** 1 4. ROA 0.512*** 0.675*** 0.131*** 1 5. Bsize -0.137*** 0.022** 0.133*** -0.043*** 1 6. Fsize -0.279*** -0.037*** 0.115*** -0.316*** 0.418*** 1 7. Lev -0.037*** -0.021** 0.069*** -0.118*** -0.037*** -0.119*** 1 8. STKdev 0.035*** 0.055*** -0.064*** 0.292*** -0.062*** 0.002 0.046*** 1 4.3. Empirical results

In this section the results of the regression analyses are presented. Table 4 provides the results of the OLS regression, which corresponds with the first and second equitation presented in section 3.6.1. Table 5 includes fixed effects in the OLS regression to control unobserved heterogeneity and corresponds with the third and fourth equation presented in section 3.6.2. Table 6 includes one year lagged independent variables in the regression, to address the problem of reverse causality. Whereas, table 7 incorporates both fixed effects and lagged variables in their model. Finally each model consists of eight different models, which all include different variables in their regression.

(26)

25 board size are statistically significant at the 1% level, but have negative coefficients, which is not in line with the expectations of this research. Furthermore, leverage (lev) is statistically insignificant with a positive coefficient; hence no inferences can be made about leverage on Tobin´s Q. The results of the control variables on the return on equity are presented in model 5. Several variables change when the ROE is used as the dependent variable. In Model 5 firm size has a positive and significant (β=0.018, p<0.01) coefficient, whereas firm size has a negative coefficient in model 1. In addition, leverage has a positive effect (β=0.086, p<0.01) on firm performance (ROE).

Hypothesis 1 predicted that board gender diversity is positively related to firm performance. To test the first hypothesis, the models 2 and 6 are relevant. In both models the coefficients of board gender diversity (BGD) are positive and significant (p<0.01). The results imply that a higher percentage of female board members are positively related to the financial performance (Tobin´s Q and ROE) of a firm. When the percentage of female directors increases with 1% the return on equity goes up with 0.001 and Tobin’s increases with 0.013. This finding is similar to those obtained by Erhardt, Werbel and Schrader (2003) and Carter, D´Souza, Simkins and Simpson (2010), who found significant evidence for a positive relation between board gender diversity and firm performance. Interestingly, several coefficients of the control variables change when comparing the results with models 1 and 5. This indicates that adding the variable board gender diversity changes the influence of other variables on firm performance.

(27)

(28)

27

Table 4. Regression results of the effects of women on boards of directors on firm performance through Tobin’s Q and ROE

The results of the OLS regression are presented in this table. Tobin’s Q and the return on equity (ROE) are the dependent variables. Tobin´s Q is calculated as the market value of equity divided by the book value of equity and the ROE is the net income divided by the book value of equity. Board gender diversity (BGD) is calculated as the ratio of females to total board members. Board size (Bsize) is the amount of board members. Firm size (Fsize) is calculated through a natural logarithm of the firm’s assets. The performance measure return on assets (ROA) is calculated by dividing the amount of earnings before interest and taxes (EBIT) through assets. Leverage (lev) is the debt to assets ratio and to calculate stock market development, market capitalization is divided by. The robust standard errors of the coefficients are shown in parentheses. ***, ** and * coefficients are statistically significant at 1, 5, and 10 %.

Dependent variables

Independent

variables Tobin’s Q Tobin’s Q Tobin’s Q Tobin’s Q ROE ROE ROE ROE

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8

(29)

28 Table 5 shows the results of OLS regression including fixed effects on the relationship between board gender and firm performance. This regression method captures the effects of unobserved heterogeneity and eliminates biased results from the OLS regression. Compared with the results of the OLS regression in table 4, the adjusted R-squared increases significantly in all models, indicating that incorporating fixed effects better fits the data than the OLS regression. Table 5 finds convincing evidence for a positive relation between bo ard gender diversity and firm performance (Tobin’s and ROE) (p<0.01). Furthermore stock market development increases the positive effect of board gender diversity on the return on equity. These finding are in line with results of the OLS regression. To the same positive result, the return on assets (ROA) and firm size (Fsize) remain significant.

In contrast, several results in table 5 have significant differences when fixed effects are employed. Specifically, it is shown that after controlling for unobserved variables the relationship between leverage and Tobin’s Q becomes positive and significant (β=0.151, p<0.05). Board size has become insignificant, whereas table 4 found evidence for a negative relation between board size and firm performance (Tobin’s Q and ROE). In addition model 4 has an insignificant interaction effect, implying that stock market development has no effect on the relationship between board gender diversity and Tobin’s Q.

(30)

29 8 is significant at the 5% level and the values of the coefficients remain negative in model 4 and positive in model 8, suggesting that the effect of stock market development varies, depending on different types of firm performance measurements.

(31)

30

Table 5. Regression results of the effects of women on boards of directors on firm performance through Tobin’s Q and ROE

Table 5 includes fixed effects in the OLS regression to control for unobserved heterogeneity. The results are presented below. Tobin’s Q and the return on equity (ROE) are the dependent variables. Tobin´s Q is calculated as the market value of equity divided by the book value of equity and the ROE is the net income divided by the book value of equity. Board gender diversity (BGD) is calculated as the ratio of females to total board members. Board size (Bsize) is the amount of board members. Firm size (Fsize) is calculated through a natural logarithm of the firm’s assets. The performance measure return on assets (ROA) is calculated by dividing the amount of earnings before interest and taxes (EBIT) through assets. Leverage (lev) is the debt to assets ratio and to calculate stock market development, market capitalization is divided by. The robust standard errors of the coefficients are shown in parentheses. ***, ** and * coefficients are statistically significant at 1, 5, and 10 %.

Dependent variables

Independent

Bsize 0.006 0.003 0.003 0.003 _0.001 _0.001 _0.001 _0.001 [0.002] [0.002] [0.002] [0.002] [0.000] [0.000] [0.000] [0.000] Fsize -0.113*** -0.120*** -0.120*** -0.120*** 0.009*** 0.008*** 0.008*** 0.008*** [0.005] [0.004] [0.005] [0.005] [0.001] [0.001] [0.001] [0.001] ROA 5.899*** 5.809*** 5.809*** 5.796*** 1.712*** 1.701*** 1.699*** 1.687*** [0.188] [0.100] [0.186] [0.189] [0.033] [0.033] [0.033] [0.034] Lev 0.151** 0.171*** 0.171*** 0.171*** 0.079*** 0.077*** 0.077*** 0.077*** [0.047] [0.046] [0.046] [0.047] [0.009] [0.009] [0.008] [0.008] BGD 0.009*** 0.009*** 0.008*** 0.001*** 0.001*** 0.001*** [0.001] [0.001] [0.001] [0.000] [0.000] [0.000] STKdev -0.025 -0.026 0.000* 0.000 [0.041] [0.042] [0.007] [0.007] STKdev*BGD -0.001 0.001*** [0.001] [0.000] Constant 2.980*** 2.980*** 3.008*** 2.293*** -0.249*** -0.185*** -0.184*** -0.188*** [0.081] [0.080] [0.094] [0.083} [0.013] [0.013] [0.016] [0.016] N 9171 9171 9171 9171 9171 9171 9171 9171 Adjusted R-squared 0.478 0.489 0.490 0.491 0.588 0.592 0.593 0.595

Country-fixed effects Yes Yes Yes Yes Yes Yes Yes Yes

Time-fixed effects Yes Yes Yes Yes Yes Yes Yes Yes

(32)

31 Dependent variables

Independent variables Tobin’s Q Tobin’s Q Tobin’s Q Tobin’s Q ROE ROE ROE ROE

Bsize _{-0.018*** -0.021*** -0.022*** -0.019*** -0.001** -0.002*** -0.002*** -0.002***} [0.003] [0.002] [0.002] [0.003] [0.000] [0.000] [0.000] [0.000] Fsize -0.041*** -0.051*** -0.049*** -0.052*** 0.016*** 0.015*** 0.016*** 0.016*** [0.005] [0.005] [0.005] [0.005] [0.001] [0.001] [0.001] [0.001] ROA 5.225*** 4.894*** 4.944*** 4.929*** 1.399*** 1.343*** 1.359*** 1.361*** [0.179] [0.179] [0.178] [0.178] [0.027] [0.028] [0.027] [0.027] Lev 0.036 -0.043 -0.045 -0.090 0.083*** 0.070*** 0.069*** 0.072*** [0.053] [0.053] [0.053] [0.053] [0.010] [0.010] [0.010] [0.010] BGD 0.009*** 0.008*** 0.009*** 0.001*** 0.001*** 0.001*** [0.001] [0.001] [0.001] [0.000] [0.000] [0.000] STKdev 0.057*** 0.023 0.019*** 0.020*** [0.014] [0.015] [0.002] [0.002] STKdev*BGD -0.011*** 0.000** [0.001] [0.000] Constant 2.163*** 2.256*** 2.189*** 2.267*** -0.238*** -0.223*** -0.245*** -0.227*** [0.085] [0.084] [0.087] [0.084] [0.013] [0.013] [0.014] [0.013] N 9171 9171 9171 9171 9171 9171 9171 9171 Adjusted R-squared 0.281 0.304 0.306 0.313 0.471 0.491 0.496 0.497 Country-fixed effects No No No No No No No No Time-fixed effects No No No No No No No No Industry-fixed effects No No No No No No No No

Table 6. Regression results of the effects of women on boards of directors on firm performance through Tobin’s Q and ROE.

(33)

32

Dependent variables

Independent

Bsize 0.006 0.004 0.004 0.004 0.001 0.001 0.001 0.001 [0.002] [0.002] [0.002] [0.002] [0.000] [0.000] [0.000] [0.000] Fsize -0.109*** -0.113*** -0.113*** -0.113*** 0.008*** 0.007*** 0.007*** 0.007*** [0.005] [0.005] [0.005] [0.005] [0.001] [0.001] [0.001] [0.001] ROA 5.925*** 5.719*** 5.719*** 5.740*** 1.707*** 1.659*** 1.659*** 1.659*** [0.188] [0.190] [0.190] [0.192] [0.033] [0.034] [0.034] [0.034] Lev 0.149*** 0.169*** 0.170*** 0.170*** 0.077*** 0.073*** 0.073*** 0.073*** [0.047] [0.047] [0.047] [0.047] [0.009] [0.009] [0.009] [0.009] BGD 0.005*** 0.005*** 0.005*** 0.001*** 0.001*** 0.001*** [0.001] [0.001] [0.001] [0.000] [0.000] [0.000] STKdev -0.001 -0.002 0.002* 0.003 [0.042] [0.042] [0.006] [0.007] STKdev*BGD -0.002** 0.000 [0.001] [0.007] Constant 2.869*** 2.892*** 2.894*** 2.889*** -0.176*** -0.171*** -0.174*** -0.171*** [0.079] [0.079] [0.093] [0.079] [0.013] [0.013] [0.015] [0.013] N 9171 9171 9171 9171 9171 9171 9171 9171 Adjusted R-squared 0.477 0.483 0.484 0.484 0.587 0.597 0.598 0.598

Country-fixed effects Yes Yes Yes Yes Yes Yes Yes Yes

Time-fixed effects Yes Yes Yes Yes Yes Yes Yes Yes

Industry-fixed effects Yes Yes Yes Yes Yes Yes Yes Yes

(34)

33

4.4. Robustness check

To verify the results, a robustness test is performed. As mentioned in section 3.6.3 the multilevel regression tests whether the results are robust to the findings of the OLS regression.

4.4.1 Multilevel analysis

Table 8 provides the results of the multilevel regression with a random intercept. Several results of the multilevel regression are robust to the findings of the OLS regression presented in table 4. For instance board size (Bsize) remains statistically insignificant, and the return on assets (ROA) again has a positive effect on Tobin’s Q and ROE. In terms of the first hypothesis, model 2 and 5 are relevant as they include the variable board gender diversity (BGD). Board gender diversity is significant and positive in both models (p<0.01). Hence, the first hypothesis is accepted, because a higher percentage of female board representation increases firm performance (Tobin’s Q and ROE). Model 4 has a negative and significant interaction effect of stock market development (Stkdev*BGD) on the relationship between board gender diversity and firm performance, implying that stock market development decreases the positive effect of board gender diversity on Tobin’s Q. On the other hand model 8 finds that stock market development increases the positive effect of board gender diversity on the return on equity. Therefore, the multilevel model accepts the second a hypothesis. Overall, the findings of the multilevel regression model are robust to the OLS regression.

(35)

34

Dependent variables

Independent variables

Tobin’s Q Tobin’s Q Tobin’s Q Tobin’s Q ROE ROE ROE ROE

Bsize -0.015*** -0.019*** -0.019*** -0.018*** _{-0.002*** -0.002*** -0.002*** -0.002***} [0.002] [0.002] [0.002] [0.002] [0.000] [0.000] [0.000] [0.000] Fsize -0.044*** -0.056*** -0.055*** -0.055*** 0.018*** 0.017*** 0.017*** 0.017*** [0.005] [0.005] [0.005] [0.005] [0.001] [0.001] [0.001] [0.001] ROA 5.191*** 4.983*** 5.020*** 5.045*** 1.406*** 1.385*** 1.401*** 1.395*** [0.177] [0.174] [0.173] [0.173] [0.027] [0.027] [0.027] [0.026] Lev 0.001 0.088 0.091 0.109 0.091*** 0.082*** 0.081*** 0.085*** [0.053] [0.053] [0.053] [0.053] [0.096] [0.096] [0.094] [0.094] BGD 0.013*** 0.012*** 0.012*** 0.001*** 0.001*** 0.001*** [0.001] [0.001] [0.001] [0.000] [0.000] [0.000] STKdev 0.049*** 0.038* 0.021*** 0.023*** [0.014] [0.014] [0.006] [0.006] STKdev*BGD _-0.005*** 0.001*** [0.001] [0.000] Constant 2.278*** 2.297*** 2.235*** 2.280*** -0.254*** -0.251*** -0.277*** -0.256*** [0.086] [0.085] [0.088] [0.085] [0.013] [0.014] [0.015] [0.014] N 9171 9171 9171 9171 9171 9171 9171 9171

Random effects Yes Yes Yes Yes Yes Yes Yes Yes

(36)

35

5. Discussion and Limitations

5.1 Discussion

(37)

36 For hypothesis 1, the results of the OLS regression models demonstrate a positive relationship between the number of female board members and firm performance (ROE & Tobin’s Q), even after controlling for reverse causality, unobserved firm characteristics. This positive result can be confirmed by both the agency theory and resource dependency theory, which imply that diverse boards have more monitoring independency and are better in securing resources, thus enabling boards of directors to increase firm performance (Hillman, Albert, Cannella and Paetzold, 2000). Consequently, I have found that board gender diversity increases firm performance.

For hypothesis 2, the results of the OLS regression suggests that stock market development has a negative interaction effect on the relationship between board gender diversity and Tobin´s Q. This negative interaction effect implies that the effect of board gender diversity on firm performance is positive for firms in countries with highly developed stock markets but the effect is more positive for firms operating in stock markets with lower development levels. Although this effect is not in line with my expectations, the positive effect of board gender diversity on Tobin’s Q is still enhanced by stock market development. On the other hand, the results of the OLS regression found that stock market development has a positive interaction effect on board gender diversity and the return on equity, implying that stock market development increases the positive effect of board gender diversity on the return on equity. Consequently, it is possible that developed stock markets increase the positive effect of gender diversity on accounting and market based performance measurements. Whereby, increased shareholder power through more investor protection and enhanced liquidity positions contribute to this positive effect.

(38)

37 implying that larger boards decrease Tobin’s Q and ROE. The return on assets is positively related to Tobin´s Q and ROE. Furthermore in the OLS regression, higher leverage rates have no effect on stock values (Tobin’s Q), but increases accounting returns (ROE), but after including fixed effects in the OLS regression leverage has a positive effect on Tobin’s Q. In addition, in the OLS regression it is found if stock markets develop, firm’s gain on their stock value and have higher accounting returns.

It is important to notice that the significance of certain coefficients in the OLS regression change, when fixed effects and lagged variables are applied. This implies that there are unobserved board or firm characteristics and that decisions made in a preceding year by a board of directors most likely will affect performance the next year.

In summary, this paper contributes to the ongoing debate about whether board gender diversity increases firm performance. Although, much research focusses on the effect of board gender diversity on firm performance, the interaction effect of stock market development has not been investigated before. The results indicate that board gender diversity increases firm performance measured through Tobin´s Q and the return on equity, implying that firms are able to create value by appointing gender diverse boards of directors. Furthermore, it is found that stock market development has an effect on the positive relation between board gender diversity and firm performance.

5.2 Limitations

(39)

38 the six years (2011-2016) might be subjected to specific firm or economic characteristics. Moreover, the sample only covers listed companies, so the results are not applicable to smaller non listed companies. Future research should use a larger sample that includes more stock markets to increase less change of significant outcomes.

Secondly, there are limitations regarding the independent variable. This thesis only examines board gender diversity by looking at the percentage of women on the board of directors, whereas several scholars use more board characteristics in their model. Furthermore, diversity can be measured through other board characteristics, like age and ethnic background. Future research should consider other board diversity characteristics when investigating the relationship between board diversity and firm financial performance.

Thirdly, there are some limitations regarding the dependent variables. This research uses ROE and Tobin’s Q as firm performance measures, the measurements represents both accounting and market performance aspects. However, ROE and Tobin’s Q are not the only financial performance measurements used to derive financial performance. There are other measures that can be used such as earnings per share, earnings before interest and taxes and return on invested capital (Carter, D´Souza, Simkins and Simpson, 2010). Besides, financial performance measurements, non-financial performance indicators can be included as well, such as customer satisfaction, since board diversity might not always influence the financial performance but could provide benefits for organisations in terms of non-financial performance. Future research could include non- financial performance indicators and some other financial performance measures.

(40)

39

6. Appendix

Conceptual model

+

Table 1. Descriptive statistics

Table 1 presents the descriptive statistics of the companies listed on the New York Stock Exchange. The sample includes 777 listed firms with 4662 observations (N) Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in table 1.

Variable Mean Min 1st Quartile 3rd Quartile Max Standard deviation N Tobin Q 1.611 0.766 1.062 1.879 4.371 0.767 4662 ROE 0.127 -0.257 0.058 0.194 0.526 0.147 4662 BGD 16.097 0 10 22.22 62.5 10.227 4662 ROA 0.044 -0.095 0.012 0.074 0.321 0.056 4662 Bsize 10.965 1 9 12 29 2.718 4662 Fsize 16.570 11.969 15.396 17.397 21.711 1.599 4662 Lev 0.237 0 0.125 0.336 0.569 0.146 4662 STKdev 1.045 0.901 0.929 1.102 1.369 0.159 4662 Stock-market development

(41)

40

Table 2 presents the descriptive statistics of the companies listed on the Deutsche Börse. The sample includes 75 listed firms with 450 observations (N) Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in table 1.

Variable Mean Min 1st Quartile 3rd Quartile Max Standard deviation N Tobin Q 1.571 0.766 1.059 1.728 4.371 0.789 450 ROE 0.121 -0.256 0.063 0.175 0.526 0.135 450 BGD 18.678 0 11.11 25 41.18 10.847 450 ROA 0.044 -0.095 0.013 0.067 0.321 0.059 450 Bsize 14.02 3 10 20 30 5.588 450 Fsize 16.522 13.096 15.068 17.760 21.838 1.789 450 Lev 0.1707 0 0.076 0.243 0.546 0.120 450 STKdev 2.392 1.938 1.967 2.654 3.171 0.426 450

Table 3 presents the descriptive statistics of the companies listed on the Japanese Stock Exchange. The sample includes 190 listed firms with 1140 observations (N) Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in table 1.

(42)

41

Table 4 presents the descriptive statistics of the companies listed on the London Stock Exchange. The sample includes 304 listed firms with 912 observations (N) Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in table 1.

Table 5 presents the descriptive statistics of the companies listed on the Hong Kong Stock Exchange. The sample includes 41 listed firms with 246 observations (N) Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in table 1.

(43)

42

Table 6 presents the descriptive statistics of the companies listed on the NASDAQ. The sample includes 81 listed firms with 486 observations (N) Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in table 1. Variable Mean Min 1st

Quartile 3rd Quartile Max Standard deviation N Tobin Q 2.747 0.811 1.782 3.814 4.3711 1.099 486 ROE 0.184 -0.257 0.121 0.261 0.526 0.159 486 BGD 16.135 0 10 20 50 9.288 486 ROA 0.127 -0.095 0.069 0.178 0.321 0.089 486 Bsize 10.337 5 9 12 16 2.066 486 Fsize 16.366 13.478 15.380 17.345 19.487 1.284 486 Lev 0.199 0 0.076 0.295 0.569 0.156 486 STKdev 0.353 0.248 0.284 0.402 0.418 0.064 486

Table 7 presents the descriptive statistics of the companies listed on the Shanghai Stock Exchange. The sample includes 37 listed firms with 222 observations (N) Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in table 1.

(44)

43

Table 8 presents the descriptive statistics of the companies listed on the Toronto Stock Exchange (TMX). The sample includes 156 listed firms with 933 observations (N) Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in table 1.

Table 9 presents the descriptive statistics of the companies listed on the Swiss Stock Exchange. The sample includes 20 listed firms with 120 observations (N) Board and firm characteristics are obtained from assets 4 and Thomson Reuters’ database. The variable definitions are presented in table 1.

(45)

44

Table 10. Intra class correlation

Level ICC Standard error

STKmarket 0.087 0.056

Table 11. Global classifications of the World Federation of Exhanges

Americas Asia - Pacific Europe - Middle East - Africa

Bermuda Stock Exchange Australian Securities Exchange Abu Dhabi Securities Exchange

BM&FBOVESPA S.A. BSE Limited Amman Stock Exchange

Bolsa de Comercio de Buenos AiresBursa Malaysia Athens Stock Exchange

Bolsa de Comercio de Santiago Colombo Stock Exchange Bahrain Bourse

Bolsa de Valores de Colombia Hochiminh Stock Exchange BME Spanish Exchanges

Bolsa de Valores de Lima Hong Kong Exchanges and ClearingBorsa Istanbul

Bolsa Mexicana de Valores Indonesia Stock Exchange Bourse de Casablanca

NYSE Group Japan Exchange Group Cyprus Stock Exchange

Nasdaq - US Korea Exchange Deutsche Börse AG

TMX Group National Stock Exchange of India LimitedDubai Financial Market

NZX Limited The Egyptian Exchange

The Philippine Stock Exchange Euronext

Shanghai Stock Exchange Irish Stock Exchange

Shenzhen Stock Exchange Johannesburg Stock Exchange

Singapore Exchange Kazakhstan Stock Exchange

The Stock Exchange of Thailand London Stock Exhange

Taipei Exchange Luxembourg Stock Exchange

Taiwan Stock Exchange Malta Stock Exchange

Moscow Exchange Muscat Securities Market Nasdaq Nordic Exchanges Nigerian Stock Exchange Oslo Børs

Palestine Exchange Qatar Stock Exchange

Saudi Stock Exchange (Tadawul) SIX Swiss Exchange

(46)

45

7. References

Abdullah, S., Ismail, K,. Nachum, L., 2016. Does having woman on boards create value? The impact of societal perceptions and corporate governance in emerging markets. Strategic Management Journal 37 (3), 466-476.

Adams, R., Ferreira, D., 2007. A theory of friendly boards. Journal of Finance 62, 217–250.

Adams, R., Ferreira, D., 2009. Women in the Boardroom and Their Impact on Governance and Performance. Journal of Financial Economics 94, 291–309. Atje, R., Boyan, J., 1993. Stock Markets and Development. European economic review 37(3), 632-640.

Bainsbridge, S., 2013. Research Handbook on Insider Trading. Cheltenham, Edward Elgar Pub Inc.

Barney, J., 1991. Firm resources and sustained competitive advantage. Journal of Management 17(1), 99-120.

Bear, S., Rahmen, 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), 207-221.

Becker, S., 1975 Human Capital. Second edition. New York: Columbia University Press.

Bertrand, M., Black, S., Jensen, S., Leras-Muney, A., 2014. Breaking the Glass Ceiling? The Effect of Board Quotas on Female Labor Market Outcomes in Norway. Unpublished working paper. National Bureau of Economic Research, 202-256.

Bøhren, Ø., Strøm, R., 2010. Governance and politics: Regulating independence and diversity in the board room. Journal of Business Finance & Accounting 37(10), 1281-1308.

Bøhren, Ø., Strøm, R., 2007. Aligned, informed and decisive: Characteristics of value-creating boards. Working paper, Norwegian School of Management.

(47)

46 Campbell, K., Minquez-Vera., A. 2008. Gender Diversity in the boardroom and Firm Financial Performance, Journal of Business Ethics, 83(3), 435-451.

Carter, D., D’Souza, F., Simkins, J., Simpson, W.G., 2010. The Gender and Ethnic diversity of US Boards and Board Comittees and Firm Financial Performance. International Review 18 (5), 396-414.

Carter, D., Simkins, B., Simpson, W., 2003. Corporate governance, board diversity, and firm value. Financial Review 38(1), 33-53.

Catalyst., 2015. Legislative Board Diversity, from 2015 http://www.catalyst.org/legislative-board-diversity

Combs, J., Crook, T., Shook, C., 2005. The dimensionality of organizational performance and its implications for strategic management research. Research Methodology in Strategy and Management 2, 259–286.

Dawe, D., 1996. A new look at the Effects of Export Instability on Investment and Growth. World Development 23 (12), 1905-1914.

Dezso, L., 2012. Does female representation in top management improve firm performance? A panel data investigation. Strategic Management Journal 33 (9), 1072-1089.

Eisenhardt, M., 1989. Agency Theory: An Assessment and Review. Academy of Management Review 14 (1), 57-74

Erhardt, N. Werbel, J., Shrader, C., 2003. Board of Director Diversity and Firm Financial Performance. Corporate Governance: An International Review 11 (2), 102-109.

Farrell, K., Hersch, P., 2005. Additions to Corporate Boards: The Effect of Gender. Journal of Corporate Finance 11, 85-106.

Freeman, R.E., 2010. Strategic Management: A Stakeholder Approach.

Goodstein, J., Guatam, K., Warren, B., 1994. The effects of board size and diversity on strategic change. Strategic Management Journal 15 (3), 241-251.

Hansen, G., Wernerfelt, B., 1989. Determinants of firm performance: The relative importance of economic and organizational factors. Strategic Management Journal 10, 399-411.