MERGERS AND ACQUISITIONS: THE ROLE OF GENDER IN EUROPE AND THE UNITED KINGDOM

) MERGERS AND ACQUISITIONS: THE ROLE OF GENDER IN EUROPE AND THE UNITED KINGDOM Ersin Güner 559370 Master Finance Supervisor: dr. P.C. (Peter) de Goeij December 2013

Abstract Evidence from the US shows that women on corporate boards affect the pricing of mergers and acquisitions. Prior to this paper, there has not been a study that investigated how gender plays a role in mergers and acquisitions in Europe or the United Kingdom. I show that the bidder cumulative abnormal announcement period returns increases by more than 20 percent for each 1 percent of female directors on the European bidder board. Second, for each 1 percent representation of independent women on the Eurozone target board the bid premium increases by 25 percent. Finally, I find that for each 1 percent representation of independent women on the UK target board decreases the bidder cumulative abnormal announcement period returns by more than 30 percent. These effects are robust to the use of larger samples and different measures of market price reaction. Whether a female CEO has an effect on the pricing of mergers and acquisitions in Europe or the UK, remains inconclusive. 2

1. Introduction Earlier studies show that women affect firm financial performance, as measured by Return on Assets, Return on Sales, Tobin s Q, and other measures. The results show that this gender diversity on board of companies has a positive influence on firm performance. These findings are not limited only to the US, but are also found in Europe. Different from these measures, there is evidence from the US that women on corporate boards affect the pricing of mergers and acquisitions (M&As) (Levi, Li, and Zhang, 2008). However, there has not been a prior study that investigates how gender plays a role in M&As in Europe or the United Kingdom. This gives this study an aim: I investigate whether the role of gender also affects mergers and acquisitions in Europe and the United Kingdom. By investigating this, I contribute with this paper to the literature how women affect firm financial performance. In this case, their impact on the pricing of M&As. Motivated by Levi, Li, and Zhang (2008), I look at how women influence M&As when serving as CEO of the bidder or target firm and via their degree of representation on boards of bidding and target companies as directors. In addition, I look whether the independence of a female director has an impact on the pricing. I use two samples of over 30 acquisition attempts. One sample covers the Eurozone for the period of 2006-2012, and the other sample covers the United Kingdom (UK) for the period of 2003-2012. For both samples it is inconclusive whether the gender of the CEO has an impact on the pricing of M&As. Further, the bidder three-day cumulative abnormal announcement period return (CAR3) increases by more than 20 percent for each 1 percent of female directors on the bidder board. This increasing effect is significant at the 10 percent level. I find more evidence when I distinguish between independent female directors and inside female directors. For each 1 percent representation of independent women on the Eurozone target board the bid premium increases by 25 percent. This effect is significant at the 10 percent level. In addition, I find that for each 1 percent representation of independent women on the UK target board decreases the bidder CAR3 by more than 30 percent. This dampening effect is significant at the 5 percent level. For robustness purposes I also employ two larger samples: For the Eurozone a sample over 60 deals and for the UK a sample over 70 deals. With these two samples I find similar results as I find with the main samples. 3

The remainder of the paper is organized as follows. The next section, I discuss the relevant literature on the role of gender in firm financial performance. I review the literature with findings on female CEOs and their effect on the performance measures. In addition, I discuss earlier findings how the presence of female directors on the board affect firm performance. Based upon the literature review I formulate and present my hypotheses in the second part of Section 2. Section 3 describes the samples I employ and the model specifications I use to investigate the role of gender in M&As. In Section 4, I present my main results. Moreover, I discuss the results and summarize them. Section 5 covers the robustness checks. In Section 6, I present the conclusion with recommendation for further work. 2.2 Literature Review & Hypothesis Development 2.1 Literature Review The current paper is related to literature that delves into the role of female directors in corporate decisions and firm performance, and the effect of board gender diversity on firm performance. One of the first studies that explored this subject is Levi, Li, and Zhang (2008). They find for US firms that the bid premium paid over the pre-announcement target share price is over 70 percent smaller when the CEO of the bidding company is a woman as opposed to when the CEO is a man. Furthermore, they showed that for each 10 percent representation of women on the board of the target results in a reduction of the bid premium by almost 15 percent. They also found that the three-day cumulative abnormal return approaches zero if the bidder CEO is a woman. Finally they showed that each 10 percent representation of woman on the target board reduces the target three-day cumulative abnormal return by almost 3 percent and the bidder three-day cumulative abnormal return by 1 percent. Marinova, Plantenga, and Remery (2010) investigated Dutch and Danish public companies by looking at the effect of board gender-diversity on firm performance measured by Tobin s Q. They show that there is a positive relation between firm performance and board genderdiversity, but this relationship is not statistically significant. Carter et al. (2010) document a similar relationship for US firms. However they find a positive and significant relation between representation of women on the board and return of assets. Continuing with mergers and acquisitions studies, Levi, Li, and Zhang (2011) find that women are less acquisitive than men. First, they look at the bids initiated by S&P 500 firms 4

during 1997-2009. They find that each 10 percent representation of women on the board results in a decrease of 7.5 percent in acquisition bids of companies. Second, with a subsample of 458 observations they also find that a 10 percent representation of women on the board of the bidder results in a reduction of the bid premium by 13.3 percent. Bugeja et al. (2012) conducted a similar study as Levi et al (2008). In this study the focus was again on US publicly listed companies. However, they investigated different deal types, over a different time period. Bugeja et al. examined acquisitions that were completed instead of attempts of acquisitions. They find that the board gender-diversity has no effect on the bid premium. Furthermore, they find that the diversity has no effect on the market reaction to the announcement of a merger or an acquisition. However, their results do show that the board gender-diversity has a positive impact on the long-term performance of the acquirers. They link this finding to the existing literature that the gender-diversity in the board may lead to economic outcomes that are beneficial to the firm. Faccio, Marchica, and Mura (2012) find that companies that have a female CEO have lower leverage, less varying earnings, and the probability of default is lower than companies that have a male CEO. These results are robust to various tests for endogeneity. Their results also suggest that female CEOs do not allocate capital efficiently. This appears so, because the results show that female CEOs do not allocate more capital to projects that create more value for well diversified shareholders. Parrotta and Smith (2013) find that companies led by female are significantly different compared to companies led by males. They show that having a female CEO has a variance reducing effect on economic outcome variables, such as investments, profits, return on equity and return on sales. Furthermore, they find that the gender of the chairman of the board of directors and the proportion of female directors on the board has relative to female CEO a smaller effect and when significant these two variables are negative. The authors argue that companies that are led by women have more stable results over time because women are less risk seeking than men. In support of this explanation, they refer to earlier findings that women value monitoring and stricter firm governance more than men. Hays, Day, and Belanus (2011) looked at the effect of CEO gender on the management and performance of credit unions. Concerning management they find that male CEOs seem to cluster with relatively larger credit unions. Female CEOs are more risk averse as the results show lower real estate loan concentrations and higher liquidity. In addition, male CEOs are more technology orientated and therefore more likely to implement web-based financial 5

services. Finally, female CEOs are striving more to increase the number of credit union members. Performance-wise the authors find there is not a statistically significant difference between the return on assets of unions led by male or female CEOs. They argue that this is due to the overall objective of credit unions as not-for-profit organizations. Davis et al. (2010) report for small and medium-sized service businesses (service SMEs) that the effect of CEO gender on performance is indirect. They find that service SMEs with a female CEO are more market oriented than service SMEs with a male CEO. Market orientation serves as a mediator between CEO gender and firm performance. Their results support that female-led service SMEs perform significantly better than male-led service SMEs due to their attention to market orientation. Mohan and Chen (2004) show that there is not a gender bias in the IPO market. They find no difference in underpricing after controlling for firm specific variables between IPOs that are led by a male CEO and IPOs that are led by a female CEO. Martin, Nishikawa, and Williams (2009) find that there is not a gender bias in the financial market measured by the three-day cumulative abnormal return around the event of an appointment of a male or female CEO. Their data consists of a matched sample of 70 female appointments and 70 male appointments. In addition, and in accordance with earlier findings of women being more risk averse, they find that changes in risk subsequent to CEO appointment are lower for female CEO. Moreover, they show that a company with relatively high total risk and idiosyncratic risk is more willing to appoint a female CEO in order to try to decrease the risk. On the other hand, Lee and James (2007) show that reactions to announcements of male CEOs are less negative than reactions to announcements of female CEOs. They show that the average threeday cumulative announcement period return (CAR3) of male led companies is less negative than the CAR3 of female led companies. In addition, they show that companies that appoint a female CEO from within the company also experience less negative CAR3 than companies that appoint female CEO from outside. Jalbert, Jalbert, and Furumo (2012) indicate that female CEOs are perceived differently by the financial markets and their management differs from their counterpart. They show that U.S. firms led female CEOs have higher sales growth, more institutional ownership, and provide higher return on investments (ROI) and on assets (ROA). Furthermore, these firms are valued higher in the market than firms led my male CEOs. The sample covers the time period 1997-2006 and female CEOs make up 1.22 percentage. 6

Triana, Miller, and Trzebiatowski (2013) find that the amount of power a female director has, is important when it comes to having an influence on the firm. They show that only when female directors on the board are powerful the relationship between board gender diversity and strategic change is the most significant positive and negative. They explain that this might be the case because power must be essential for women to have an influence. The authors also suggest that it might not make sense for having female directors on the board unless they have enough power to change. Cotter et al. (1997) show that target firms with a board that is occupied in majority by independent directors experience higher shareholder gains when facing an acquisition. They find that the target shareholders gain on average 20 percent points more than target firms that do not have a board with the bulk of the directors being independent. Moreover, the authors note that these gains come at the expense of the bidder shareholders. In addition, they find that when the board is independent the shareholder gains are greater in resisted offers as opposed to when the board is not independent. There have been studies that focused on differences between male en female directors instead of looking at the diversity of the board. Kulich et al. (2011) find that bonuses of male directors are higher than of female directors. What they also find is that the remuneration of the male directors depends more on the firm performance than the compensation of female directors. The magnitude of the gender wage gap in the UK is 17.0 percent (TUC Women s Conference, 2008). While the gap in continental Europe is also 17.0 percent, there is a huge variation per country. In the United States, women in management professions earn 28.4% less than men. In financial professions this gap is even wider (Bureau of Labor Statistics, 2013). One peculiar finding that develops in the same way is that corporate women earn less and less compared to men as they are promoted to a higher function. Put differently, the gap widens in top managerial positions, for example the position of CEO (Arulampalam, Booth, and Bryan, 2005). While there is a gap between what male directors and female directors get as reward, there is evidence, as I have reviewed above, that shows that the presence of women as CEO and/or director on the board has substantial and positive effects on firm performance and on stock 7

price movements. Moreover, the impact that female directors have on stock price movement in case of an M&A activity has not been examined with UK and Eurozone data. 2.2 Hypothesis Development As I have mentioned in the literature review, the role of gender in mergers and acquisitions has not been researched prior to the current paper in the Eurozone and the United Kingdom. That makes it an interesting topic to investigate. Moreover, there is evidence that suggests that female CEOs make better decisions than male CEOs. As a consequence, these decisions result in better firm financial performance. Therefore, it may be that female CEOs make better acquisition decisions than their male counterparts. In the current paper the performance of the CEO will be measured by the bid premium and the three-day cumulative abnormal return (CAR3). Based upon earlier findings I expect that the gender of the CEO will have an effect on the bid premium and on the CAR3 in mergers and acquisition activity. Therefore, I want to check whether there is a relation with the following: Hypothesis 1: The presence of a Female CEO has an effect on the Bid Premium and/or CAR3. I will check whether this is true by looking at the statistically significance of the coefficient of the dummy variable that indicates whether a CEO is female in equation (1) and (2). If the estimate is not significant I cannot accept hypothesis 1. For example, in the case that the bid premium is the dependent variable and the estimate of the dummy variable Target CEO is Female is significant and positive, I can say that the presence of a female CEO on the target board results in a higher bid premium. Earlier research also shows gender diversity (female directors on the board) results in positive outcomes for the shareholders. With the literature that I reviewed in the parts above, I expect that the presence of female directors on board of directors has an effect on the financial performance of a firm as measured by the bid premium and CAR3. Hypothesis 2: The presence of female directors on the board of directors has an effect on the bid premium and/or CAR3. 8

I will check this by looking at the statistically significance of the estimate of the variables Bidder/Target Proportion of Female Directors in equation (1). When the estimate of the variable is significant I can accept hypothesis 2. Finally, independent directors are more actively involved in decision making than inside directors (Cotter et al. (1997)). From this earlier finding I expect the following: Hypothesis 3: The effect of the presence of Independent female directors is larger than the effect of the presence of inside female directors on the bid premium and/or CAR3. I will compare the estimates of the variables Bidder (Target) Proportion of Independent Female Directors with Bidder (Target) Proportion of Inside Female Directors in equation (2) in absolute terms. For example, the following might be found: The estimate of the target independent female directors is positive and larger than the positive estimate of target inside female directors when the dependent variable is the bid premium. This would imply that the presence of independent female directors on the target board has a larger (and positive) effect on the bid premium than the presence of inside female directors on the target board. 3. Data and Research Framework I describe in this section how my samples are formed. Furthermore, I present the empirical specifications to explore the role of gender in mergers and acquisitions. 3.1 Samples Employed Before forming the two main M&A samples, I collect all Eurozone and United Kingdom (UK) acquisition attempts with announcement dates between January 1, 2003 and December 31, 2012 as identified by the Mergers and Acquisitions database of Thomson Financial s Securities Data Company (SDC) database. I choose the above time frame because the Eurozone countries have been using the Euro as currency for at least one year by January 1, 2003. The same time frame is chosen for UK because a comparison cannot be made otherwise. I identify all deals where both bidder and the target are public firms and where the form of the deal was coded as a merger (SDC deal form M), an acquisition of majority of 9

interest (AM), or an acquisition of assets (AA). After applying these filters I obtain 989 deals. The filters that are applied are motivated by a prior study by Levi, Li, and Zhang (2008). Next, I gather the market and share price returns from DataStream for the period 2003-2012. For accounting information I use the Worldscope database. The data source for board characteristics are the annual reports of companies. The major part of the annual reports is available via the site of the companies. The other part of the annual report is supplemented by the Morningstar annual reports database. Because there is not a specific database of board characteristics for European and UK companies, each annual report has been checked and data on each director has been collected manually. The director information allows me to compute the board size, independence of directors, the proportion of female directors, and the proportion of independent or inside female directors. My final sample consists of 72 acquisition attempts. Because there is a difference in the structure of the board of directors between UK companies and Eurozone companies I choose to split my sample of 72 in two samples. One sample consists of European acquirers and targets, and counts 32 acquisition attempts. The other sample consists of UK acquirers and targets, and counts 33 acquisitions attempts. The remaining 7 deals, which are acquisitions attempts between a UK company and a company from the Eurozone, will not be a part of this study. There is a huge decrease from the initial number of 989 deals to my final two samples that sum up to 65 deals. First reason is because the data retrieved from SDC contained data with missing values concerning the transaction value, the bid premium, and the Datastream company code. These missing values resulted in dropping from 989 deals to 460 deals. Second, the retrieved accounting data from Worldscope contained missing values. After dropping the missing values the total sample of deals downsized to 216 deals. Of these 216 deals I managed to find for 72 deals the annual reports for both the bidder company and of the target company. Finally, as I have explained above, I end up with two samples that together sum up to 65 deals. For a robustness check, I also employ two larger samples of acquisition attempts for the period 2003-2012 with data only of bidder board characteristics. I end up with 61 deals for the Eurozone and 71 deals for UK. 10

3.2 Sample Overview Table 1 Panel A presents the descriptive statistics on the acquisition attempts for the Eurozone. I show that on average the bidder pays a premium of 24.5 percent above the market price of the target, measured four weeks before the bid. The average three-day cumulative abnormal announcement period return (CAR3) for the target is 15.6 percent. The average CAR3 for the bidder is 1.5 percent. The average CAR7, CAR11, and CAR21 for the target are 17.6, 17.8, and 15.4 percent respectively. The average CAR7, CAR11, and CAR21 for the bidder are 1.4, 1.8, and 2.0 percent respectively. [Insert Table 1 Panel A about here] Table 1 Panel B presents the descriptive statistics on the acquisition attempts for the UK. I show that on average the bidders in UK pay a much higher premium than in the Eurozone of 63.6 percent above the market price of the UK target, measured four weeks before the bid. The average three-day cumulative abnormal announcement period return (CAR3) for the target is 17.1 percent. The average CAR3 for the bidder is 1.3 percent. The average CAR7, CAR11, and CAR21 for the target are 18.4, 20.3, and 18.6 percent respectively. The average CAR7, CAR11, and CAR21 for the bidder are -0.8, 1.5, and 1.0 percent respectively. These numbers are similar to the numbers of the Eurozone. [Insert Table 1 Panel B about here] I focus on the bid premium and cumulative abnormal announcement period return as measures of CEO and board member performance in the M&A process. These measures are motivated by Levi, Li, and Zhang (2008) to be valid and meaningful. They argue this by providing two examples. First, it would be a success to acquire a target for a smaller premium over the target share price prior to the bid. A smaller premium results in a smaller cumulative abnormal return for target shareholders around the announcement date. This positive outcome could be attributed to the competence of the bidder CEO and/or the board member. Second, a bigger premium also results in a bigger cumulative abnormal return around the announcement date. This could be attributed to the competence of the target CEO and/or board member. In this study the gender factor plays an important role. The Eurozone sample fraction of female bidder CEOs is 3.1 percent and the sample fraction of female target CEOs is 6.3 11

percent. The comparable numbers for the UK are as follows: 6.1 percent of the bidder CEOs is female, but in my sample there has not been a target firm with a female CEO. On the European boards of bidder companies the representation of women is 7.1 percent whilst the representation of women on the target companies is 5.8 percent. On the UK boards of the bidder women make up 5.2 percent and on the boards of the target women have a representation of 7.1 percent. When calculating these proportions the female CEO is excluded if that is the case. The average fraction of independent female directors of Eurozone bidder companies is 4.0 percent, whilst the fraction of inside female directors is 3.1 percent. For the targets these are 1.8 and 4.0 percent respectively. On average, the fraction of independent female directors for UK bidders is 3.9 percent and the fraction of inside female directors is 1.3 percent. The numbers for the targets are 4.4 and 2.7 percent respectively. The Eurozone bidder board size is 14.4 and the target board size is 10.7. The UK bidder board size is 7.7 and the target board size is 7.1. This is consistent with bidders being larger than targets, and that board size is positively correlated with firm size. Furthermore, on average 49.8 percent of the bidder boards in the Eurozone exist of independent directors and 43.5 percent of the target boards is made up of independent directors. In the UK the numbers are 50.0 percent and 48.1 percent respectively. In European bidder firms the CEO is also the COB (Chairman of the Board) in 28.1 percent of the cases. On average, the targets CEOs are also the COB in 18.8 percent of the cases. In the UK the numbers are smaller. Bidder CEOs are in 9.1 percent of the cases also the COB. Target CEOs are in 12.1 percent of the cases the COB. In the Eurozone, 46.9 percent of the deals are made with cash only as the method of payment. In the UK this number is 45.5 percent. Less than 16 percent of the European deals are pure stock swaps. In the UK this is almost twice as large: 30.3 percent. Less than half of the European deals are diversifying according to industry classifications of Fama and French (1997). Almost 40 percent of the deals in the UK are diversifying. Finally, I discuss the summary statistics on the bidders. On average, the Tobin s Q of Eurozone bidders and UK Bidder is 1.50 and 1.85 respectively. The mean book leverage ratio of Eurozone bidder is 20.2 percent and that of the UK bidders is 16.0 percent. On average, the operating cash flow to totals assets of European bidders is 9.0 percent. For UK bidders the 12

ratio is 10.9 percent. The Eurozone bidder has on average a market capitalization of 11 million and total assets of 20 million. The average UK bidder has a market capitalization of almost 2 million and total assets of 1.6 million. 3.3 Model Specification I run the following cross-sectional regressions to analyze the influence of gender in mergers and acquisitions: (1), where bid characteristics can be the bid premium, target CAR3, or bidder CAR3. The gender variables of CEOs and directors in the bidder and target firms are the main variables of interest in this study. The board-related variables and controls for bid-characteristics and firmcharacteristics are motivated by prior findings of the study of Levi, Li, and Zhang (2008). Characteristics that explain the bidder include Tobin s Q, book leverage, operating cash flow, and firm size. I report my results with three different specifications. I gather these results by estimating the model through employing equation (1). First, to provide a baseline result I run a regression excluding measures of CEO and director gender. Second, I run the regression as in equation (1) to estimate the model without year fixed effects. Finally, I run another regression with year fixed effects. By including year fixed effects I can control for unobserved heterogeneity per year and by doing so get results that are more accurate. Similar to Levi, Li, and Zhang (2008) I am also interested in the independence of female directors and the role of their independence. They argue that independent female directors make decisions that are probably better for their shareholders. These decisions that are in the 13

interest of shareholders may affect the bid premium and CARs. Therefore, I run the following cross-sectional regressions: (2) where Bid Characteristic could be the bid premium, target CAR3, and the bidder CAR3. The gender variables of CEOs and independent and inside directors of the bidder and target firms are the main variables of interest. I end this section with presenting the correlations between the dependent variables and all the right-hand side variables. I present the correlation matrix in Table 2, where Panel A represents the correlations of the variables for the Eurozone and Panel B represents the correlations for the UK. I show for the Eurozone deals that there is a positive correlation between the bid premium and the target CAR3. This is not unusual because the size of the bid premium is the main driving force behind the target price reaction. I also show that the target (bidder) CAR3 and CAR7 (or CAR11 or CAR21) are highly correlated, which is in line with expectations. Most correlation among pair of variables does not raise concern for multicollinearity in my regressions analysis. However, there are pairs that exhibit high correlation. The pairs are combinations between the bidder board size, size and bidder total assets. [Insert Table 2 Panel A about here] I present the same for the UK in Panel B of Table 2. The results are very similar to the Eurozone correlations. Concerns for multicollinearity arise again only through bidder board size, size, and bidder total assets. 14

[Insert Table 2 Panel B about here] 4. Gender Effects in Mergers and Acquisitions In this section I present my main findings on the role of gender in acquisitions attempts 4.1 Main Results [Insert Table 3 Panel A and Panel B about here] Table 3 Panel A presents the results of the regressions where the bid characteristic according to equation (1) is the bid premium and the main variables of interest are the gender variables of the bidder and target CEOs and directors for Eurozone companies. Table 3 Panel B presents the equivalent regression results for the UK companies. I establish a benchmark by excluding the main variables in the baseline regression (column (1)). I find that European bids using all-cash payment result in a higher bid premium. I present my main findings in column (3) of Panel A and Panel B. Column (3) of Panels A and B include year fixed effects. For UK bids I find no statistically significant estimates in column (3). For bids that took place in the Eurozone I find that the presence of independent directors on the target board is significant and negatively associated with the bid premium. The sample average of the bid premium is 24.5 percent and the negative coefficient is 1.038. This implies an economic effect that a 10 percent representation of independent directors on the target board would decrease the bid premium by more than 40 percent (= 0.1 1.038 / 0.245). This effect is significant at the 10 percent level. Furthermore, I can note that the inclusion of the main variables does not change the sign of the control variables. [Insert Table 3 Panel A about here] In Panel C and D of Table 3 I represent the regression results of equation (2) where the main variables of interest also include the independence of female directors. The main findings are again in column (3) that has the year fixed effects included. I find no statistically significant results for UK bids as can be seen in Panel D. I find for European bids that the presence of a 15

female bidder CEO is statistically significant and positively associated with the bid premium. In economic terms this presence would drive the bid premium up by almost 560 percent (= 1.386/0.245). This effect is significant at the 5 percent level, but it is important to note that in the entire sample there is just one female bidder CEO and therefore this result is not representative of the presence of a female CEO on the bidder board. [Insert Table 3 Panel C about here] In Panel C column (2), I find that the presence of independent female directors on the bidder board significantly reduces the bid premium. However, when I include the year fixed effects in column (3) to control for unobservable factors that are time-invariant the statistical significance of the coefficient disappears. The presence of inside female directors does not have a significant effect. Furthermore, I show that the presence of independent female directors on the target board significantly increases the bid premium. In economic terms this implies that for each 1 percent representation of independent women on the target board the bid premium increases by 25 percent (= 0.01 6.142 / 0.245). This effect is significant at the 10 percent level. Table 4 represents the regression results where the dependent variable is the target CAR3. In panel A and B the main variables of interest are the gender of the CEO and the directors. First I present a baseline regression in column (1), but neither the European (Panel A) nor UK (Panel B) results are significant. [Insert Table 4 Panel A and B about here] I present my main findings in column (3) of panel A and B. For the Eurozone bids (Panel A) I find no statistically significant results at all. For the UK bids (Panel B) I find no statistically significant coefficients on any of the main variables when I include year fixed effects. Without the year fixed effects, I find that the presence of female directors on the target board has a positive effect on the target CAR3 (column (2) of Panel B). The economic effect of each 10 percent representation of women on the target board results in an increase of almost 70 percent (= 0.1 1.184 / 0.171). This effect is significant at the 10 percent level. 16

In Table 4 I add a fourth column with the bid premium as an extra independent variable. This lets me explore the role of CEO and director gender indirectly via the bid premium. This is meaningful because the bid premium is a measure for CEO and directors competence like I stated in the previous section. In Panel A of Table 4 I find for European bids that the bid premium is significant and positively associated with the target CAR3. This effect is significant at the 5 percent level. In Panel C and D of Table 4 I present the results where there is a distinction between independent and inside female directors. Inside female directors do not seem to have an effect in the European bids. I show in Panel C column (3) that independent female directors on the target board have a significant and positive effect on the target CAR3. The economic effect on the target CAR3 is an increase of almost 28 percent (= 0.01 4.755 / 0.171) for each 1 percent representation of independent women on the board. This effect is significant at the 10 percent level. However, when I include the bid premium as an independent variable in column (4), this effect of independent female directors is not statistically significant anymore. [Insert Table 4 Panel C about here] In Table 4 Panel D I find for UK bids that inside female directors on the target board have a significant and positive effect on the target CAR3. Even when I include the bid premium, this effect remains positive and significant. In economic terms, for each 1 percent representation of inside female directors on the target board, the target CAR3 increases with 17 percent (= 0.01 2.922 / 0.171). This effect is significant at the 5 percent level. [Insert Table 4 Panel D about here] I present the regression results in Table 5 where the dependent variable is the bidder CAR3. In Panel A and B I focus just on the CEO and director gender. In panel C and D I add the control variables of independent and inside female directorship. I only include the variables that measure the general board characteristics and the bid and firm characteristics. I find in Panel A for European bids that the presence of independent directors on the bidder board has a significant negative association with the bidder CAR3. In Panel B I find for the UK bids that an all-equity payment method, a higher operating cash flow, and more total assets are 17

increasing the bidder CAR3. Furthermore I find that bidders with a bigger market capitalization and higher Tobin s Q face a smaller bidder CAR3. [Insert Table 5 Panel A and Panel B about here] Again the main results can be found in column (3) of Panel A and B. I show that female directors on the board of the European bidder results in a higher bidder CAR3 (Panel A). Even when I introduce the bid premium in column (4), the effect remains. In economic terms, the bidder CAR3 increases by more than 20 percent (= 0.01 0.322 / 0.015) for each 1 percent of female directors on the bidder board. This effect is significant at the 10 percent level. I find similar results in column (2) of Panel B for the UK bidder. However when I include the year fixed effects in column (3), the effect of female directors disappears. In Table 5 Panel C and Panel D I present the regressions results where I distinguish between independent female directors and inside female directors. I find no significant results in the Eurozone bids (Panel C). For the UK bids I find in column (2) of Panel D that there is a positive relation between the independence of female directors on the bidder board and bidder CAR3. However this is not statistically significant anymore when I include the year fixed effects. [Insert Table 5 Panel C and Panel D about here] I show in column (3) of Panel D that there is a negative association between independent female directors on the target board and the bidder CAR3. This effect remains even when I include the bid premium in column (4) as explanatory variable. The economic effect is that for each 1 percent representation of independent women on the target board decreases the bidder CAR3 by more than 30 percent (= 0.01 0.499 / 0.015). This effect is significant on the 5 percent level. 4.2 Discussion In this section I discuss the results of the previous section and their implications on hypothesis 1, 2 and 3. Although coefficients might be statistically significant it should be taken into consideration that the size of both samples is small. 18

Overall, I find no support for hypothesis 1. Overall, I find no statistically significant estimates for the effect that a female CEO has on the firm performance as measured by the bid premium and/or the CAR3. However, the significant estimate that I do find implies that a female CEO on the European bidder board has an increasing effect on the bid premium (Column (3) of Panel C, Table 3). However, in the entire sample there is just one female CEO. In addition the sample is limited to just 32 observations. Therefore, this result is weak to support hypothesis 1 and not representative for the effect that female CEO have on the bid premium. I find some evidence in support of hypothesis 2. In the case of the bidder CAR3 as dependent variable, I find that a representation of 1 percent of female directors on the European bidder board increases the bidder CAR3 by more than 20 percent. However, the statistically significance of the estimate is at the 10 percent level. The most of the evidence I find is in support of hypothesis 3. I find that independent female directors on the Eurozone target board are positively correlated with the bid premium and the estimate of the inside female directors is not significantly different than zero (column (3) of Panel C of Table 3). One interpretation could be that the independent female directors try to make the acquisition harder for the bidder and in doing so they uphold the interest of the shareholders, whilst the presence of inside female directors has no effect on the bid premium. I find a similar result when the dependent variable is the European target CAR3, but when I introduce the bid premium as an explanatory variable the estimate is not statistically significant anymore (column (3) and (4) of Panel C of Table 4). Furthermore I find that target independent female directors of UK firms have a larger and negative effect on the bidder CAR3 than the inside female directors of the target (column (4) of Panel D of Table 5). Again, this suggest that the presence of independent female directors on the target board makes it harder for the bidder to acquire the target (lowering the CAR) whilst the presence of the inside female directors seems not to affect the bidder CAR3 (not significant). Finally, I find that the presence of the inside female directors on the UK board is larger than the presence of the independent female directors (column (3) and (4) of Panel D of Table 4). This result does not support hypothesis 3. It suggests that inside female directors have an increasing effect on the target CAR3 whilst the presence of the independent female directors is not significantly different than zero and not affecting the target CAR3 at all. 19

5. Robustness Check on the Gender Effects in M&As In this section I present the additional investigation as robustness checks on my main findings. 5.1 Additional Investigation I employ two larger samples for the Eurozone and the UK to check whether my results are robust. The samples consist of acquisition attempts for the period 2003-2012. By not including information of the target board the samples are larger and have almost double the observations compared with the main sample. The sample that covers the Eurozone counts 61 attempts. The other sample that covers the United Kingdom (UK) consists of 71 deals. In Table 6 Panel A I present the sample descriptive statistics for the Eurozone. It can be observed that the average premium that has been paid is 30.5 percent. The average target CAR3 is 16.7 percent and the average bidder CAR3 is 0.4 percent. In Table 6 Panel B I present the similar statistics for the UK. The average premium paid in these deals is 44.5 percent. The average target CAR3 is 15.7 percent and the average bidder CAR3 is -0.2 percent. [Insert Table 6 Panel A and Panel B about here] In the Eurozone sample the fraction of bidder female CEO is 1.6 percent whilst the sample fraction of bidder female CEO in UK sample is 2.8 percent. In the Eurozone companies, on average, women make up 6.6 percent of the board. On UK boards this number is slightly smaller namely 6.3 percent. Furthermore, it can be observed that on the Eurozone bidder board the fraction of independent female directors is 4.2 percent and the fraction of inside female directors is 2.4 percent. The same numbers for the UK companies are 5.2 percent and 1.1 percent respectively. The average size of the bidder Eurozone board is 14.6 directors and 50.7 percent of these directors are on average independent. The CEO is also the Chairman of the Board (COB) in 24.6 percent of the cases. The average number of directors on the bidder UK board is 7.8 and 48.7 percent of these directors are on average independent. In 14.1 percent of the cases the CEO also acts as COB. 20

Close to 40 percent of the Eurozone deals are settled with cash as the payment method and more than 20 percent of the deals are pure stock swaps. The UK deals use only cash as method of payment in 45.1 percent of the cases. On average, 31 percent of the deals are settled with stock. More than 40 percent of the Eurozone deals are diversifying according to industry classifications of Fama and French. This number in the UK deals is 35.2 percent. The average market capitalization of the Eurozone bidder and the UK bidder is 9.6 million and 3.3 million respectively. The Tobin s Q of the Eurozone bidder is on average 1.6 and the average Q of the UK bidder is 2.0. The book leverage of the European bidder is 20 percent and that of the UK bidder is 17 percent. The average ratio of operating cash flow to total assets of the European bidder and the UK bidder is 9.5 percent and 6.9 percent respectively. The Eurozone bidder has on average total assets of almost 16 million. The average total assets of the UK bidder are almost 4 million. Comparing my larger samples with the main samples I can note that the fraction of bidder female CEO is smaller in my larger samples. The fraction of female directors on the board is similar. Furthermore I can note that in the larger sample the Eurozone bidder are slightly smaller and the UK bidders are bigger than in the main samples. The correlation pairs of the variables are also similar to the main samples. [Insert Table 7 Panel A and Panel B about here] With these larger samples I run the following cross-sectional regressions to explore the role of gender in M&As: (3) where Bid Characteristics could be the bid premium, target CAR3, or bidder CAR3. I also run regressions that extend equation (3) by classifying female directors into independent and inside female directors on the bidder board. 21

In Table 8 I present my main results. To conserve space, I present only the results regarding the gender variables. I choose to do this due to space constraints. Panel A and Panel B show the results for the Eurozone and the UK respectively when the dependent variable is the bid premium. As in my main sample I find no statistically significant results for the Eurozone or the UK regarding the gender variables. [Insert Table 8 Panel A and Panel B about here] In Panel C of Table 8 I present the results when the dependent variable is the target CAR3 for the Eurozone companies. I show in column (5) that there is a positive relation between fraction of independent female directors on the bidder board and the target CAR3. However, when I introduce the bid premium as an extra explanatory variable in column (6) this finding is not significant anymore. The coefficient of the bid premium in column (6) is statistically significant at the 1 percent level, which means that independent female directors have an effect on the target CAR3 indirectly via the bid premium. Moreover in column (6), I find that the presence of inside female directors on the bidder board affects the target CAR3 negatively at the 10 percent level. Recalling that the sample average of the target CAR3 is 16.7 percent and the coefficient is -1.112, this implies an economic effect that a 10 percent representation of inside female directors on the bidder board will reduce the target CAR3 by more than 65 percent (= 0.1 1.112/0.167). [Insert Table 8 Panel C about here] In Panel D I present the results for the UK companies when the dependent variable is the target CAR3. I find in column (5) that there is a positive effect of the presence of independent female directors on the bidder board and the target CAR3. This effect remains significant even when I introduce the bid premium as an extra independent variable in column (6). Recall that the sample average of the target CAR3 of the UK companies is 15.7 percent and the coefficient of the independent female directors in column (6) is 0.887.The economic effect of a 10 percent representation of independent female directors on the bidder board is that it increases the target CAR3 by more than 55 percent (= 0.1 0.877 / 0.157). This effect is significant at the 10 percent level. 22

[Insert Table 8 Panel D about here] In Table 8 Panel E and Panel F I present the regression results where the dependent variable is the bidder CAR3. For the Eurozone companies I show in Panel E column (4) and (5) that the fraction of inside female directors on the bidder board has an increasing effect on the bidder CAR3. The effect of inside female directors remains statistically significant even after controlling for the bid premium in column (6). The coefficient of the variable is 0.366 and the average bidder CAR3 is 0.4 percent. This implies an economic effect that a 1 percent representation of inside female directors on the bidder board increases the bidder CAR3 by more than 90 percent (= 0.01 0.366 / 0.004). This effect is significant at the 5 percent level. [Insert Table 8 Panel E about here] In Panel F I present the results for the UK companies. I find in column (2) that the presence of women on the bidder board has a positive and significant effect on the bidder CAR3. This effect remains even when I include the year fixed effects in column (3). Recall that the average bidder CAR3 for the UK companies is -0.2 percent. From this it can be showed that the economic effect of a 0.5 percent representation of female directors on the bidder board can lead to an increase of the bidder CAR3 of more than 62 percent (= 0.005 0.249 / 0.002). This effect is significant at the 5 percent level. In columns (4) (6) where I distinguish between independent and inside female directorship I find that the presence of independent female directors also has a positive and significant effect on the bidder CAR3. I show in column (4) that the presence of independent female is significant and positive, and when I control for year fixed effects in column (5) this effect does not disappear. In addition, I take the bid premium as an extra explanatory variable in column (6) and the effect still remains. The economic significance of a 0.5 percent representation of independent female directors on the bidder board increases the bidder CAR3 by more than 61 percent (= 0.005 0.245 / 0.002). This effect is significant at the 10 percent level. [Insert Table 8 Panel F about here] 23

5.2 Discussion Overall I find no significant coefficients for the Eurozone and UK that are in support of hypothesis 1. Although the size of both samples is almost twice as larger than the main sample it does not result in significant results. In the European sample, 1 of the 61 bidder firms has a female CEO. Conversely, from the 71 UK bidder firms 2 firms have a female CEO. In addition, I do not find support for hypothesis 2 and 3 when the dependent variable is the bid premium. The coefficients of the variables of interest are not significant to support the two hypotheses. However, I find support for hypothesis 2 when the dependent variable is the bidder CAR3. The presence of female directors on the bidder board in the UK has a positive effect on the bidder CAR3 (column (3) of Panel F of Table 8). I find in Panel C, D, and F support for hypothesis 3. In all three cases the coefficient of the inside female directors is not significant whilst the coefficient of the independent female directors is statistically significant. I find for the Eurozone and UK that the presence of independent female directors on the bidder board has an increasing effect on the target CAR3. I find for the UK that the presence of independent female directors has an increasing effect on the bidder CAR3 whilst the presence of inside female directors does not seem to have effect on the bidder CAR3. In column (5) of Panel E of Table 8 I find that the presence of inside female directors on the European board has a significant increasing effect on the bidder CAR3 whilst the effect of the independent female directors is not significant. This result does not support hypothesis 3. In addition to CAR3, I performed extensive checks with CAR7, CAR11, and CAR21 as the dependent variable according to equation (1) and (2). To conserve space, I do not report the regression results in this paper. Altogether, I find similar results as in Section 4. 6. Conclusion Consistent with earlier studies, I have found evidence that supports that gender also plays a role in M&A activity in the Eurozone and the United Kingdom. Although there are more statistically significant results in the current paper, I present only the main findings in this section. I find that the presence of female directors on the European bidder board is in the interest of the shareholders of the bidder company. In addition, I find that in general the bidder (target) independent female directors have an influence benefitting the bidder (target) 24