The announcement effects of adding a female to the board as CEO on shareholder value in Europe

The announcement effects of adding a female to

the board as CEO on shareholder value in Europe

Name: Lisa Beekhuis Studentnumber: 10335854 Date: 24-06-2016 Bachelor thesis: Economics and Business Specialization: Finance and Organization Supervisor: Guilherme Vala Elias Pimentel de Oliveira

ABSTRACT

This thesis examines the announcement effects of appointing a woman as CEO, for 154 European companies listed on the Euronext Amsterdam, Euronext Brussels, Euronext Lisbon and Euronext Paris. Using an event study, the cumulative abnormal returns are estimated for the event window [-1,+1], for the time period 2001-2016. This research finds that the announcement of the appointment of a new CEO affects the cumulative average abnormal returns with -0.4119%; this is not significant. Several regression analyses have been executed on the effects of the announcement of the appointment of a female CEO. The research finds no significant effects of the announcement of the appointment of a female CEO on shareholders value. Keywords: gender; CEO; event study; announcement STATEMENT OF ORIGINALITY This document is written by Lisa Beekhuis who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

TABLE OF CONTENTS

1. INTRODUCTION ... 4 2. LITERATURE REVIEW AND HYPOTHESES ... 5 2.1 CORPORATE GOVERNANCE ... 5 2.2 GENDER DIVERSITY ... 6 2.3 HYPOTHESES ... 9 3. METHODOLOGY ... 9 3.1 DATA ... 9 3.2 EVENT STUDY ... 10 3.3 REGRESSION ... 12 4. RESULTS ... 13 4.1 EVENT STUDY RESULTS ... 13 4.2 REGRESSION RESULTS ... 14 5. CONCLUSION ... 18 REFERENCES ... 19 APPENDICES ... 22 APPENDIX A ... 22 APPENDIX B ... 23

1. Introduction

In 2015 15% of all directorships were held by women (covering 4218 companies around the world), which shows that women are in minority in corporate boards (MSCI, 2015). This percentage is astonishing if results about performance are shown: research shows that companies perform better when they have female leadership. Generated Return on Equity equals 10.1% for companies with strong female leadership against 7.4% for those without (MSCI, 2015).

In 2010 the European Commission started to give some pressure to companies to have more females at corporate boards, because research showed that in 2010 only 11.9% of board seats in the largest publicly listed companies in the EU were held by women (European Commission, 2015). In 2015 an increase was shown to 21,2% female board members, but this is still a low percentage, taking into account that 60% of university graduates are female (European Commission, 2015).

The minority of women in corporate boards could be a consequence of the negative effect that appointment of a woman has on shareholder value, which can be affected by a lot of factors. Shareholders want managers to maximize their value, while managers may have other preferences (also known as Agency theory (Jensen and Meckling, 1976)). Shareholder reaction on the announcement of a female CEO appointment has been investigated before with data from the United States, for instance by Lee and James (2006), who did research on the announcement effects of top executives. They found that the announcement of a female executive results in more negative stock market reaction than the announcement of a male executive.

From the research of Lee and James (2006) the following research question is derived: What are the announcement effects of adding a woman to the board of directors

as a Chief Executive Officer on shareholder value in Europe? Due to existing literature I

expect the stock market returns to be negative when a female is appointed as CEO. In line with the expectations, the Cumulated Abnormal Returns (CAR) in this research are negative, but insignificant. Therefore there is no evidence of any negative shareholder reaction after an announcement of an appointment of a female CEO.

The sample of this empirical research consists of European companies included in the Euronext stock market. I use data of announcements of appointments of CEO’s between 2001 and 2016, while these CEO’s are still in function at research date.

The remainder of this thesis is organized in the following manner. In the second section I discuss some existing literature and research on corporate governance and gender diversity. Section three provides the data and the methodology of the event study and the regressions, to give the results of the empirical research in section four. After the results I give a conclusion and advice for future research, and I add some appendices to this study for more understanding.

2. Literature Review and Hypotheses

This section provides information of existing literature on corporate governance and gender diversity. Literature shows that corporate governance is a key determinant for firm performance, and that corporate governance can differ among countries. An important difference in corporate governance between Europe and the United States is that in the United States companies have one-tier boards while in Europe two-tier boards are more common. Furthermore, literature shows ambiguous results regarding CEO announcements, both positive and negative shareholder reactions are found. After the literature review the hypotheses are mentioned. 2.1 Corporate Governance Research on announcement effects has been done before in the United States, but results for Europe could differ due to differences in corporate governance. Corporate governance can be defined as the protection of suppliers of capital and shareholders by using a set of control mechanisms and institutions (Kirchmaier et al., 2005). Research shows that the structure of a corporate board has a lot of influence on how companies perform and managerial characteristics have significant effects (Bertrand an Schoar, 2003; Arena and Braga-Alves, 2013).

According to Kirchmaier et al. (2005), a major difference between corporate governance in the United States and Europe is that shareholders in the United States have stronger rights from law than Europeans have - which results in more protection for minority shareholders - while in Europe it is common that there are shareholders with a majority of the shares (dominant shareholders), so that they have a lot of input in the company.

There are two important corporate governance models, the one-tier board and the two-tier board. In a one-tier board the Chief Executive Officer can be Chairman too (this is called CEO-duality). In a two-tier board these functions are separated. According to Maassen (1999), companies in the United States have one-tier boards, and Towers Watson (2012) argued that for example in the Netherlands one-tier boards were not legally allowed until 2012, so nowadays most companies still have two-tier boards. In a two-tier corporate governance model CEO’s are more supervised compared to a one-tier board, while in a one-tier board the relationship between members is closer and information flows more easy (Towers Watson, 2012).

Hermalin and Weisbach (1998) did research on corporate boards using a bargaining game system between the CEO and the rest of the board. They examine whether board effectiveness is related to the appointment process, because previous research did not find consistent stock market reaction to management changes. Hermalin and Weisbach (1998) argue that the reaction of the stock market depends on the type of CEO change. In their research the authors find a negative relation between CEO turnover and firm performance, and the more independent the board is, the stronger this relationship.

Besides literature about corporate governance in general, also research has been done to CEO appointment and CEO experience. Research shows that job performance is strongly related with job-specific experience, according to the human capital theory (Elais et al., 2011). Baysinger and Butler (1985) find that if women have no experience, appointing them has no positive effect on firm performance. Elais et al. (2011) find, using an event study, that if a firm has bad financial performance, stock market reactions are positive if the firm hires a CEO with some previous experience. 2.2 Gender diversity By nature, men and women are different. Bowman-Upton and Sexton (1990) argue that women have a significantly lower energy level and are less likely to take risk, relative to men. But on the other hand, they show that women have a higher need for independence. Beside the differences, Bowman-Upton and Sexton (1990) also show some similarities in business characteristics between men and women. Women and men have similar social competences and both men and woman want to avoid problems.

Nowadays, women are still in minority in boardrooms. Adams and Ferreira (2004) show that changing gender composition has several benefits, e.g. the different perspective women bring into the board room. According to Kim and Starks (2016) women can bring another type of expertise into the board due to their unique skills, and therefore the board will be more effective in their decision making. The conclusion they give is that a more gender-diverse board can increase firm performance.

Furthermore, Campbell and Minguez-Vera (2008) also find a positive relation between gender diversity in boardrooms and firm performance, measured by an approximation of Tobin’s Q, using panel data from Spain. Their research suggests that arguments for greater presence of female can be split up into ethical and economic arguments. An example of an ethical argument is the fact that in principle men and women are equal, so also women deserve a chance to participate in corporate functions. An economic argument the authors give is that women can lead to more innovative and creative solutions and gender-divers boards are better in finding alternative solutions to problems.

Rossi and Cebula (2015) did research on the announcement of board of director appointments. In their study, Rossi and Cebula (2015) use a sample of 100 announcements during the time period 2012-2014, for Italian listed companies. They find positive significant cumulative abnormal returns for four of the examined event windows. Rossi and Cebula (2015) argue that the results can be linked mostly to the composition of the board of directors.

Martin et al. (2009) did research on the announcement of a female CEO appointment. They use 70 announcements of female CEO appointments and 70 announcements of male CEO appointments to examine whether there is a difference in the stock market reaction. They find positive significant abnormal returns after the announcement of a female CEO appointment, and they find no significant differences between male and female CEO appointments. For that reason Martin et al. (2009) conclude that shareholders are as confident in men as in women.

Also Farrell and Hersch (2005) did research on the announcement effects of additions of corporate boards for women compared to men, by performing an event study. They use a sample consisting of firms listed on the Fortune 500 and Service 500 in 1990. Farrell and Hersch (2005) conclude in their research that while women serve

on better performing firms very often; the abnormal returns around the announcement dates are not significant.

Besides the researches that show firm performance is positively related to gender diversity, also negative relations are found. Adams and Fereirra (2009) investigate the effect of gender diversity on firm performance for firms listed on S&P500, S&P MidCaps, and S&P SmallCap, using panel data. The authors find that the presence of a more gender diverse board results in higher CEO turnover sensitivity to stock market reactions, relative to boards without diversity in gender. Beside this, they also find that attendance records of women are better and therefore women are more likely to monitor, but too much monitoring can reduce shareholder value. Adams and Fereirra (2009) conclude that overall, a greater gender diversity worse firm performance.

Furthermore, Lee and James (2006) show significant negative stock market reaction for female appointments, in contrast to the results of Martin et al. (2009). An explanation Martin et al. (2009) give for those different results is that Martin et al. (2009) use more recent data and have a larger proportion of women in their dataset. Lee and James (2006) did research using a sample of top executive female announcements in the United States between 1990 and 2000. They perform an event study using cumulative abnormal returns. An explanation given by Lee and James (2006) for their results is that due to the minority of women in top executive functions, women are still seen as anomalies. Lee and James (2006) state that if women take more positions in top executive jobs, the difference in shareholder reaction on male and female will become smaller.

To increase the number of women in executive boards, several European countries enforced a women quota, both legislated and voluntary. Adams and Ferreira (2004) provide evidence that additional costs may arise when changing the gender composition in corporate boards, which can be an argument for the slow increase of women in corporate boards. They argue that women quotas can result in lower shareholder value. According to Ahern and Dittmar (2012) Norway was the first country that set a women quota for the board of directors of 40% for all public-limited firms, effective from July 2005. In their research Ahern and Dittmar (2012) use panel data on 248 publicly listed companies and show that the announcement of this new law in Norway resulted in a negative stock market reaction for all firms, but firms with no

female directors react more negative compared to firms with at least one female director. The reaction of publicly listed firms was to change their corporate structure, such that the law does not affect them (Fortune, 2014). 2.3 Hypotheses According to the literature, the following hypothesis are adopted: H1: Announcement of adding a female to the board as CEO results in negative stock market reactions. H2: Announcement of adding a female to the board as CEO results in negative stock market reaction when she is a shareholder too. H3: Announcement of adding a female to the board as CEO results in less negative stock market reaction when she has only a few other executive positions.

3. Methodology

This section will provide information on the methodology of this empirical research. Using the market model, also known as the Capital Asset Pricing Model (now: CAPM), I explain how to perform an event study to get (cumulative) abnormal returns. Furthermore, I show and explain the regression model and the variables included. 3.1 Data The sample of this empirical research consists of companies listed on the Euronext stock market, which made an announcement for appointing a new CEO in the period 2001 – 2016. For every company included in the sample, company- and individual information is obtained from the Orbis database. Since I need information on shareholders reaction, financial information (i.e. returns) is obtained from DataStream. After cleaning the data, the dataset consists of 154 observations1_. 1 _{In the dataset 1294 companies were included. After cleaning the data for missing- and wrong} information the dataset is reduced to 362 observations. Announcement dates are obtained from press releases from the companies. Since not all companies provide their full archives, some CEO’s were founder too and some CEO’s were in their function before the IPO, the dataset reduced to 167 observations. For 14 of the 167 companies returns were not available.

3.2 Event Study

To measure shareholder reaction, I perform an event study. According to MacKinley (1997), an event study is used to measure the effects on firm value of an economic event, such as the announcement of a new CEO. A reason for this, given by MacKinley (1997), is that important events have an immediate effect on security prices. A major part of an event study is to define the period of interest that will be examined: the event window. The event window includes several days before the event, the event day itself and several days after the event.

Different studies have different event windows. Lee and James (2006) and Farrel and Hersch (2005) both used an event window of [-1,+1], Ahern and Dittmar (2012) used an event study of five days surrounding the event date, so [-2,+2], and Rossi and Cebula (2015) used six different event windows: [-10,0], [0,+5], [-5,+5], [0,+10], [-10,+10] and [-1,+1]. Since all of them investigate stock market reactions on changes or additions in corporate boards, this research focuses on the event window most of them use: [-1,+1], and additionally I investigate the event windows [-3,+3] and [-5,+5] to see if there are differences between the different event windows. To perform an event study, abnormal returns are needed. The abnormal return can be defined as the actual return minus the expected or normal return, i.e. the realistic return in case of no event (MacKinley, 1997; de Jong, 2007):

𝐴𝑅

_!,!

= 𝑅

_!,!

− 𝑅

_!,!

The normal returns have to be estimated using an estimation window, which differs for every research, just as the event window. The estimation window should not overlap the event window, because then the event is taken into account (MacKinley, 1997). Rossi and Cebula (2015) used an estimation window of [-250, -11], Farrel and Hersch (2005) used an estimation window of [-150, -20], and MacKinley (1997) uses different estimation windows: [-120, “day before start event window”] and [-250, -20]. After balancing the estimation windows for above-mentioned researches, this research will use the following estimation window: [-200, -20].

-200 -20 -1 0 +1 To estimate the normal returns, this research uses the market model, since MacKinley (1997) and Farrel and Hersch (2005) used it in equal research too and is proved to be a well-performing model. The market model, also known as CAPM, described by MacKinley (1997) is a statistical model that relates the return of one particular company to the market return, assuming there is a stable linear relationship. I use the following formula to estimate the normal returns (MacKinley, 1997; de Jong, 2007):

𝑅

_!,!

= 𝛼

_!

+ 𝛽

_!

𝑅

_!"#,!

+ 𝜀

_!,!

In this formula, Ri,t and Rmkt,t are the returns at time t for company i and the market, respectively, and 𝜀_!,! is an error term. The expected error term is assumed to be zero and its variance is assumed to be constant.

To draw overall inferences for the event window, the abnormal returns have to be aggregated through time and across securities (de Jong, 2007; MacKinley, 1997). I aggregate the abnormal returns across securities to the unweighted average of abnormal returns using the following formula (MacKinley, 1997; de Jong, 2007):

𝐴𝐴𝑅

_!

=

1

𝑁

𝐴𝑅

!,! ! !!!

Figure 1: Period of interest Estimation window [-200,-20], the event window [-1,+1] and the event date, 0.

To calculate the aggregation through time I use the following formula (MacKinley, 1997; de Jong, 2007):

𝐶𝐴𝐴𝑅 𝑡

_!

, 𝑡

_!

=

𝐴𝐴𝑅

_! !_! !! !_!

3.3 Regression

To run a regression, there have to be dependent and independent variables. The dependent variable in this research is for every regression the same: the cumulative abnormal return (from now: CAR). The CARs for an event window of [-1,+1] can be calculated by the following formula (MacKinley, 1997; de Jong, 2007):

𝐶𝐴𝑅(−1, +1) =

𝐴𝑅

_!,! ! !!!!

For the first hypothesis, the regression equitation equals: 𝐶𝐴𝑅_!,! = 𝛼 + 𝛽_!𝐹𝐸𝑀𝐴𝐿𝐸 + 𝛽_!𝐴𝐺𝐸 + 𝛽_!𝑆𝐻𝐴𝑅𝐸𝐻𝑂𝐿𝐷𝐸𝑅 + 𝛽_!𝐻𝐼𝐺𝐻_𝑃𝑂𝑆 + 𝛽_!𝐿𝑂𝑊_𝑃𝑂𝑆 + 𝛽_!𝐹𝐼𝑅𝑀𝑆𝐼𝑍𝐸 + 𝛽_!𝐶𝑅𝐼𝑆𝐼𝑆 + 𝜀_!

In this regression the dependent variable 𝐶𝐴𝑅_!,! denotes the cumulative abnormal return, 𝛼 is the constant in the model and 𝜀_! is the error term. The beta’s are multiplied by independent and control variables.

The independent variables in this research are the dummy variables female,

shareholder and low other executive positions. The variable female denotes 1 if the CEO is

a female and 0 if the CEO is a male. The variable shareholder denotes 1 if the CEO is shareholder of the company too and 0 if not. The variable low other executive positions denote 1 if the CEO has less than ten other executive positions, and 0 if more than ten.

In the regressions that test the second and third hypothesis, the variables

shareholder and low other executive positions are both multiplied with the variable female. The intersection variable female multiplied by shareholder denotes 1 if the CEO is

female and shareholder, and 0 otherwise. The intersection variable female multiplied with low other executive positions denotes 1 if the CEO is a female and she has less than ten other executive positions, and 0 otherwise. The intersection variables are the explanatory variables for the second and third hypothesis, respectively.

The control variables in this research are the variables age, high other executive

positions, firm size and crisis. Age is a dummy variable that denotes 1 if the CEO is older

than the average age, and 0 if the CEO is younger than the average age. High other

executive positions is a dummy variable that denotes 1 if the CEO has equal or more than

twenty other executive positions, and 0 if less than twenty. Firm size is the natural logarithm of the firms’ total assets. Finally, crisis is a dummy variable which denotes 1 if the announcement is made during the European crisis, and 0 if not.

4. Results

This section discusses the results of the empirical study. The event study shows no significant cumulative average abnormal returns around the announcement date of the appointment of a new CEO. The regression analyses show that the announcement of the appointment of a female CEO does not significantly affects the cumulative abnormal returns.

4.1 Event Study results

In this section I discuss the results of the event study. The methodology of the event study is discussed in section 3. For the event study I did a t-test on the Cumulative Average Abnormal Returns (CAARs) to see whether they are significantly different from zero. As Table 1 shows, the CAAR for the event window [-1,+1] is -0.4119% and the CAAR for the event window [-5,+5] is -0.30339%, but both not significant, which means that no significant abnormal returns are found after the announcement of the appointment of a new CEO for the used sample. A possible explanation for these results is that shareholders tend to react negative on changes in corporate governance, but still have enough believe in the appointment process and selection, resulting in no significant effects.

The results from this research can be related to the results existing literature shows. Lee and James (2007) found significant negative cumulative abnormal returns

for the event window [-1,+1] of -2.47% for the appointment of a female CEO, and Farrel and Hersch (2005) found insignificant negative cumulative abnormal returns for the time window [-1,+1] of – 0.0981% for adding a female to an all male board.

Table 1 – Cumulative Abnormal Returns

Event date AAR t-value

-5 0.00446% 0.031155 -4 0.03283% 0.207497 -3 -0.23123% -0.80138 -2 0.02187% 1.23341 -1 -0.06844% -0.34204 0 -0.24092% -0.61907 1 -0.10253% -0.32889 2 -0.06933% -0.32914 3 0.08191% 0.442496 4 0.15261% 0.853021 5 -0.08144% -0.4396 CAAR[-1,+1] -0.4119% -0,097206 CAAR[-5,+5] -0.30339% -0.13701 This table gives the results of the event study for the event windows [-1,+1] and [-5,+5]. For this study the whole sample is used. Note: Significance levels at 1%, 5% and 10% presented by ***, ** and * respectively. 4.2 Regression results In this section I discuss the results of the regressions. Before I analyse the regressions I give some descriptive statistics of the variables of the basic regression in Table 2. In the sample 23.4% of the appointed CEOs is female, which is corresponding with the percentage of females in board seats in Europe. Beside that, only 8.44% of the CEO’s in the sample is also a shareholder of the company. Furthermore, an interesting statistic is that the largest number of other executive positions is 59, while the mean of other executive positions is 9.66.

Table 2 – Descriptive Statistics of the variables of the Basic Regression

Obs Mean Std. Dev. Min Max

CAR 154 -0.00422 0.07341 -0.53145 0.29525 Female 154 0.23377 0.42461 0 1 Age 154 53.1039 6.60082 33 71 Shareholder 154 0.08442 0.27892 0 1 Other Positions 154 9.66013 10.7869 1 59 Ln_Assets 154 14.3956 3.01959 3.332205 20.7161 Year 154 2011.461 3.26425 2001 2016

This table gives the descriptive statistics that are used in the basic regression. It shows the number of observations, the mean, the standard deviation, the minimum value and the maximum value. In the table above, the CAR is de cumulative abnormal return of the event window [-1,+1], age gives the age of the CEO at the announcement date, other positions gives the number of other executive positions, ln_assets is the natural logarithm of the total assets (in euro’s) and female and shareholder are dummy variables that can equal 0 or 1. This research tests three different hypotheses, and for these hypotheses four different regressions are tested. For all regressions the CAR [-1,+1] is the dependent variable. In model 1, as Table 3 shows, there are no significant variables. I did this regression to test the announcement effects for gender only, without any control variables. For the first hypothesis, model 2 is used. The variable of interest in this model is “Female”, and as Table 2 shows, this variable is not significantly different from zero, so for that reason the hypothesis that stock market returns are negative when an announcement is made for appointing a woman as CEO is rejected. The significant variables in model 2 are the natural logarithm of the assets, which controls for firm size, and the variable that addresses whether the announcement is made during the European crisis. As we conclude from this model shareholder reaction is negatively affected by firm size, but is positively affected when the announcement is made during crisis. Even while the variable coefficients are not significant, similarities are found in the literature. Lee and James (2007) found significant negative stock market reaction for adding a female to the board of directors, from a sample between 1990 and 2000. Martin et al. (2009) found in their research a significant positive stock market reaction, using

more recent data than Lee and James did. A possible explaining for the results in this research is that business cultures are changing, and women are breaking the glass ceiling, and right now shareholder reaction is changing regarding women.

For the second hypothesis, model 3 is used. The variable of interest in this model is the intersection variable in which ”Female” is multiplied with the variable “Shareholder”. As Table 2 shows, the variable is negative, as expected, but also this variable is not significant, so that the second hypothesis has to be rejected too. In this research there is no prove that adding a woman to the board as CEO when she is a shareholder too, results in negative stock market reactions. These results could be biased, because very few CEO’s in the sample are shareholder too. Although this coefficient is insignificant, it seems to have a tendency to be negative, as expected due to agency theory, i.e. other preferences for managers and shareholders.

For the third hypothesis, model 4 is used. The variable of interest in this model is the intersection variable in which the variable “Female” is multiplied with the variable that counts for CEOs with less than ten other executive positions. Also this variable is negative, but again, insignificant, such that also the third hypothesis is rejected. Literature at this part explains that a lot of experience can have a negative effect on CEO success, because job-specific experience can make a CEO less risk-loved and rigid. An explanation for this is that every firm is different and all events have to be approached differently.

A possible explanation for the turbid results could be the difference in corporate governance in Europe compared to the United States. Since most of the European companies have a two-tier board, shareholders may trust on the effectiveness of the supervisory board.

Also regressions have been done on CARs for the event windows [-3,+3] and [-5,+5], for the basic regression. For the event window of [–3,+3] the coefficient for the variable female is -0.0172, which is not significant. For the event window of [-5,+5] the coefficient for the variable female is -0.0116, which is also not significant. Results of these regressions are tabulated in Appendix A and Appendix B.

Table 3 – Ordinary Least Squares Regression (1) (2) (3) (4) Constant -0.0036 (0.0068) 0.0869 (0.0555) 0.0834 (0.0556) 0.0894 (0.0557) Female -0.0024 (0.0140) -0.0062 (0.0144) -0.0023 (0.0149) 0.0103 (0.027) Age -0.0001 (0.001) -0.0001 (0.0009) -0.0002 (0.001) Shareholder 0.0281 (0.0218) 0.0373 (0.0233) 0.0274 (0.0218) High_other_pos -0.0321 (0.0289) -0.0272 (0.0292) -0.0328 (0.029) Low_other_pos -0.00004 (0.0131) 0.0003 (0.0139) 0.0046 (0.0153) Ln_assets -0.0062** (0.0020) -0.006*** (0.0020) -0.0064*** (0.0020) Year_crisis 0.0205* (0.0131) 0.0199 (0.0131) 0.0207 (0.0131) Female*shareholder -0.0625 (0.0574) Female*lowotherpos -0.0229 (0.0315) Observations 154 154 154 154 R-squared 0.0002 0.1020 0.1093 0.1053 This table shows the results of the regressions tested for the hypotheses mentioned in section 2.5. Dependent variable: CAR[-1,+1] Significance levels at 1%, 5% and 10% presented by ***, ** and * respectively. Standard errors are between brackets

5. Conclusion

This thesis examines the announcement effects of the appointment of a female as CEO. This examination is done by doing an event study on announcement of CEO appointments for European companies listed on the Euronext, in a time period of 2001-2016. For this time period a cumulative average abnormal return of -0.4119% is found for the event window [-1,+1] and a cumulative average abnormal return of -0.30339% is found for the event window [-5,+5], both are not significantly different from zero. Even while they are not significant, they were expected to be negative due to existing literature.

For testing the hypotheses four regressions are performed. The second model predicts whether the CEO being a female has a negative effect on shareholder value. The coefficient found equals to -0.0062. This is insignificant, and for that reason the first hypothesis is rejected. The third model predicts stock market reaction when the appointed female is also a shareholder. The coefficient found equals to -0.0625. This is insignificant, and for that reason also the second hypothesis is rejected. The fourth model predicts stock market reaction when the appointed female has only a few other executive positions, to test the third hypothesis. The coefficient found equals to -0.0229. This is again insignificant, and therefore, also the third hypothesis is rejected.

Main conclusion from the event study is that there is no significant effect in shareholder value around the announcement date of the appointment of a new CEO in general for companies in Europe. Main conclusion from the regression analysis is that there is no significant effect in shareholder value around the announcement date of the appointment of a woman as CEO in Europe. Differences in results for Europe versus the United States can be a consequence of differences in corporate governance. For further research, it would be interesting to do research on the announcement effects of female appointments again for Europe, but then with some more information on the announcement. For example information is needed about whether the announcement was expected or not, if the appointed CEO was an in- or outsider, in which stage the company is and about the percentage of shares a CEO holds in the company.

References

Adams, R., & Ferreira, D. (2004). Gender Diversity in the Boardroom. Finance Working

paper 57, 1-22.

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

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(1), 137-197.

Arena, M.P., & Braga-Alves, M.V. (2013). The discretionary effect of CEOs and board chairs on corporate governance structures. Journal of Empirical Finance, 21, 121-131. Baysinger, B.D., & Butler, H.N. (1985). Corporate Governance and the Board of Directors: Performance Effects of Changes in Board Composition. Journal of Law, Economics, & Organization. 1(1), 101-124. Bertrand, M., & Schoar, A. (2003). Managing with style: The effect of managers on firm policies. The quarterly journal of economics, 118(4), 1169-1208. Campbell, K., & Minguez-Vera, A. (2008). Gender Diversity in the Boardroom and Firm Financial Performance. Journal of Business Ethics, 83, 435-451. Elsaid, E., Wang, X., Davidson III, W.N. (2011). Does experience matter? CEO successions by former CEOs. Managerial Finance, 37(10), 915-939.

European Commission. (2015). Gender balance on corporate boards: Europe is cracking

the glass ceiling. Retrieved from: http://ec.europa.eu/justice/gender

equality/files/womenonboards/factsheet_women_on_boards_web_2015 10_en.pdf

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

Fortune. (2014). Women on boards: Are quotas really the answer? Retrieved from: http://fortune.com/2014/12/05/women-on-boards-quotas/ Hermalin, B.E., & Weisbach, M.S. (1998). Endogenously Chosen Boards of Directors and Their Monitoring of the CEO. The American Economic Review, 88(1), 96-118. Jensen, M.C., & Meckling, W.H. (1976). Theory of the firm: Managerial Behavior, Agency Costs and Ownership structure. Journal of Financial economics, 3, 305-360.

Jong, de F. (2007). Event Studies Methodology. Retrieved from: BlackBoard, Bachelor’s Thesis and Thesis Seminar Finance and Organization

Kim, D., & Starks, L.T. (2016). Gender Diversity on Corporate Boards: Do Women Contribute Unique Skills? American Economic Review, 106(5), 267-271

Kirchmaier, T., Owen, G., Grant, J. (2005). Corporate Governance in the US and Europe:

Where Are We Now? Retrieved from: http://ssrn.com/abstract=708461 Lee, P.M., & James, E.H. (2007). SHE’-E-OS: Gender effects and investor reactions to the announcement of top executive appointments. Strategic Management Journal, 28, 227-241. Maassen, G.F., (1999). An international comparison of corporate governance models: a study on the formal independence and convergence of one-tier and two-tier corporate boards of directors in the United States of America, the United Kingdom and the Netherlands. ISBN: 9090125914

MacKinlay, A.C. (1997). Event Studies in Economics and Finance. Journal of Economic

Martin, A.D., Nishikawa, T., Williams, M.A. (2009). CEO Gender: Effects on Valuation and Risk. Quarterly Journal of Finance and Accounting, 48(3), 23-40

MSCI. (2015). Women on boards. Retrieved from

https://www.msci.com/documents/10199/04b6f646-d638-4878-9c61 4eb91748a82b Rossi, F., & Cebula, R.J. (2015). Stock market reactions to announcements of the board of directors: evidence from Italy. Applied Economics, 47(20), 2102-2118. Sexton, D., & Bowman-Upton, N. (1990). Female and male entrepreneurs: Psychological characteristics and their role in gender-related discrimination. Journal of Business Venturing, 5(1), 29-36.

Towers Watson. (2012). European Corporate Governance: A Lack of Shared Identity

versus a Move towards Convergence. Retrieved from: https://www.towerswatson.com/nl-NL/Insights/IC-Types/

Ad-hoc-Point-of-View/2012/European-Corporate-Governance

Appendices

Appendix A Ordinary Least Squares regressions with dependent variable CAR[-3,+3] Table 4 – Ordinary Least Squares Regression Constant 0.1267 (0.0852) Female -0.0172 (0.0221) Age -0.0001 (0.0015) Shareholder 0.017 (0.0334) High_other_pos -0.0386 (0.0443) Low_other_pos -0.0052 (0.0213) Ln_assets -0.0082*** (0.0031) Year_crisis 0.0035 (0.0201) Observations 154 R-squared 0.0581 Dependent variable: CAR[3,+3] Significance levels at 1%, 5% and 10% presented by ***, ** and * respectively. Standard errors are between brackets

Appendix B Ordinary Least Squares regressions with dependent variable CAR[-5,+5] Table 5 – Ordinary Least Squares Regression Constant 0.1552 (0.0981) Female -0.0116 (0.0255) Age -0.0006 (0.0017) Shareholder 0.0076 (0.0384) High_other_pos -0.0289 (0.0511) Low_other_pos -0.0072 (0.0246) Ln_assets -0.0082** (0.0036) Year_crisis 0.0036 (0.0231) Observations 154 R-squared 0.0581 Dependent variable: CAR[-3,+3] Significance levels at 1%, 5% and 10% presented by ***, ** and * respectively. Standard errors are between brackets