Consequences of the Dutch gender quota on board composition and firm performance

(1)

0

Consequences of the Dutch gender quota on

board composition and firm performance

MSc A&C Accountancy

University of Groningen, Faculty of Economics and Business

June 25, 2018 TIM MEIJER Student number: 2227339 Email: t.meijer.7@student.rug.nl Supervisor: S. Mukherjee Co-assessor: N. Hussain Word count: 7598

(2)

1 Abstract. This thesis builds on the growing stream of literature on gender quotas by studying the consequences of the Dutch gender quota on board composition and firm performance. I hypothesize, based on previous literature, that the gender quota law significantly increases board gender diversity. Furthermore, based on previous literature and the agency theory, I expect that the quota law significantly impacts the board composition characteristics and positively impacts firm financial performance, measured by Tobin’s Q and ROA. I perform a quantitative study with a sample of 4496 unique hand-collected director observations and 588 unique firm observations. My results suggests that the quota increased the percentage of women on boards in firms that need to comply with the Dutch quota law. Furthermore, the results indicate that the quota, besides increasing the amount of academic degrees on the board, did not have any significant effect on board composition characteristics. Lastly, I show that the quota also did not have any significant effects on firm performance, which implies that the decision for implementing a board gender quota should not be based on the criteria (future) financial performance.

Keywords: Netherlands * Gender quota * Gender diversity * Board composition * Firm performance * Board employment diversity * Tobin’s Q * ROA.

(3)

2

INTRODUCTION

On the 24th of March 2017, the Dutch Government announced the extension of the gender law that expired on the first of January 2016 to the end of 2019 (Staatsblad, 2017). The law targets a balanced distribution in the executive and supervisory boards of large public and private companies, which means that at least 30 per cent of the seats in these boards must be occupied by men and at least 30 per cent by women. Firms are required to comply with the law or explain non-compliance in their annual report, otherwise known as ‘comply-or-explain’. The Minister of Security and Justice, Ard van der Steur, mentioned in the announcement: “Although progress has been made, the current situation does not do justice to the potential of female talent. Without an active approach, the increase in the number of women takes too long and too much talent remains underused” (Rijksoverheid, 2016). Because quotas are implemented based on equality reasons but can have a serious impact on the governance of firms, I examine the consequences of a gender quota in this thesis. More specifically, I look at the impact of the Dutch gender quota regarding governance and financial aspects of firms.

Literature on the consequences of gender quotas in the boardroom is growing, with five studies on its consequences in Norway (Ahern & Dittmar, 2012; Bøhren & Staubo, 2014; Bøhren & Staubo, 2016; Bøhren and Strøm, 2010; Matsa & Miller, 2013), one on the consequences in Spain (Reguera-Alvarado, de Fuentes, & Laffarga, 2017) and one on the consequences in Europe (Lending & Vähämaa, 2017). Until now, findings suggests that gender quotas, implemented with or without non-compliance sanctions, improve board gender diversity. Furthermore, Ahern and Dittmar (2012) show that board composition characteristics changed drastically after the implementation of the Norwegian gender quota as a result of a small pool of suitable new female directors. Ahern and Dittmar (2012), Bøhren and Strøm (2010) and Reguera-Alvarado et al. (2017) studied the impact of a gender quota on firm performance, but with mixed results.

From their findings it is difficult to gain a thorough understanding of the consequences of a gender quota in the boardroom, because the results are not easily generalizable across countries. The countries are different in gender equality, have different governance structures and deal differently with non-compliance of the law. With this thesis, I contribute to the growing literature on board gender quotas by examining the consequences of the Dutch quota. To do so, I answer the following questions: (1) Does the Dutch gender quota increase the proportion of women directors in the

(5)

4 Dutch boardroom? (2) Does the Dutch gender quota have a significant impact on overall board characteristics? (3) What are the financial consequences of the Dutch gender quota?

To answer my research questions, I use a dataset with a sample period of 2010-2015 that includes board and director level data, as well as market and accounting data. Board data on firm level comes from the BoardEx database and the Thomson Reuters Worldscope database was used to match market and accounting data with the board data. Director level data was hand-collected by a group of five students of the University of Groningen through annual reports, with supplementary public sources such as the company’s website and LinkedIn. For my hypothesis testing I mainly used an difference-in-differences approach, for which I chose Finland as the control group because of similar gender equality levels and economic performance (Hofstede, 1980, 1984). Overall, this dataset has 588 firm observations and 4496 director observations.

To answer the first research question, I hypothisize that that the Dutch gender quota increases the proportion of female directors in the boardroom. I use an approach in line with Ahern and Dittmar (2012) based on the approach of Stevenson (2010). In this approach, I use the variation in board gender diversity before the quota as an instrument to seize the impact of the implementation. My results suggests that the implementation of the Dutch gender quota had a positive relationship with the proportion of female directors in the Dutch boardrooms.

Next, to answer the second research question, I hypothesize that the implementation of the Dutch gender quota had an significant impact on the composition characteristics of the Dutch boards. In answering this hypothesis, using insights from the research of Gray and Nowland (2017), I developed a new variable measuring the diversity of employment backgrounds of directors in a board. To solve any endogeneity problems, I use a difference-in-differences design that uses the implementation of the Dutch gender quota as an exogenous shock on firms to increase their board gender diversity. My results suggested that the quota only had a positive impact on the amount of academic degrees on a board, thus not drastically impacting the board composition characteristics of large Dutch firms.

Lastly, I hypothesize that the Dutch gender quota had a positive impact on the performance of Dutch firms. By using a difference-in-differences design and controlling for board size, board independence, firm size and firm leverage, I find no relationship between the implementation of

(6)

5 the Dutch gender quota and firm performance. Thus, my evidence implies that the decision for implementing a board gender quota should not be based on the criteria (future) financial performance.

This thesis contributes to the literature in two ways. First, this thesis contributes to the literature on board gender diversity, with a focus on board gender diversity in Europe and the Netherlands, as most existing literature is based on US samples (e.g. Adams & Fereirra, 2009; Carter, Simkins & Simpson (2003). Second, not much is known about the effects of gender regulation (Kirsch, 2017). This thesis contributes to the growing stream of literature about gender quotas and their effects and is the first study that looks at the Dutch gender quota in a quantitative perspective. Next to the contribution to literature, my results can support regulators making decisions about gender quotas. First, Dutch legislators can use this thesis to evaluate the existing quota. Furthermore, this thesis can also support foreign legislators to make decisions about quota laws that they wish to implement.

The remainder of my thesis is structured as follows. First, I elaborate on the background of gender legislation and quotas. In the section following the background I develop three hypotheses based on literature and theory. Next I describe the data used, the data collection, the variables used and the methodology of my thesis. After describing the methodology I discuss the results and in the last section I conclude and mention some suggestions for future research.

BACKGROUND

In this section I first describe the background of gender quotas in Europe, after which I go more in depth about the legislation in the Netherlands.

Gender legislation in Europe

In 1976 the Council of the European Union published a new Council Directive “on the implementation of the principle of equal treatment for men and women as regards access to employment, vocational training and promotion, and working conditions” (European Parliament, 1976). This was the first European law targeting the equal treatment of men and women in all of its Member States. Although the Directive was legally binding, it lacked a clear goal. Furthermore,

(7)

6 individual countries needed to devise their own views and laws on how to reach the objectives set by the Directive.

It was not until 2007 that the principle of equality between men and women became clearly established in the European Union. In the treaty of Lisbon, article 8 declares: “In all its activities, the Union shall aim to eliminate inequalities, and to promote equality, between men and women”. Directives on this equal treatment laid down a general framework for inequalities on the ground of among other things age and sex as regards employment (European Union, 2007).

About 6 years later, on 20 November 2013, the European Parliament backed a proposal with a quantitative objective of a 40 per cent presence of the under-represented sex among non-executive directors of large listed companies by 2020. In this case, a large listed company is defined as a company that employs more than 250 persons. It is, however, for each Member State of the Union to define how they accomplish this objective (Jourová, 2016).

In response to European legislation, multiple countries introduced boardroom gender quotas. Norway became the first country in 2003 to impose such a quota to combat gender inequalities in highly influential director positions. The Norwegian Public Limited Liabilities Companies Act requires both men and women to hold at least 40 per cent of the seats on the board. Because of this quota, in 2015, women hold more than 45 per cent of the board seats (Credit Suisse, 2016).

Dutch gender legislation

The first initiative for gender equality in top positions started by the Dutch Government was initiated in 2001 when the Ministries of Social Affairs and Employment and Economic Affairs started a network of ambassadors that needed to promote the flow of women to top positions. During their term, well-known business leaders aimed to carry out a number of actions points, both internally focused on their own company and externally focused on their branch (de Geus & van Gennip, 2006).

It took until 2006 before the Dutch Government quantified their goals of having more women at top positions in business with a revised ‘emancipation policy’, which stated that emancipation is not only beneficial for women, but also for the society as a whole (Lückerath-Rovers, 2015). Where the goal in 2001 was ‘more’ women at the top, the emancipation policy of 2006 declared a

(8)

7 target of 20 per cent women as executive and non-executive directors by 2020 (Second Chamber of the States General, 2015).

The first steps to legislation were made in 2008 when the ‘Partij van de Arbeid’ (the Labour Party) filed for a motion to include equality targets into the Dutch Corporate Governance Code. The target required that on both boards each gender must occupy at least 30 per cent of the seats (Kalma, van Vroonhoven-Kok, & Weekers, 2009). The motion was adopted after the majority of the Parliament voted in favor of the motion. However, the motion was ignored by the Minister of Finance that argued that the motion would affect the self-regulatory character of the Code (Lückerath-Rovers, 2015). In response, the Labour Party filed a bill to introduce the gender equality targets into the Dutch Civil Code. This resulted in the inclusion of the targets into the

Management and Supervision Act. This Act comprised of several additional proposals, some of

which were highly controversial (Lückerath-Rovers, 2015). These highly controversial and often highly critical proposals caused a lot of debate, and as a result it took until 1 January 2013 before the Act became active. The termination date of the Management and Supervision Act, 1 January 2016, was already set before they agreed about the implementation date. This meant that the gender quota was only active for three years before it became inactive. It took until 13 April 2017 before the gender quota was reinitiated and therefore extended to 1 January 2020. Between 1 January 2016 and 13 April 2017 companies were asked to act in accordance with the intention to extend the statutory quota.

The Dutch gender quota has no legal sanctions and is in accordance with the ‘comply-or-explain’ principle. This means that companies are expected to comply with the gender quota, but if that is not the case, these companies should explain in their annual report (1) why the director roles are not evenly distributed between the two genders, (2) what the company has done to achieve this balanced distribution and (3) what the company will do to achieve this balanced distribution.

HYPOTHESES DEVLOPMENT

In this section I develop my hypotheses based on literature and theory.

Studies suggests that a gender quota significantly increases board gender diversity (Ahern and Dittmar, 2012; Reguara-Alvarado et al., 2017). However, the evidence from Norway is not

(9)

8 decisive for the results in the Netherlands. First, because the evidence from Norway is based on a quota law that had large penalties for non-compliance. Second, the quota law of Norway has a different structure than that of the Netherlands. The Netherlands mandates a 30 per cent representation of each gender and does not have sanctions for non-compliance, while Norway mandates 40 per cent representation of each gender and non-compliance can lead to dissolution of the firm. Third, the time of implementation of the Norwegian quota law was around ten years earlier. The evidence from Spain however, is more indicative, as both quota laws do not have sanctions on non-compliance and work on a ‘comply-or-explain’ basis. Therefore, based on previous literature, I propose the following hypothesis about Dutch companies that need to comply with the Dutch gender quota:

H1: The Dutch gender quota has a significant effect on the proportion of female directors on the board.

At the implementation date of the Dutch gender quota, the only firm in compliance with the Dutch gender quota was Wolters Kluwer (Lückerath-Rovers, 2015). This meant that all firms, except one, suffered an exogenous impact of the quota to increase board gender diversity levels. Previous literature on Norway saw a significant effect of the Norwegian gender quota and changes in board characteristics (Ahern & Dittmar, 2012). Ahern & Dittmar (2012) argued that because the pool of new female directors is limited, the quota had a significant impact on several board characteristics. Their results suggested that boards after implementation of the quota were younger and had less CEO experience. They further state that characteristics of directors such as professional experience, age and education can have an impact on the advisory role of directors, but these characteristics have had little attention in literature on boards.

Next to board composition characteristics that are often present in today’s literature such as independence, size and tenure, this thesis looks at the consequences of a quota on the composition of the board regarding employment background diversity. Diversity on the board regarding professional expertise is argued to be an important component for the advising and monitoring functions of a board (Gray & Nowland, 2017). When a board has directors with different professional backgrounds, its advising ability improves, it acquires more perspectives and it gains a broader knowledge (Klein, 1998; Williams & O'Reilly III, 1998).

(10)

9 Despite the differences in enforcement regime, the time of implementation and the lower board gender diversity compliance obligation, I argued in the section above that the Dutch gender quota improves the proportion of women directors on the board. Following literature on Norway, I therefore hypothesize the following:

H2: The Dutch gender quota has a significant effect on the board characteristics.

Until now, literature on the relationship between board composition and firm financial performance showed mixed results (Kirsch, 2017; Rhode & Packel, 2011). The research of Carter, et al. (2003) resulted in the first empirical evidence of the relationship of board diversity and firm performance. Using a sample of 1000 US Fortune firms in 1997 and controlling for industry, size, and other governance measures, they found a statistically significant positive relationship between the percentages of women on boards and firm performance. In contrast with the positive relationship from Carter et al. (2003), Adams and Ferreira (2009) found a negative relationship between board gender diversity and firm performance, also by using an US sample. Lastly, Carter, D’Souza, Simkins and Simpson (2010) did not find results in line with neither Carter et al. (2003) nor Adams and Ferreira (2009) by concluding that there is no relationship between the proportion of women directors and financial firm performance.

One of the first papers that looked at the relationship between board gender diversity and firm performance in a Dutch setting was the paper of Lückerath-Rovers (2013). She built on the empirical methods of Joy, Carter, Wagner and Narayanan (2007) and Company (2007) with return on equity (ROE) as the measure for firm financial performance. Her results suggests that boards with female directors generally outperform boards without female directors. Next, Marinova, Plantenga and Remery (2016) did not only study the relationship in a Dutch setting, but also included Danish firms, as they argued that both countries are similar with regards to corporate governance and gender equality. In contrast with Lückerath-Rovers (2013) they did not find a statistically significant relationship between board gender diversity and firm performance.

The first empirical evidence from the consequences of mandated board gender diversity and firm performance came from the Norwegian Quota. The results of Bøhren and Strøm (2010) suggest that board gender diversity in a mandated setting has a negative effect on firm performance. In line with Bøhren and Strøm (2010), Ahern and Dittmar (2012) found, by also using a Norwegian

(11)

10 sample, that the gender quota resulted in a large decrease in the Tobin’s Q of the companies. Reguera-Alvarado et al., (2017) found, using a sample of 125 non-financial public listed Spanish Companies from 2005 to 2009, that the increase of female directors by the gender quota improved the economic results of the companies, which is in contrast with the Norwegian results.

Agency theory has been the main theoretical approach arguing that performance can increase when increasing diversity within the board (Jensen & Meckling, 1976). Jensen and Meckling (1976) describe that the independent directors play the monitoring and control function of the principal, thus reducing agency costs and thereby increasing performance.

Dalton and Dalton (2011) argue that the ability and willingness of a board to strive for the interest of the owners by monitoring the firm’s managers is related to their independence from management. There are two ways in which board independence can be increased. First, the appointment of more independent directors. Nygaard (2011), Terjesen, Couto and Fransisco (2015) and Bøhren & Staubo (2016) showed that an increase in the proportion of female directors increased the proportion of independent directors, which suggests that female directors are more often independent directors. Second, a more heteregeneous board is seen as a more independent board, with the heterogeneity being increased by board diversity. Board diversity is composed of multiple attributes, characterstics and skills of individuals, including gender (Van Der Walt & Ingley, 2003). As the heterogeneity on a board increases, the board gains a wider range of views which results in increased board independence (Adams, De Haan, Terjesen, & Van Ees, 2015; Hillman & Dalziel, 2003). This increased independence from management results in a board that is more effective in monitoring, controlling and advising, thus reducing agency costs and increasing performance (Jensen & Meckling, 1976).

As mentioned before, literature on the relationship between board gender diversity and firm performance, both in a worldwide and in a Dutch setting showed inconclusive results, which means that it is difficult to create an expectation of the gender quota on firm performance on past literature. Therefore, based on the agency theory I argue that because women increase the heterogeneity of a board and because women are more often independent directors, the increase in board gender diversity increases the monitoring function of the board. The better monitoring of the board, in return, reduces agency costs and therefore increases firm performance. Consequently, I propose the following hypothesis:

(12)

11

H3: The Dutch gender quota has a positive relationship with firm financial performance.

DATA AND METHODOLOGY

In this section I describe how the data for this thesis was collected, after which I describe the variables used and then follow up with the methodology used for testing my hypotheses.

Data collection

For the testing of my hypotheses, I use a sample that includes companies that are defined as a ‘large entity’ in the Dutch Civil Code. To qualify as a large entity, two of the following three criteria must be met: (1) the value of the total assets must be more than €17.5 million; (2) a net turnover of more than €35 million and/or (3) more than 250 number of employees on average. Firms with missing observations were excluded from the sample. In line with similar studies (e.g. Marinova et al., 2016; Reguera-Alvarado et al., 2017), I excluded financial firms because of their specific accounting. Since the gender quota became inactive from 1 January 2016 onwards, the cut-off date for the data is restricted to 2015, which is three years from when the quota was implemented. The final sample contains observations from 2010 to 2015, three years before and three years after the quota came into effect. By excluding companies that delisted during the sample period (2010-2015), survivorship bias could occur. Therefore, I choose not to exclude these companies and thus acquire an unbalanced yearly panel dataset.

Director level (employment background) data was hand-collected by a group of five students of the University of Groningen through annual reports, with supplementary public sources such as the company’s website and LinkedIn. Of the 7066 directors that were in our dataset, we hand-collected 4963 observations, around 70 per cent. Every unique director was classified by their employment backgrounds before they came an executive or non-executive director – academic, politician, social, HR, military, scientist, technical, accounting, business, financial, lawyer, or marketing. Directors that did not have publicly available information were labeled as ‘invisible’. APPENDIX B shows the students and amounts of data collected, the classification scheme used and the public sources from which we gathered information. Board data on firm level comes from the BoardEx database. Lastly, the Thomson Reuters Worldscope database was used to match market and accounting data with the board data.

(13)

12 These criteria, combined with a substantial time of data collection due to the timely nature of collecting all employment background information of the directors, results in a unique dataset of 588 firm observations and 4496 director observations. The final sample statistics regarding the firms and directors are reported in TABLE 1.

| Insert TABLE 1 here |

Dependent Variables

For testing hypothesis one, I use the dependent variable board gender diversity, measured as the number of female directors divided by the total board size, which is in line with previous literature (e.g. Adams & Ferreira, 2009, Ahern & Ditmar, 2012).

For the second hypothesis, to measure the impact of the gender quota on board characteristics, I make use of several dependent variables. First, unique to today’s literature is the investigation of a variable that looks at the diversity within a board regarding employment backgrounds of directors. For this, I devised a proxy which is the amount of different employment backgrounds that the directors have. In total we distinguished twelve different employment backgrounds. When the amount of different employment backgrounds in the board increases, the employment background diversity also increases. For the classification scheme refer to the previous section. The other dependent variables used to measure the consequences of the quota in regards to board composition are board size, board independence, board qualifications (amount of academic degrees), board networks, firm tenure, board tenure, experience in listed boards, current affiliations in listed boards, foreign board members and time until retirement.

Lastly, for the third hypothesis, to test whether the gender quota has a positive relationship with firm performance, I use two measures of firm performance. In line with previous research on quotas (e.g. Ahern & Dittmar, 2012; Reguera-Alvarado et al., 2017; Marinova, et al., 2016) and in line with most research on gender diversity and firm performance (e.g. Adams & Ferreira, 2009; Carter et al., 2003) I use Tobin’s Q as my first proxy of firm performance. I measure Tobin’s Q by dividing the market capitalization of a firm by the sum of the total assets and total liabilities of the firm. Furthermore, I use ROA as my second firm performance measure. Hagel, Brown, Samoylova and Lui (2013) state that ROA is not a perfect measure, but when assessing company

(14)

13 performance is it the most effective and broadly available measure there is. I measure ROA by dividing the earnings before interest and taxes by the total assets of the firm.

In my thesis I use both measures as proxies for firm performance because they complement each other, ROA to measure historical financial performance and Tobin’s Q to measure future financial performance.

Control variables

Consistent with existing literature on board gender diversity and firm performance I also include several control variables that have been shown to have a relationship with firm performance, of which two are board-specific and two are firm-specific (e.g. Campbell & Mínguez-Vera, 2008; Lui, Wei, & Xie, 2014)

The first board-specific control variable is board size, measured by number of directors. The relationship between board size and firm performance is extensively researched. For example, Hermalin and Weisbach’s (2003) findings from empirical literature on boards showed a predominantly negative relationship board size and firm performance. Yermack (1996) results suggests that board size and Tobin’s Q are inversely related, and Guest (2009) found that a firm’s profitability declines with larger boards, stating that that boards become less effective when having more directors. The other board-specific control variable is board independence, measured by the number of non-executive (independent) directors divided by the total number of directors in the board. In line with the agency theory, board independence has been found to be positively related with firm performance (e.g. Aggarwal, Erel, Stulz, & Williamson, 2010; Leung, Richardson, & Jaggi, 2014; Zhu, Ye, Tucker, & Kam, 2016).

The first firm-specific control variables is firm size. Most empirical evidence, although there are some exceptions, suggests that a positive relationship between firm size and firm financial performance exists (e.g. Lee, 2009; Stierwarld, 2009) In line with previous research I measure firm size by the natural logarithm of total assets (e.g. Lückerath-Rovers, 2013; Reguera-Alvarado et al., 2017). Next, I include leverage as a firm-specific control variable, which is calculated as the total debt divided by the total assets. Studies that found a negative relationship between leverage and firm performance include Chen, Firth, and Zhang (2008) and Ogebe, Ogebe and Alewi (2013).

(15)

14 Furthermore, I include ROA as control variable in the regressions with Tobin’s Q as dependent variable and vice versa. Not only are they both seen as valid indicators of financial performance (Gentry & Shen, 2010), but literature also shows that they are positively related to each other (e.g. Hoskisson, Johson, & Moesel, 1994; McGuire & Matta, 2003).

Lastly, I include board size as control variable in the difference-in-differences estimation in regards to the impact of the quota on employment background diversity in the board. Gray and Nowland (2017) found that the amount of directors with different expertise in the board is positively correlated with board size. This suggests that when a company has more board positions to fill, it is likely to diversify their professional expertise. I argue that the professional expertise of a director is closely related to their employment background, therefore I use board size as a control variable. APPENDIX B shows all variables used in this study, including their definitions and their sources.

(H1) Methodology influence of the Dutch gender quota on board gender diversity To determine whether the Dutch gender quota had any effect on the board gender diversity of large Dutch firms I use an approach similar to Ahern and Dittmar (2012), who based their approach on the approach of Stevenson (2010). I follow their approaches by using the variation in board gender diversity before the implementation of the quota (2012) as an instrument to seize the exogenous impact of the gender quota regarding board gender diversity. To verify whether the announcement and/or speculation of the gender quota did not impact the proportion of female directors in 2012 I compared the board gender diversity of Dutch listed firms in 2011 and 2012. I find that the gender diversity was identical in the majority of the firms. In my estimation I include fixed effects which account for individual firm characteristics that have a chance of influencing board gender diversity. Moreover, it reduces the risk of variables that are being omitted (Reeb, Sakakibara, & Mahmood, 2012).. I estimate the following equation:

𝐵𝐺𝐷_𝑖,𝑡 = 𝛽₀+ 𝛽₁2013𝐷𝑈𝑀𝑀𝑌 + 𝛽₂2014𝐷𝑈𝑀𝑀𝑌 + 𝛽₃2015𝐷𝑈𝑀𝑀𝑌 + 𝛽₄2013𝐷𝑈𝑀𝑀𝑌 ∗ 𝐵𝐺𝐷2012_𝑖 + 𝛽₅2014𝐷𝑈𝑀𝑀𝑌 ∗ 𝐵𝐺𝐷2012_𝑖 + 𝛽₆2015𝐷𝑈𝑀𝑀𝑌 ∗ 𝐵𝐺𝐷2012_𝑖 + 𝛼_𝑖 + λ_𝑡+ 𝑢_𝑖,𝑡 (1)

Where 𝐵𝐺𝐷_𝑖,𝑡 is the amount of female directors of firm i divided by the total amount of directors of firm i in year t. 2013𝐷𝑈𝑀𝑀𝑌 is coded ‘1’ if the year is 2013, else coded ‘0’. 2014𝐷𝑈𝑀𝑀𝑌 is

(16)

15 coded ‘1’ if the year is 2014, else coded ‘0’. 2015𝐷𝑈𝑀𝑀𝑌 is coded ‘1’ if the year is 2015, else coded ‘0’. 𝐵𝐺𝐷2012_𝑖 is the amount of female directors of firm i divided by the total amount of directors of firm i in the year 2012. 𝛼_𝑖 represents the firm fixed effects, which include but are not limited to industry effects. λ_𝑡 represents the year fixed effects and 𝑢_𝑖,𝑡 represents the idiosyncratic error term.

(H2) Methodology influence of the Dutch gender quota on board characteristics To test whether the implementation of the gender quota on 1 January 2013 had any effect on the board characteristics of Dutch firms I use a difference-in-differences approach. In this approach the implementation of the gender quota law is seen as an exogenous shock on firms to increase their board gender diversity, which eliminates any endogeneity problems that other statistical tests may have. The difference-in-differences approach, assuming that the control group is suitable, solves the problems of unobserved firm-specific heterogeneity and reverse-causality (Reeb, Sakakibara, & Mahmood, 2012). Furthermore, in line with the previous section I also include fixed effects. The difference-in-differences design model is estimated in equation (2):

𝐵𝑂𝐴𝑅𝐷𝐶𝐻𝐴𝑅𝑖,𝑡 = 𝛽0+ 𝛽1𝑇𝑅𝐸𝐴𝑇𝑀𝐸𝑁𝑇𝑖+ 𝛽2𝑃𝑂𝑆𝑇𝑄𝑈𝑂𝑇𝐴𝑡+ 𝛽3𝑇𝑅𝐸𝐴𝑇𝑀𝐸𝑁𝑇𝑖

∗ 𝑃𝑂𝑆𝑇𝑄𝑈𝑂𝑇𝐴𝑡+ 𝛼𝑖 + λ𝑡+ 𝑢𝑖,𝑡 (2)

Where 𝐵𝑂𝐴𝑅𝐷𝐶𝐻𝐴𝑅𝑖,𝑡 represents the different board composition variables used in this thesis –

board size, board independence, board qualifications, board networks, firm tenure, board tenure, experience in listed boards, current affiliations in listed boards, foreign board members and time until retirement. 𝑇𝑅𝐸𝐴𝑇𝑀𝐸𝑁𝑇𝑖 is a dummy variable coded ‘1’ if firm i is a Dutch firm, ‘0’ if firm

i is a Finnish firm. 𝑃𝑂𝑆𝑇𝑄𝑈𝑂𝑇𝐴_𝑡 is a dummy variable coded ‘1’ if the year of the observation is 2013, 2014 or 2015 and 𝑇𝑅𝐸𝐴𝑇𝑀𝐸𝑁𝑇𝑖∗ 𝑃𝑂𝑆𝑇𝑄𝑈𝑂𝑇𝐴𝑡 is the interaction term. 𝛼𝑖 represents the

firm fixed effects, which include but are not limited to industry effects. λ_𝑡 represents the year fixed effects and 𝑢𝑖,𝑡 represents the idiosyncratic error term.

Finland as a control group

Because all large entities in the Netherlands need to follow the gender quota law, it is difficult to find a perfect suitable control group for the difference-in-differences approach. This control group functions as a proxy for how large public listed Dutch firms would have behaved if the quota was

(17)

16 not implemented on 1 January 2013. As large private listed Dutch firms do not have all the necessary information available and the differences between the larger and smaller Dutch firms are too large, large public listed companies of another, similar country had to be selected.

To find the best suitable country, I imposed several criteria. First, all non-European countries were eliminated, as the cultural and economical differences would be too significant. Second, all European countries that had a gender quota or gender target from 2010 until 2015 were also eliminated. Third, the country has to be comparable with the Netherlands based on the gender equality index and in line with the results of two published works of Hofstede (1980, 1984) about international cultural differences.

Based on these criteria I had two countries to choose from, Austria and Finland. First looking at the labor force participation rates of females, I found that Finland and the Netherlands are almost perfectly similar in the whole sample period (Both countries had 72.6 per cent in 2010, 74.4 per cent in Finland and 74.7 per cent in the Netherlands in 2015, whereas Austria had 67.6 per cent in 2010 and 71 per cent in 2015, source: online database ILOSTAT). In addition, after visually comparing both countries with the Dutch sample on the firm performance measures ROA, Tobin’s Q and as a robustness check ROE, I accepted Finland as the most suitable control group as a proxy for how the Dutch firms would have been if there had been no gender quota. The three years average ROA, Tobin’s Q and ROE from the Dutch firms and comparable (see the upcoming section for criteria) Finnish firms are shown in respectively figure 1, 2 and 3 in APPENDIX A.

(H3) Methodology influence of the Dutch gender quota on firm performance

To measure the impact of the Dutch gender quota on firm performance, I use a difference-in-differences approach in line with the previous section. Furthermore, in line with previous research I control for various variables that have shown to have a relationship with firm performance (e.g. Reguara-Alvarado et al., 2017). The difference-in-differences approach with control variables is estimated in equation (3):

𝑃𝐸𝑅𝐹𝑂𝑅𝑀𝐴𝑁𝐶𝐸_𝑖,𝑡 = 𝛽₀+ 𝛽₁𝑇𝑅𝐸𝐴𝑇𝑀𝐸𝑁𝑇_𝑖 + 𝛽₂𝑃𝑂𝑆𝑇𝑄𝑈𝑂𝑇𝐴_𝑡+ 𝛽₃𝑇𝑅𝐸𝐴𝑇𝑀𝐸𝑁𝑇_𝑖

∗ 𝑃𝑂𝑆𝑇𝑄𝑈𝑂𝑇𝐴𝑡+ 𝛽4𝑅𝑂𝐴/𝑇𝑂𝐵𝐼𝑁𝑆𝑄𝑖,𝑡+ 𝛽5𝐹𝑆𝐼𝑍𝐸𝑖,𝑡+ 𝛽6𝐿𝐸𝑉𝐸𝑅𝐴𝐺𝐸𝑖,𝑡

(18)

17 Where 𝑃𝐸𝑅𝐹𝑂𝑅𝑀𝐴𝑁𝐶𝐸_𝑖,𝑡 represents either Tobin’s Q, the market capitalization of firm i divided by the sum of total assets and total liabilities of firm i in year t or ROA, the earnings before interest and taxes of firm i divided by the total assets of firm i in year t. 𝑇𝑅𝐸𝐴𝑇𝑀𝐸𝑁𝑇𝑖 is a dummy variable

coded ‘1’ if firm i is a Dutch firm, ‘0’ if firm i is a Finnish firm. 𝑃𝑂𝑆𝑇𝑄𝑈𝑂𝑇𝐴_𝑡 is a dummy variable coded ‘1’ if the year of the observation is 2013, 2014 or 2015 and 𝑇𝑅𝐸𝐴𝑇𝑀𝐸𝑁𝑇_𝑖 ∗ 𝑃𝑂𝑆𝑇𝑄𝑈𝑂𝑇𝐴_𝑡 is the interaction term. 𝛽₄𝑅𝑂𝐴/𝑇𝑂𝐵𝐼𝑁𝑆𝑄_𝑖,𝑡 represents the control variable ROA when the dependent variable is Tobin’s Q and vice versa. 𝐹𝑆𝐼𝑍𝐸_𝑖,𝑡 is the natural logarithm of the total assets of firm i in year t. 𝐿𝐸𝑉𝐸𝑅𝐴𝐺𝐸_𝑖,𝑡 is the total liabilities of firm i divided by the total assets of firm i in year t. 𝐵𝑆𝐼𝑍𝐸_𝑖,𝑡 is the amount of directors at firm i in year t. 𝐵𝐼𝑁𝐷𝑃_𝑖,𝑡 is the amount of independent (non-executive) directors of firm i divided by the total amount of directors of firm i in year t. 𝛼_𝑖 represents the firm fixed effects, which include but are not limited to industry effects. λ_𝑡 represents the year fixed effects and 𝑢_𝑖,𝑡 represents the idiosyncratic error term.

The main coefficient of interest in these equation is 𝛽3, that can also be written as:

𝛽₃ = (∆𝑃𝐸𝑅𝐹𝑂𝑅𝑀𝐴𝑁𝐶𝐸̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ 𝑇𝑟𝑒𝑎𝑡𝑒𝑑_{) − (∆𝑃𝐸𝑅𝐹𝑂𝑅𝑀𝐴𝑁𝐶𝐸}_{̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅}𝐶𝑜𝑛𝑡𝑟𝑜𝑙₎

If the coefficient is statistically significant it indicates that the implementation of the quota had an influence on the financial performance of Dutch firms. Conversely, if the coefficient is not significant, it indicates that the quota had no effect on the financial performance of Dutch firms.

RESULTS

In this section I start with describing the descriptive statistics of my sample, after which I conduct a Pearson Correlation Matrix and lastly, I discuss the results of my three hypotheses.

Descriptive statistics

In TABLE 2 I report the descriptive statistics for my sample of Dutch and Finnish firms, along with a difference in mean test. Looking at the Dutch firm and board characteristics in column (1)-(4), the table shows that the average Tobin’s Q was 0.61 with a standard deviation of 0.46 and that the average ROA was 0.05 with a standard deviation of 0.09. On average, 11 per cent of the three to thirteen director seats were occupied by female directors. Furthermore, the descriptive statistics show that not all firms were profitable, as ROA has a minimum value of -0,29, that not all firms

(19)

18 had at least one woman director in their board in the period 2010 to 2015 and that at least one firm did not have any debt in 2010 to 2015.

Next to the Dutch statistics, I present the descriptive statistics for the Finnish firms in columns (5)-(8) and column (9) and (10) report the difference in mean test results between the Dutch and Finnish board and firm characteristics. It is notable that there are quite a lot of differences between the two countries in terms of board composition. I find that Finnish boards had on average more academic degrees in the board, more experience in listed boards, more current affiliations with listed boards, are younger, have a higher proportion of female directors, have a higher proportion of independent directors and lastly, Finnish firms have, on average, a higher leverage. Dutch firms were on average larger, had more foreign board members and carried more different types of employment backgrounds in their boards.

TABLE 3 shows the descriptive statistics and difference in mean test for characteristics of Dutch boards before and after the implementation of the quota. It is notable that the differences between both groups only exists on the characteristic board gender diversity. Furthermore, there are still boards after the implementation of the quota that do not have any female directors. Lastly, notable is that the firm with the smallest board (three directors) appointed a new male director while they had already three male directors.

Next, TABLE 4 shows the descriptive statistics and difference in mean test for female directors before and after the implementation of the quota in panel A. Although there are some small differences between the groups, none of them are significant. Consequently, panel B shows the descriptive statistics and difference in mean test for female directors employed before the implementation of the quota and those appointed after the implementation. The differences between the two are significant. The female directors appointed after the quota are, on average, less tenured in the board and in the firm, have less lifetime experience in listed boards and have fewer current affiliations in listed boards. It is notable that the amount of female directors that were

(20)

19 newly appointed after the quota is quite low (n=13). Thus, it seems that most of the appointed female directors were already active as a director at another listed firm.

Lastly, TABLE 5 shows the descriptive statistics and difference in mean test for male and female directors. The results show that, on average, male directors have more board experience, are more tenured in their board and firm and are older than the female directors. Female directors, on the other hand, have on average more academic degrees and have a larger network.

Pearson correlation coefficient

I conducted a Pearson correlation coefficient (TABLE 6) to gather an initial thought about the relationship of the variables used for my research. Overall, there seems to be no statistically significant correlation between board gender diversity and the two measures of firm financial performance, Tobin’s Q and ROA (p>0.1). Furthermore, it seems that there are significant correlations between board independence, firm size and firm leverage and one or both of the two firm financial performance measures. ROA and Tobin’s Q also seem to be significantly correlated. Lastly, the results suggests that board gender diversity is related to almost all board characteristics that I used for the testing of hypothesis two, except for ‘current affiliations with listed firms’ and ‘foreign board members’.

(H1) Influence of the Dutch gender quota on board gender diversity

TABLE 7 shows the results of the approach similar to Ahern & Dittmar (2012). The results show significantly increased board gender diversity levels after the implementation of the quota, with roughly the same coefficients for all three years. The coefficients of the year dummies are respectively 0.042, 0.063 and 0.082, all significant at the 1 per cent level. The coefficients of the interaction terms consisting of the year dummies multiplied by the board gender diversity in year 2012 are respectively -0.199 (p<0.05), -0.162 and (p<0.1) and -0.216 (p<0.05). TABLE 7 shows results in support of hypothesis 1, which is also in line with the evidence from Norway and Spain (Reguera-Alvarado et al. 2017, Ahern & Dittmar, 2012).

(21)

20 It is notable that the second and third year coefficients are very different (0.063 and 0.082 for the Netherlands, 0.128 and 0.212 for Norway, all significant at the 1 per cent level). This can be attributed to the fact that the Dutch gender quota is according to the ‘comply-or-explain’ principle and has no sanctions in place.

(H2) Influence of the Dutch gender quota on board characteristics

I report the results of the impact of the Dutch gender quota on board characteristics through a difference-in-differences research design with Finland as the control group in TABLE 8. From TABLE 8, I note that the interaction terms for all board characteristics are statistically insignificant at the 10 per cent level except for board qualifications, which is statistically significant at the 10 per cent level and has a coefficient of 0.023.

The results suggest that the Dutch gender quota did not have any significant effect on board characteristics of the large, public Dutch listed firms, thus not supporting hypothesis 2, which is not in line with previous studies from Ahern & Dittmar (2012). This can be attributed to the lower levels of board gender diversity in the Netherlands compared with Norway. Dutch firms are less constrained to find female directors with the same characteristics of the current directors. This can also be concluded by the low amount of newly appointed female directors (n=13) in my sample.

(H3) Influence of the Dutch gender quota on firm performance

TABLE 9 reports difference-in-differences design results for the effects of the different independent and control variables on firm financial performance. The results of equation (3) without control variables is shown in model (1) and (2). The results of the same equations with control variables are shown in model (3) and (4).

Looking at TABLE 9, model (1) and (2) without the control variables, the interaction term (DiD), my main variable of interest, did not show any significant results. The treatment dummy shows significant results in all models, although it is notable that the coefficient in the model including

(22)

21 control variables with Tobin’s Q as performance measure is negative. The quota dummy for the period 2013-2015 shows a positive coefficient significant at the 1 per cent level for the dependent variable Tobin’s Q, for ROA the results are insignificant. When adding the control variables in model (3) and (4) the main variable of interest does not change significance levels, thus resulting in insignificant coefficients of -0.012 and 0.007 (p>0.10). ROA and Tobin’s Q both have significant coefficients at the 1 per cent level, respectively 0.660 and 0.061. Firm size is only significant in model (4) (p<0.01), while leverage is in both models significant at the 1 per cent level with negative coefficients. The constant is not significant in the model with Tobin’s Q as dependent variable, while it is significant at the 10 per cent level for ROA. The coefficients are respectively 1.798 and 0.529. Board size and board independence are in line with the fixed effects regression not statistically significant.

The results of the difference-in-differences model showed no statistically significant interaction term. This means that my results suggest, contrary to my expectations based on the agency theory, that there is no relationship between the Dutch gender quota and the increase in firm financial performance, thus also rejecting hypotheses 3.

Robustness

First, next to including fixed-effects in the regressions, creating lagged variables is also a commonly used way to reduce endogeneity (e.g. Aschhoff & Schmidt, 2008; Bania, Gray, & Stone, 2007), although it is not empirically justified (Reed, 2015). I use the difference-in-differences model on firm performance with lagged variables to verify my results, from which the results are shown in TABLE 10. Second, I use a multivariate OLS regression with fixed effects to verify the results of the difference-in-differences design. As this test does not need a control group, I eliminate bias that the Finnish sample introduced, although the approach includes endogeneity problems. The result of this test are shown in TABLE 11. Lastly, in several unreported analyses I conducted all main tests with samples that were (1) not winsorized and (2) winsorized at the 5 per cent level (5 95).

I report that the results of the robustness checks, in line with my main empirical results, showed no significant changes in board composition after the implementation of the quota and no significant relationship between the quota and firm performance.

(23)

22

CONCLUSION

The aim of my thesis was to seek more clarity of the consequences of the Dutch gender quota regarding gender diversity, board composition and financial aspects. First, by using the variation in board gender diversity before the implementation of the quota, I found that the Dutch gender quota, despite not having legal sanctions for non-compliance, had a positive impact on board gender diversity. This is in line with previous literature (e.g. Ahern & Dittmar, 2012; Reguera-Alvarado et al., 2017). Furthermore, by using a difference-in-differences design to examine the impact of the gender quota on board composition, my results suggests that only the amount of academic degrees on the board has increased. Previous literature showed a significant change in board composition (Ahern & Dittmar, 2012). I argue that this can be attributed to the lower amount of board gender diversity in the Netherlands, which means that Dutch listed firms are less constrained to find female directors with the same characteristics as the current directors. Lastly, also by using a difference-in-differences design, my results suggests that the Dutch gender quota did not have any impact on firm performance. This indicates that gender in the boardroom alone does not impact firm performance.

The main limitation of my thesis is the shortcoming regarding external validity of the results. As discussed earlier in my thesis, most quotas have different structures, countries have different cultures and the economic performances between countries differ also. This means that it is difficult to generalize the findings to other countries that implemented or want to implement a quota. Future research should focus on doing a qualitative study regarding quotas whereby the main findings in Norway, Spain and the Netherlands are discussed whereby new insights can be gathered on quotas in general.

(24)

23

REFERENCES

Adams, R., & Ferreira, D. (2009). Women in the boardroom and their impact on governance and performance. Journal of Financial Economics, 94(2), 291-309.

Adams, R., De Haan, J., Terjesen, S., & Van Ees, H. (2015). Board Diversity: Moving the Field Forward. Corporate Governance: An International Review, 23(2), 77-82.

Aggarwal, R., Erel, I., Stulz, R., & Williamson, R. (2010). Differences in governance practices between US and foreign firms: Measurement, causes, and consequences. Review of

Financial Studies, 23(3), 3131-3169.

Ahern, K. R., & Dittmar, A. K. (2012). The Changing of the Boards: The impact on Firm Valuation of Mandated Female Board Representation. Quarterly Journal of Economics, 127, 137-197.

Aschhoff, B., & Schmidt, T. (2008). Empirical evidence on the success of R&D cooperation - Happy together? Review of Industrial Organization, 33, 41-62.

Bania, N., Gray, J. A., & Stone, J. A. (2007). Growth, taxes, and government expenditures: Growth hills for U.S. states. National Tax Journal, LX, 193-204.

Bøhren, Ø., & Staubo, S. (2014). Does mandatory gender balance work? Changing organizational form to avoid board upheaval. Journal of Corporate Finance, 28, 152-168.

Bøhren, Ø., & Staubo, S. (2016). Mandatory gender balance and board independence. European

Financial Management, 22(1), 3-30.

Bøhren, Ø., & Strøm, Ø. (2010). Governance and Politics: Regulating Independence and Diversity in the Board Room. Journal of Business Finance & Accounting, 37(9), 1281-1308.

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

(25)

24 Carter, D. A., D'Souza, F., Simkins, B. J., & Simpson, W. (2010). The Gender and Ethnic Diversity of US Boards and Board Committees and Firm Financial Performance. Corporate

Governance: An International Review, 18(5), 396-414.

Carter, D. A., Simkins, B., & Simpson, W. (2003). Corporate Governance, Board Diversity, and Firm Value. Financial Review, 38(1), 33-53.

Chen, G., Firth, M., & Zhang, W. W. (2008). The efficiency and profitability effects of China's modern enterprise restructuring programme. Review of Accounting, 16(1), 74-91.

Company, M. &. (2007). Women matter. Gender diversity, a corporate performance driver. Paris.

Retrieved from

http://www.mckinsey.com/locations/swiss/news_publications/pdf/women_matter_english .pdf

Credit Suisse. (2016). The CS Gender 3000: The reward for change.

Dalton, D., & Dalton, C. (2011). Integration of micro and macro studies in governance research: CEO duality, board composition, and financial performance. Journal of Management,

37(2), 404-411.

de Geus, A. J., & van Gennip, C. E. (2006). Rapport Ambassadeursnetwerk gevolgd. Retrieved April 12, 2018, from https://zoek.officielebekendmakingen.nl/kst-27099-17.html

European Parliament. (1976). Council Directive of 9 February 1976, 1976 O.J. L 039 of 14/02/1976, on the principle of equal treatment for men and women as regards access to employment, vocational training and promotion, and working conditions. P0040-0042.

European Union. (2007, December 13). Treaty of Lisbon Amending the Treaty on European Union and the Treaty Establishing the European Community. 2007/C 306/01. Retrieved April 11, 2018, from https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:12007L/TXT

Gentry, R. J., & Shen, W. (2010). The Relationship between Accounting and Market Measures of Firm Financial Performance: How Strong Is It? Journal of Managerial Issues, 12(4), 514-530.

(26)

25 Gray, S., & Nowland, J. (2017). The diversity of expertise on corporate boards in Australia.

Accounting & Finance, 57(2), 429-463.

Guest, P. M. (2009). The Impact of Board Size on Firm Performance. The European Journal of

Finance, 15(4), 385-404.

Hagel, J., Brown, J., Samoylova, T., & Lui, M. (2013). Succes or struggle: ROA as a true measure of business performance. Report 3 of the 2013 Shift Index series. Deloitte University Press.

Hermalin, B. E., & Weisbach, M. S. (2003). Boards of directors as an endogenously determined institution: A survey of the economic literature. Economic Policy Review, 9, 7-26.

Hillman, A., & Dalziel, T. (2003). Boards of directors and firm performance: integrating agency and resource dependence perspectives. Academy of Management Review, 98(2), 383-396.

Hofstede, G. (1980). Motivation, leadership, and organization: do American theories apply abroad? Organizational dynamics, 9(1), 42-63.

Hofstede, G. (1984). Culture's consequences: International differences in work-related values (Vol. 5). Sage.

Hoskisson, R., Johson, R., & Moesel, D. (1994). Corporate Divestiture Intensity in Restructuring Firms - Effects of Governance, Strategy, and Performance. Academy of Management

Journal, 37, 1207-1251.

Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305-360.

Jourová, V. (2016). Gender balance on corporate boards, Europe is cracking the glass ceiling.

European Commision. Retrieved April 11, 2018, from

ceiling.ec.europa.eu/newsroom/document.cfm?doc_id=46280

Joy, L., Carter, N. M., Wagner, H. M., & Narayanan, S. (2007). The bottom line: corporate performance and women's representation on boards. Catalyst, 3(1).

(27)

26 Kalma, van Vroonhoven-Kok, & Weekers. (2009). Gewijzigd amendement van het lid Kalma C.S.

ter vervanging van dat gedrukt onder nr. 8. Retrieved April 12, 2018, from

https://zoek.officielebekendmakingen.nl/kst-31763-14.html

Kirsch, A. (2017). The gender composition of corporate boards: A review and research agenda.

The Leadership Quarterly.

Klein, A. (1998). Firm performance and board committee structure. The Journal of Law and

Economics, 41(1), 275-304.

Lee, J. (2009). Does Size Matter in Firm Performance? Evidence from US Public Firms.

International Journal of the Economics of Business, 189-203.

Lending, C. C., & Vähämaa, E. (2017). European board structure and director expertise: The impact of quotas. Research in International business and Finance, 39, 485-501.

Leung, S., Richardson, G., & Jaggi, B. (2014). Corporate board and committee independence, firm performance, and family ownership concentration: An analysis based on Hong Kong firms.

Journal of Contemporary Accounting and Economics, 10(1), 16-31.

Lückerath-Rovers, M. (2013). Women on boards and firm performance. Journal of Management

& Governance, 17(2), 491-509.

Lückerath-Rovers, M. (2015). Gender quota in the boardroom: the Dutch approach. Deakin Law

Review, 20(1), 75-94.

Lui, Y., Wei, Z., & Xie, F. (2014). Do wemen directors improve firm performance in China?

Journal of Corporate Finance, 28, 169-184.

Marinova, J., Plantenga, J., & Remery, C. (2016). Gender Diversity and Firm Performance: Evidence from Dutch and Danish Boardrooms. Discussion Paper Series / Tjalling C.

Koopmans Research Institute, 10(03).

Matsa, D., & Miller, A. (2013). A female style in corporate leadership? Evidence from quotas.

(28)

27 McGuire, J., & Matta, E. (2003). CEO Stock Options: The Silent Dimension of Ownership.

Academy of Management Journal, 46, 255-265.

Nygaard, K. (2011). Forced board changes: Evidence from Norway. Manuscript. Norwegian School of Economics and Business Administration (NHH), Bergen, Norway.

Ogebe, P., Ogebe, J., & Alewi, K. (2013). The Impact of Capital Structure on Firm's Performance in Nigeria. Munch Personal RePec Archiv (MPRA).

Reeb, D., Sakakibara, M., & Mahmood, I. P. (2012). From the Editors: Endogeniety in International Business Research. Journal of International Business Studies, 43, 211-218.

Reed, W. (2015). On the practice of lagging variables to avoid simultaneity. Oxford Bulletin of

Economics and Statistics, 77(6), 897-905.

Reguera-Alvarado, N., de Fuentes, P., & Laffarga, J. (2017). Does Board Gender Diversity Influence Financial Performance? Evidence from Spain. Journal of Business Ethics,

141(2), 337-350.

Rhode, D. L., & Packel, A. K. (2011). Diversity on Corporate Boads: How Much Difference Does Difference Make? SSRN Electronic Journal.

Rijksoverheid. (2016, 01 15). Opnieuw wettelijk streefcijfer voor vrouwen en mannen in bestuur

ondernemingen. Retrieved from

https://www.rijksoverheid.nl/actueel/nieuws/2016/01/15/opnieuw-wettelijk-streefcijfer-voor-vrouwen-en-mannen-in-bestuur-ondernemingen

Second Chamber of the States General. (2015). Meerjarrenbeleidsplan Emancipatie 2006-2010. Emancipatie: vanzelfsprekend, maar het gaan niet vanzelf! Retrieved 12 4, 2018, from https://zoek.officielebekendmakingen.nl/kst-30420-2.html

Staatsblad. (2017). Artikel 166 van de Wet van 10 februari 2017. 188.

Stevenson, B. (2010). Beyond the Classroom: Using Title IX to Measure the Return to High School Sports. Review of Economics and Statistics, 92, 284-301.

(29)

28 Stierwarld, A. (2009). Determinants of Firm Profitability - The Effect of Productivity and its

Persistence. Melbourne Insititute of Applied Economic and Social Research, 1-25.

Terjesen, S., Couto, E. B., & Fransisco, P. M. (2015). Does the presence of independent and female directors impact firm performance? A multi-country study of board diversity. Journal of

Management & Governance, 1-37.

Van Der Walt, N., & Ingley, C. (2003). Board dynamics and the influence of professional background, gender and ethnic diversity of directors. Corporate Governance, 11(3), 218-234.

Williams, K. Y., & O'Reilly III, C. A. (n.d.). Demography and Diversity in Organisations: A Review of 40 years of research in BM Staw and LL Cummings (eds) Research in Organisational Behaviour Vol. 20. Connectivut: Jai Pres.

Yermack, D. (1996). Higher Market Valuation of Companies with a Small Board of Directors.

Journal of Financial Economics, 40(2), 185-221.

Zhu, J., Ye, K., Tucker, J. W., & Kam, J. C. (2016). Board hierarchy, independent directors, and firm value: Evidence from China. Journal of Corporate Finance, 41, 262-279.

(30)

29

TABLES

TABLE 1

In this table I report the amount of yearly observations after the criteria reported in the ‘Data collection’ section.

Sample statistics

Firms (Obs.) Directors (Obs.)

Year Dutch Finnish Total Dutch Finnish Total

2010 49 41 90 411 296 707 2011 48 47 95 404 344 748 2012 48 49 97 406 352 758 2013 46 49 95 372 345 717 2014 51 55 106 403 383 786 2015 51 54 105 404 376 780 Total 293 295 588 2400 2096 4496

(31)

0 TABLE 2 Summary Statistics Variables Netherlands Finland Netherlands - Finland (N = 293) (N = 295)

Mean SD Min Max Mean SD Min Max Diff. (t-stat)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

Board Qualification 0.21 0.09 0 0.68 0.26 0.07 0.05 0.66 -0.05*** -6.63

Board Networks 68.97 47.53 8 216 70.07 41.65 4.47 261 -1.10 -0.30

Firm Tenure 0.91 0.56 0 3.60 0.87 0.48 0.08 2.66 0.05 1.05

Board Tenure 0.75 0.49 0 3.52 0.79 0.43 0.03 2.50 -0.05 -1.19

Lifetime Exp. Listed Boards 0.39 0.15 0.20 1.08 0.47 0.19 0.10 1.44 -0.08*** -5.46

Curr. Aff. Listed Boards 0.22 0.07 0.09 0.64 0.27 0.09 0.08 0.77 -0.04*** -5.94

Foreign Board Members 0.27 0.24 0 1 0.20 0.19 0 0.75 0.07*** 3.81

Time Until Retirement 1.25 0.82 -1.72 4.62 1.88 0.84 0 6.14 -0.63*** -9.23

Board Gender Diversity 0.11 0.11 0 0.40 0.24 0.11 0 0.50 -0.13*** -13.83

Board Empl. Back. Diversity 0.45 0.15 0.13 1 0.35 0.15 0.13 0.80 0.10*** 7.91

Tobin’s Q 0.61 0.46 0.06 2.79 0.67 0.57 0.06 2.79 -0.06 -1.39 ROA 0.05 0.09 -0.29 0.34 0.06 0.08 -0.29 0.34 -0.01 -1.33 Board Size 8.17 2.45 3 13 7.10 1.51 3 11 1.07*** 6.35 Board Independence 0.55 0.20 0 0.90 0.88 0.15 0 1 -0.33*** -22.27 Firm Size 14.20 1.89 9.44 19.28 14.00 1.52 9.44 17.44 0.20 1.42 Leverage 0.20 0.12 0 0.58 0.24 0.12 0 0.58 -0.03*** -3.27

In this table I report the summary statistics of my final sample. Columns (1)-(4) reports the summary statistics of the Dutch firms, columns (5)-(8) reports the summary statistics of the Finnish firms and finally, columns (9) and (10) reports the results of the difference in mean t-test.

Continuous variables are normalized using board size.

(32)

1

TABLE 3

Difference between various board characteristics before and after the implementation of the quota

Before quota After quota

(N = 148) Before - After

Variables (N = 145)

Mean SD Min Max Mean SD Min Max Diff. (t-stat)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

Board Qualification 0.21 0.09 0 0.42 0.22 0.09 0.06 0.69 -0.02 -1.51

Board Networks 66.91 47.74 8.96 216.43 70.98 47.41 8 203.92 -4.06 -0.73

Firm Tenure 0.94 0.56 0 3.60 0.89 0.57 0.02 2.79 0.05 0.71

Board Tenure 0.74 0.48 0 3.52 0.76 0.50 0.02 2.47 -0.02 -0.32

Lifetime Exp. In Listed Boards 1.01 0.37 0.20 2.48 1.00 0.40 0.20 3.04 0.01 0.27

Current Aff. In Listed Boards 0.23 0.09 0.09 0.64 0.23 0.07 0.11 0.48 0.01 0.81

Foreign Board Member 0.25 0.24 0 1 0.29 0.24 0 1 -0.04 -1.38

Time Until Retirement 1.27 0.84 -1.52 4.62 1.24 0.82 -1.72 4.42 0.02 0.24

Board Gender Diversity 0.10 0.10 0 0.40 0.14 0.11 0 0.40 -0.04*** -2.99

Board Empl. Background diversity 0.45 0.16 0.13 1 0.44 0.14 0.14 0.83 0.01 0.58

Board Size 8.40 2.54 3 13 7.97 2.35 4 13 0.45 1.56

Board Independence 0.55 0.20 0 0.91 0.55 0.21 0 0.91 0.00 0.01

In this table I report the difference in means t-test results for various board and firm characteristics between the three year period before the implementation of the quota and the three year period after the implementation of the quota.

(33)

0

TABLE 4

Board attributes of female Dutch directors

Mean SD Mean SD Diff (t-stat)

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

Panel A

Variables Before the Quota After the Quota Before - After

(N = 126) (N = 174)

Board Qualifications 0.24 0.16 0.26 0.17 -0.02 -0.85

Board Networks 94.08 107.29 118.51 130.85 -24.43 -0.19

Firm Tenure 0.60 0.91 0.55 0.84 0.05 0.50

Board Tenure 0.56 0.82 0.55 0.80 0.01 0.10

Lifetime Exp. In Listed boards 0.86 0.69 0.78 0.61 0.08 1.06

Current Aff. in Listed boards 0.47 0.37 0.43 0.32 0.03 0.84

Foreign Board Member 0.36 0.48 0.40 0.49 -0.04 -0.8

Time Until retirement 1.76 0.91 1.70 1.00 0.06 0.52

Panel B

Before Quota Appointed after Quota _{Before - After}

(N = 287) (N = 13)

Qualifications 0.26 0.17 0.19 0.11 0.07 1.41

Network 109.33 122.72 84.33 103.80 25.00 0.72

Board Tenure 0.57 0.82 0.10 0.04 0.47** 2.05

Firm Tenure per Board 0.59 0.88 0.12 0.15 0.47* 1.93

Lifetime Exp. In Listed boards 0.83 0.65 0.25 0.25 0.58*** 3.2

Current Aff. in Listed boards 0.46 0.35 0.21 0.17 0.25** 2.54

Dummy: Foreign B. Member 0.38 0.49 0.46 0.52 -0.08 -0.59

Time Until retirement 1.71 0.97 2.11 0.77 -0.40 -1.46

In this table I report the attributes of female directors in my sample. Column (5) and (6) reports the results of the difference in mean t-test. In Panel A I report the mean and standard deviations of women before and after the implementation of the quota. In Panel B I report the mean and standard deviations of women appointed before the implementation of the quota and the new female directors appointed after the implementation of the quota.

(34)

1

TABLE 5

Board attributes male and female directors

Variables

Men Women

Men - Women

(N=2100) (N=300)

Mean SD Mean SD Diff (t-stat)

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

Qualifications 0.20 0.15 0.25 0.16 -0.06*** -5.95

Network 67.81 88.63 108.25 121.92 -40.44*** -7.01

Board Tenure 0.72 0.69 0.55 0.81 0.16*** 3.78

Firm Tenure 0.88 0.90 0.57 0.87 0.31*** 5.55

Lifetime Exp. In Listed boards 0.98 0.85 0.81 0.65 0.17*** 3.35

Current Affiliations in Listed boards 0.46 0.39 0.45 0.35 0.01 0.54

Dummy: Foreign Board Member 0.30 0.46 0.36 0.48 0.06 -3.03

Time Until Retirement 1.05 1.08 1.73 0.96 -0.68*** -10.22

In this table I report the attributes of women and male directors in my sample. Column (5) and (6) reports results of the difference in mean t-test.

Consequences of the Dutch gender quota on board composition and firm performance