• No results found

The effect of board composition on takeover performance

N/A
N/A
Protected

Academic year: 2021

Share "The effect of board composition on takeover performance"

Copied!
42
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The effect of board composition on

takeover performance

Master Thesis

MSc. Finance, Corporate Finance track

Koen Galjaard - 10430067

Supervisor: Vladimir Vladimirov

(2)

Statement of Originality

This document is written by Koen Galjaard, 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.

(3)

Abstract

This research examines takeover performance in relation to the number of women and founding-family members present in the board of directors. The sample consists of 4,490 takeovers by US firms in 1986 till 2010; all deals have a value of at least $10 million. For founding-family members no significant effect is found. I find that the number of women present in the board is positively influencing takeover performance. The positive effect is more significant in long-run takeover returns than in short-run takeover returns.

Furthermore, the positive response of takeover performance to the presence of female directors weakens as the number of female directors increases, indicating a diminishing effect. The results are robust for deal specific controls and industry and year fixed effects. This study concludes that gender diverse boards outperform single gender boards in takeovers.

(4)

Table of Contents

Statement of Originality ... 2

Abstract ... 3

1. Introduction ... 5

2. Literature Review ... 7

2.1 Motives for acquisitions ... 7

2.2 Women in the board ... 8

2.3 Family Firms ... 11

2.4 Welfare effect of acquisitions ... 13

3. Hypotheses & Rationale for hypotheses... 15

4. Methodology ... 17 4.1 Abnormal returns ... 17 4.2 Regressions ... 18 5. Data ... 20 5.1 Data collection ... 20 5.2 Variable construction ... 20 6. Results ... 23 6.1 Interpretation of results. ... 23

6.2 Interpretation of results for controls ... 27

7. Robustness ... 29

8. Discussion ... 33

9. Conclusion ... 34

9.1 Advice on board diversity regulation ... 35

Appendix ... 36

(5)

5

1. Introduction

The market for corporate control has been subject to extensive research. While being such a widely-investigated topic, the relation between board composition and the performance of firms on the market for corporate control has seen little exposure in academic research. This paper will analyze the effect of different board compositions on the performance of firms in acquisitions. Specifically, research is conducted to determine whether presence of women and family members on the board of directors leads to significant performance deviations for takeovers in both the short and long-term.

The relevance for research on female board representation is proven by laws that force firms to assign female directors in Norway, Spain, the Netherlands and France among others (Ahern & Dittmar, 2012). The implementation of these laws and the debate over similar laws in the rest of the world emphasize that board diversity is regarded an important topic in modern society (Singh et al., 2008). Given the great relevance of board diversity in corporate governance, the effects of female board representation on shareholder value deserves both empirical and theoretical analysis (Carter et al., 2003).

The call for diversity has sparked a debate between proponents and opponents of positive discrimination for women. Proponents argue that adding women to the board changes the working dynamics of the board, causing better monitoring (Adams & Ferreira, 2009). Secondly, the variety of perspectives in a diversified board causes the evaluation of more alternatives and better decision making (Carter et al., 2003). Thirdly, Malmendier and Tate (2008) show that women tend to suffer less from overconfidence, while overconfidence is accountable for poor takeover performance via the hubris motive (Berkovitch & Narayanan, 1993). Consequently, the hypothesis in this paper is that an increased number of women on the board improves takeover returns.

Opponents argue that boards are assigned to maximize the firm’s value (Farrel & Hersch, 2005). They warn that tokenism, the phenomenon that women are merely assigned as board members due to their gender rather than their skill, is harmful for the firm (Farrel & Hersch, 2005). These arguments show the relevance for research on the effects of a diversified board on firm performance. This paper attempts to identify those effects in the context of the market for corporate control.

(6)

6 In family firms the agency problems are reduced, as the interest of a family member is better aligned with the firms’ interest (Morck et al., 1988). Therefore, assigning family directors can create value by mitigating agency costs (Villalonga & Amit, 2006). Consequently, the hypothesis in this paper is that the presence of founding-family board members increases the short and long-term gains of takeovers. On the other hand, assigning family directors can also destroy firm value, if the assigned family director is less competent than a hired professional (Burkart et al., 2003). Family firms have seen a decent amount of exposure in the academic literature, yet it has often been overlooked in the market for corporate control context (Anderson & Reeb, 2003). Evaluating family-firm performance in takeovers can deliver valuable new insights in the dynamics of value creation in takeovers.

Summarizing, adding women to the board and founding-family directors can both potentially increase takeover performance. While their potential effect is the same, both factors influence takeover returns in different ways. Family directors can affect takeover returns by decreasing agency motives, while women can do so by reducing the hubris motive, among other channels.

(7)

7

2. Literature Review

The literature review starts an analysis of different possible motives and important factors in takeovers. After which arguments are provided for potential differences in takeover motives and dynamics in gender diverse boards. Then, a summary of literature on family firms is provided and how they might differ from non-family firms in a takeover perspective. The literature section is rounded off by a summary of important papers in the merger and acquisition literature.

2.1 Motives for acquisitions

Berkovitch and Narayanan (1993) identify three general motives for an acquisition: synergy, agency and hubris. Synergy is the most commonly known motive for acquisitions and implies that the value of two firms combined is larger than the stand-alone value of those firms, because they complement each other (Andrade et al., 2003). The agency motives represent motives for directors to pursue a takeover, because they enhance the director’s welfare, while the takeover is not beneficial for the acquiring shareholders (Berkovitch & Narayanan, 1993). The hubris motives for takeovers revolve around overvaluation by the acquiring director. These valuation mistakes lead to the acquiring director overpaying in takeovers, thereby destroying shareholder value (Berkovitch & Narayanan, 1993). The hubris motive could also be described as director’s overconfidence.

Takeovers that are undertaken from a synergy motive should yield gains to the acquiring party’s shareholders, while takeovers that are undertaken from an agency or hubris motive are unfavorable for the acquiring party’s shareholders (Berkovitch & Narayanan, 1993). Previous literature suggests that takeovers are disadvantageous on average for the acquiring party (Moeller et al., 2004) (Loughran & Vijh, 1997). Essentially, the evidence in the literature then suggests that the hubris and agency play a bigger role on the acquiring side of the table than the synergy motive.

The agency motive is driven by several factors, including the phenomenon of empire building. Empire building is defined as the director’s search for status, power, compensation and prestige (Jensen, 1986). Morck, Shleifer and Vishny (1990) find that directorial objectives such as empire building are an important factor in bad acquisitions. Reducing or

(8)

8 handling the need for empire building would result in a better performance of a firm in the market for corporate control.

Malmendier and Tate (2005) analyze the effect of directorial overconfidence on corporate investments. Their qualification for an overconfident director is a director that does not diversify his stock portfolio to reduce his exposure to company specific risk (Malmendier & Tate, 2005). A strong relation is found between the overconfidence of a director and the sensitivity of investment to cash flow, meaning that overconfident directors invest more if they can do so (Malmendier & Tate, 2005). Given these findings, Malmendier and Tate suggest to limit the CEO’s investment opportunities by debt overhang or assigning more independent and cautious directors. The latter is of interest for this paper, as women are traditionally considered to be more cautious and less overconfident (Carter et al., 2003).

Summarizing, there are many possible motives for a takeover, which Berkovitch and Narayanan divided in three sub-categories; synergy, agency and hubris motives. Given that agency and hubris motives have the potential of driving unfavorable takeovers, a reduction of agency and hubris motives should positively influence takeover performance. Empire building and directorial overconfidence are part of the agency and hubris motive respectively. An increase of the number of women in the board can weaken those harmful motives and thereby increase takeover performance. The next section will provide arguments for this claim.

2.2 Women in the board

Carter et al. (2003) find a positive significant effect of female board representation on firm value. Their sample consists of fortune 1000 firms and they define board diversity as the percentage of minorities and women on the board, while measuring firm value in terms of Tobin’s Q. Two measures are used to capture the effect of women board representation on firm value, namely a dummy variable that equals one when there is at least one woman present on the board and the percentage of women on the board. Running two separate regressions, one for the dummy and one for the percentage coefficient, yields significant results for both coefficients at a five percent level. These results provide substantial evidence for a positive significant effect of women board representation on firm value (Carter et al., 2003). The researchers theoretically back their results by claiming that women

(9)

9 add new perspectives to the decision-making progress, which causes the consideration of more alternatives and better analysis of the consequences of decisions (Carter et al., 2003). Translating that theory to a takeover perspective, the presence of women on the board could positively influence takeover performance by enhancing the analysis of takeovers consequences. Thereby blocking potentially harmful takeovers, which would be undertaken in an all-male board.

In line with the findings of Carter et al. (2003), Adams and Ferreira (2009) find that women on the board tend to monitor more than men on the board. Female directors influence their male colleagues so that they have less attendance and monitoring problems than male directors without female colleagues (Adams & Ferreira, 2009). Consequently, boards that are diverse in gender give higher priority to monitoring than boards that are not (Adams & Ferreira, 2009). As previously mentioned, this could positively influence takeover performance, by better distinguishing harmful takeovers from beneficial takeovers.

In contrast of Carter et al. (2003), Ahern and Dittmar (2012) find that an increase in female board representation decreases firm value. Ahern and Dittmar use a Norwegian law for their research, the law required Norwegian companies to have at least 40% female directors. The law was initially imposed in a voluntary fashion, but after voluntary compliance failed, it was made compulsory in January 2006, with a two-year transition period. By January 2008, all firms that did not comply with the law would be forced to dissolve. Ahern and Dittmar use this exogenous shock for their research on the impact of mandated female board change in 2012. Their sample exists of all firms that are listed on the Oslo Stock Exchange from 2001 to 2009. They find that the imposed quota had a significantly negative effect on the Tobin’s Q of Norwegian companies. A mandatory 10% increase of female board representation results in a 12.4% decrease in terms of Tobin’s Q (Ahern & Dittmar, 2012). The results are supporting the hypothesis that boards are picked to maximize shareholder value, by showing that constraints on the composition of the board leads to a significant loss of value.

One might be inclined to think that female board representation is destroying value, based on the work of Ahern and Dittmar (2012). That might not be the case, however. Ahern and Dittmar highlight that, due to the constraint, not only the gender composition of the board changed, but also the composition in age, nationality, experience as CEO, education level and number of inside directors changed. These changes could also have

(10)

10 affected the performance of the firm, instead of the gender composition of the board. Additionally, the focus of this paper is performance regarding acquisitions, while Ahern and Dittmar investigate the overall firm performance. Their work does describe that firms undertake more acquisitions, but does not elaborate on that subject. This paper can identify whether female board representation has a significant effect on returns from takeovers and thereby supplement the work of Ahern and Dittmar. A small side note is mandatory here; there are no laws that force U.S. companies to assign female board members, as was the case in Norway. This is reflected in the number of companies in the sample that have one or more women on the board, which is 21%.

Farrel and Hersch investigate the likelihood of a firm assigning a female to the board, given the number of women already on the board. They use a Poisson model and distribution to determine whether the decision of adding a board member is gender neutral. The probability of a firm assigning a new female director is negatively influenced by the number of women currently on the board (Farrel & Hersch, 2005). Also, when a female director resigns from the board, the probability of the firm adding a woman to the board increases significantly (Farrel & Hersch, 2005). The underlying theory, according to Farrel and Hersch, is that firms respond to calls for diversity. Implicitly, this would mean that women directors are hired, at least to some extent, for their gender rather than their skill (Farrel & Hersch, 2005). The announcement of hiring a woman for the board yields no significant changes in the firm’s valuation, even though better performing firms have women on their board (Farrel & Hersch, 2005). This highlights the key question that arises; do women on the board cause firms to perform better or can better performing firms allow having women on the board?

Levi, Li and Zhang performed research on the relation between director gender and performance of mergers and acquisitions. They show that per female director, the amount of merger and acquisition related bids drops by 7.6% (Levi, Li & Zhang, 2014). Also, per female director on the acquiring side, the bid premium is reduced by 15.4% (Levi, Li & Zhang, 2014). The theoretical framework that is built around these findings revolves around males being more overconfident than females. Men are more confident in the precision of their predictions than women are (Barber & Odean, 2001). Additionally, men tend to see possible outcomes of their actions more positively than women do (Malmendier & Tate, 2005). Both findings regarding overconfidence affecting decisions of women in mergers and

(11)

11 acquisitions, lead to the conclusion that women are expected to be less affected by the need for empire building (Levi, Li & Zhang, 2014). In conclusion of their findings, Levi et al. state that having female directors has a favorable effect on the performance of a firm and its acquisition policy. Remarkably, they do not directly test whether the presence of women on the board of directors indeed boosts acquisition performance. The results in this paper will be able to confirm or refute the conclusion of Levi et al.

2.3 Family Firms

When the interest of shareholders and management are strongly deviated, the agency motive is more prominently present (Jensen, 1986). The shareholder’s main interest is the value of his shares, while the director’s interest is also focused on his career and personal welfare. The director’s search for personal welfare and a great career is known as empire building (Jensen, 1986). The more the directors care about the shareholder’s value, the more the interest of shareholders and directors are in agreement, the less strong is the agency problem (Jensen, 1986). The interests of a founding-family director differ from a director that is not part of the founding-family. In the sense that a family member of the founding family has a stronger bond to the firm and therefore cares about the future of the firm, rather than only caring about his future career and personal welfare (Anderson & Reeb, 2003). Hence, agency motives should be less significant in family-managed firms. Consequently, a less significant agency motive should lead to better takeover performance, according to the theory of Berkovitch and Narayanan (1993).

When the firm is managed by one or more family members, interest of management and owners is expected to be more aligned than under non-family management (Morck et al., 1988). Family management is thereby mitigating the first agency problem; the conflict of interest between owners and management. Hence, according to the agency theory family management is value enhancing by mitigating agency costs (Villalonga & Amit, 2006).

However, hired professionals could be significantly better directors than the family members present in the management. In that case the higher skill of the professionals could be more valuable than the mitigation of the agency costs by the family management. Therefore, family directors are potentially value decreasing (Burkart, Panunzi & Shleifer, 2003).

(12)

12 Family directors have a longer investment horizon than non-family directors, because they are more invested in the company (Anderson & Reeb, 2003). Also, family members have a longer tenure as director and manager, which can be derived from their lower turnover rate (Anderson & Reeb, 2003). Due to the long investment horizon, family directors are less likely to excavate the company’s resources for short term gains that are value destroying in the long run. Hence, family management mitigates value destructing via its long investment horizon (Anderson & Reeb, 2003). Given these findings and translating them to takeovers, one would assume family directors act more prudently on the market for corporate control than non-family directors. Miller et al. (2009) find that family firms M&A activity is indeed more prudent and is based on continuity rather than empire building.

Large shareholders can use their power to extract benefits of the firm to their own advantage (Jensen, 1986). For example; force the director to make a deal with a firm that is owned by the same large shareholder. If the large shareholder is entitled to ten percent of the profits of the firm, but is also entitled to sixty percent of the profits of another firm, he will generate additional gains for himself by moving the profits from the prior firm to the latter. Essentially transferring money from the firm to their own pockets, at the cost of the other shareholders in the firm. This agency problem, where one shareholder is making profits at the cost of the other shareholders, is labelled the second agency problem (Villalonga & Amit, 2006). This problem is more prominently present in cases where the large shareholder is an individual or a family (Villalonga & Amit, 2006). In that case, the incentive for extracting gains from the firm is bigger, because the benefits are shared among a limited amount of people. If the large shareholder is an entity owned by many individuals, for example a pension fund or bank, the incentive to expropriate money from the firm is smaller, because the gains from expropriating money from the firm are smaller. On the other hand, the incentive for monitoring the director is also smaller, whereas that incentive is bigger in when the large shareholder is an individual or family (Burkart et al., 2003).

Summarizing, if the large shareholder is an institution the first agency problem, between directors and owners, is bigger than the second agency problem. The second agency problem, between shareholders, is more substantial if the large shareholder is an individual or a family (Villalonga & Amit, 2006).

These agency problems both influence shareholder value, however it is unclear which of the two has a more significant effect on shareholder value. As Morck and Yeung

(13)

13 (2003) point out, ‘much more research is required on the welfare and shareholder value effects of family control in a business’. Shleifer and Vishny (1997) argue that concentrated ownership leads to better corporate governance, but they do not take the second agency problem into account. Concentrated ownership in a family setting may indeed solve the classic agency problem as described in the literature by the likes of Jensen and Meckling (1976). However, new agency problems arise in such a setting. Among these problems are the second agency problem as described in the previous paragraph, also a director can neglect the wishes of minority shareholders and solely act in favor of the block holder. If the director and the block holder are both founding-family members, their private relation increases the probability that the director will solely act in favor of the block holder (Villalonga & Amit, 2006). Thereby potentially harming the overall shareholder value, to the benefit of one large block holder.

2.4 Welfare effect of acquisitions

Even though literature on the topic of acquisitions is quite rich, the results regarding acquisition performance in different leading papers are contradictory. The inconsistency in conclusions by previous literature can be attributed to the difference in measurement of performance, as well as the use of different time periods (Fama, 1998). In this paper both short- and long-run performance are analyzed with regard to takeovers. After analyzing the takeover performance, the effect of different board compositions on the takeover performance is studied more extensively.

Andrade, Mitchell and Stafford (2001) find that the combined stockholders of the acquiring and target firm gain from a merger or acquisition. However, they state that most the gains are transferred to the stockholders of the target firm. Many other papers find that target stockholders are on average the biggest winners of an acquisition (Bradley, Desai & Kim, 1988) (Loughran & Vijh, 1997). Andrade, Mitchell & Stafford (2001) first focus on the immediate market response following the completion of a merger, in which they find positive returns for both the target and acquiring firm on average. Later, they analyze the long-term performance of merged firms relative to their Industry peers and find that mergers improve the long-term performance of merging parties (Andrade et al., 2001). Theoretically the findings are backed by the synergy theory; the view that increased

(14)

14 efficiency between two firms leads to increased future performance, which is reflected in the stock price of the firms (Andrade et al., 2001). This view is recognized as the traditional view and has been criticized and tested by the academic community in recent decades.

Rau and Vermaelen (1998) represent the view of the new school, who argue that short term abnormal returns do not capture the full effects of an acquisition. By claiming that short term abnormal returns do not represent the full effect of the market’s response to an acquisition, Rau and Vermaelen (1998) are essentially claiming that the market is unable to correctly value the effects of an acquisition upon completion. Therefore, the discussion whether short term market reactions capture the full effect of an event, is implicitly a discussion about the efficient market hypothesis. In their results Rau & Vermaelen (1998) show that firms with high book-to-market ratios underperform in the short run and over perform in the long-run, compared to firms with low book-to-market ratios. The difference in performance is attributed to the market’s overly positive view on a firm with a low book-to-market ratio and too negative view on a high book-to-market ratio firm (Rau & Vermaelen, 1998). This plausible explanation of their findings combined with the discrepancy in short- and long-run results for acquisitions indeed form a strong argument against the efficient market hypothesis. Consequently, the call for long-run performance analysis to capture the full effect of an acquisition is justified.

This paper will focus on the gains or losses for acquiring firms in takeovers, where Andrade, Mitchell and Stafford (2001) focused on the net gains of a merger or acquisition by reviewing the gains and losses of both parties involved. The focus on acquiring firm’s gains or losses is justified in a takeover, because the target firm’s gains should be equal to the premium paid in the deal (Loughran & Vijh, 1997). The consensus in the literature is that target shareholders benefit from acquisitions, but acquisitions are not beneficial for acquiring shareholders (Moeller et al., 2004) (Loughran & Vijh, 1997). Yet, acquisitions continue to take place on a massive scale. In 2015, a total of 44,000 M&A deals took place with a value of over 4.5$ trillion dollars (Institute for Mergers, Acquisitions and Alliances).

(15)

15

3. Hypotheses & Rationale for hypotheses

In line with Levi, Li & Zhang (2014) and backed by the findings of Adams & Ferreira (2009), Malmendier and Tate (2008) and Carter, Simkins & Simpson (2003), I expect female directors to have a positive influence on the takeover performance of firms. The overconfidence argument brought forward by Malmendier and Tate (2008) indicates that boards with female presence should perform better takeovers than boards without. Secondly, the monitoring argument proposed by Adams & Ferreira (2009) indicates that diversified boards, in terms of gender, should outperform non-diversified boards in takeover performance. Thirdly, the argument of Carter, Simkins & Simpson (2003) that boards with women evaluate more alternatives implies that boards with women are better in evaluating possible takeovers, hence they should perform better. Concluding these three arguments, I expect the presence of women on the board to have a positive influence on takeover performance in both the short- and long run.

H1: Firms with a woman or multiple women on the board of directors perform better in takeovers than firms with an all-male board.

The second hypothesis is that, for the same arguments, the more women in the board, the better takeover performance will be.

H2: The higher the percentage of women on the board the better a firm performs in takeovers.

A family firm differs from a non-family firm in many ways. Some of the differences can potentially influence takeover returns on both the short- and long-term. Firstly, the relation of a family member to the firm can possibly reduce agency costs, as the interests of the family member are more in line with the interests of the firm (Anderson & Reeb, 2003). Reducing the agency motive for takeovers can increase takeover performance (Berkovitch & Narayanan, 1993). Secondly, family directors have a longer investment horizon than non-family directors, because they tend to have a longer tenure at the firm (Anderson & Reeb,

(16)

16 2003). Therefore, they are less likely to seek takeovers that yield short-run returns but are harmful in the long-run. Thirdly, the corporate control strategy of family firms is based on continuity rather than empire building (Miller et al., 2009). Consequently, less harmful takeovers should be undertaken, due to a lower level of empire building behavior among directors (Berkovitch & Narayanan, 1993). On the other hand, by selecting directors from the pool of family members, the competence of family directors can be lower than a hired professional (Burkart, Panunzi & Shleifer, 2003). Also, other wealth decreasing agency problems, as discussed in the literature review, can arise from a firm being run by founding-family members. Combining all the pros and cons of founding-family firms and non-founding-family firms in takeovers, the expectation is that family firms outperform non-family firms. The definition of a family firm in this paper is related to the presence of a founding-family director, the following hypotheses are formed.

H3: Firms with a family member or multiple family members on the board of directors perform better in takeovers than firms without a family member on the board.

H4: The higher the percentage of family members on the board the better firms perform in takeovers.

(17)

17

4. Methodology

4.1 Abnormal returns

An important aspect of an event study is the computation of abnormal returns (Andrade et al., 2001). To compute abnormal returns, a model is chosen to predict the normal returns. In this paper the market model is chosen, introduced by Brown and Warner (1985). As Fama (1998) points out, every model has its limitations and not a single model completely describes expected returns. The market model is a firm specific model that uses regressions of a stock’s return on the market return in an estimation period, to predict a stock’s expected return in an event period (Moeller et al., 2003) (Fama et al., 1969). The estimation period in this paper is 230 days prior to the takeover till 30 days prior to the takeover. The event period is not included in the estimation period, to prevent the event from biasing the normal performance model estimates in the direction of the event effect (Brown & Warner, 1985). Regressions during this period are used to predict the ‘normal’ return of a firm, had it not completed the takeover. The regression for the normal return is:

𝐸𝐸(𝑅𝑅)𝑗𝑗𝑗𝑗 =∝𝑗𝑗+ 𝛽𝛽𝑗𝑗∗ 𝑅𝑅𝑚𝑚𝑚𝑚𝑗𝑗+ 𝜀𝜀𝑗𝑗𝑗𝑗

𝐸𝐸(𝑅𝑅)𝑗𝑗𝑗𝑗 = Predicted ‘normal’ return for stock j at date t.

𝛽𝛽𝑗𝑗 = Sensitivity of stock j to the market return.

𝑅𝑅𝑚𝑚𝑚𝑚𝑗𝑗 = Market return.

The abnormal return is calculated by subtracting the normal return from the actual return. 𝐴𝐴𝑅𝑅𝑗𝑗𝑗𝑗 = 𝑅𝑅𝑗𝑗𝑗𝑗− 𝐸𝐸(𝑅𝑅)𝑗𝑗𝑗𝑗

𝐴𝐴𝑅𝑅𝑗𝑗𝑗𝑗 = Abnormal return for stock j at time t.

𝑅𝑅𝑗𝑗𝑗𝑗 = Actual return of stock j at time t.

(18)

18 The normal and abnormal returns are computed for four different event periods; the announcement, short term, three year and five year periods. The announcement period consists of five days, two days prior to the takeover until two days after the takeover. The short-term period runs from twenty days pre-takeover till the same amount of days after the takeover. The three year and five-year event windows are self-explanatory. Nota bene: for the five-year window one extra day is included, because a leap-year is very likely to be included in a five-year window. Cumulative abnormal returns are calculated for all periods to test takeover performance.

𝐶𝐶𝐴𝐴𝑅𝑅𝑗𝑗(𝑗𝑗1,𝑗𝑗2) = � 𝐴𝐴𝑅𝑅𝑗𝑗 𝑗𝑗2 𝑗𝑗=𝑗𝑗1

𝐶𝐶𝐴𝐴𝑅𝑅𝑗𝑗(𝑗𝑗1,𝑗𝑗2) = Cumulative abnormal return of stock j from t1 till t2.

𝐴𝐴𝑅𝑅𝐴𝐴 = Abnormal return of stock j 4.2 Regressions

OLS regressions are used to estimate the effect of the variables of interest on cumulative abnormal returns for all event windows. The computation of all variables is reported in the data section. Both dummy and continuous variables for the presence of women and family members on the board are used.

For robustness purposes, more variables have been added to the regressions. Due to the addition of these control variables, the sample size has decreased. Nevertheless, the sample group still contains 2239 takeovers. Several control variables have been investigated in the existing literature. For example, the size of the acquiring firm is significant for the announcement return for acquiring-firm shareholders (Moeller et al., 2004). Smaller firms report 2% higher announcement returns than bigger firms do, which makes controlling for size essential (Moeller et al., 2004).

Secondly, the literature suggests correcting for book-to-market ratio when analyzing abnormal returns (Fama & French, 1993) (Mitchell & Stafford, 2000) (Andrade et al., 2001). Fama and French show that the book-to-market ratio, combined with size, is a good proxy for sensitivity to common risk in stock returns. Hence, it is vital to include book-to-market ratio in the regressions.

(19)

19 Thirdly, the deal value is added as an explanatory variable. The bigger the deal, the bigger the impact of the deal on the acquiring firms’ financials. Also, the deal value is equal to the value of the target’s assets and a premium and is therefore a good proxy for the target company’s size.

Also, regressions are run with industry fixed effects in table 11 through 14, to control for possible industry effects in the results. The industries are defined in eight sub-categories via sic codes. The different categories are: 1. agriculture, forestry and fishing (sic:0100 – 0999), 2. mining (sic:1000-1499), 3. construction (sic:1500-1799), 4. manufacturing (sic:2000-3999), 5. transport, communication and energy (sic: 4000-4999), 6. wholesale (sic:5000-5199), 7. retail (sic:5200-5999) and 8. services (sic:7000-8999).

Lastly, year fixed effects have been added to the regressions, to control for year specific effects in the sample. The sample period 1986 till 2010 is divided in five periods of five years to control for trends in different periods. The results of these regressions are found in table 14.

(20)

20

5. Data

5.1 Data collection

The data on acquisition for this paper is acquired from the Thomson one M&A database. The sample period for the acquisition data is January 1st 1986 till December 31th 2010. Only

deals with a value of ten million and above are taken into consideration, also all acquirers are U.S. based firms. Additionally, only takeovers are considered; the acquirer has less than the majority of shares prior to the deal and controls all of the shares after the deal.

Daily stock data is collected from the CRSP database for the period January 1st 1985 till December 31th 2016. The chosen period allows constructing both an estimation window prior to all takeovers and a five-year performance window thereafter. The stock data and M&A data can’t be merged directly, since Thomson One contains no identifiers that are present in CRSP except for cusip codes. However, the cusip codes in CRSP are updated regularly, while the cusip codes in Thomson One are historical. Simply merging those two would lead to mismatching companies and losing many observations. Therefore, historical cusip codes have been obtained and merged with the CRSP data. Finally, the stock data and M&A data have been merged via the historical cusip codes. Director data is acquired from the ISS database. The data on directors is on a yearly basis and supplies information to construct variables for family members and women in the board of directors.

5.2 Variable construction

Relative

Relative is a dummy variable that equals 1 when one or more founding-family members are present in the board of directors of a firm.

Woman_board

Woman_board is a dummy variable that equals 1 when one or more women are present in the board of directors of a firm.

(21)

21

% Relative

% Relative is the percentage of founding-family members present in the board of directors.

% Woman

% Woman represents the percentage of women present in the board of directors.

Total assets

Total assets owned by the acquirer in millions of USD, in the year before the deal.

Deal value

Deal value is the amount paid to acquire the target in millions of USD.

Book / market

Book/market is the book-to-market ratio of the acquirer in the year before the deal.

Total assets(ln)

The ln value of Total assets, this variable is used for total assets in the rest of the paper.

Deal value(ln)

(22)

22

Table 1: Summary Statistics.

All values are reported in USD. The summary statistics show that in 14.3% of the 4490 deals in the sample, a family member was active in the board of the acquirer at the time of the deal. Also in 21.07% of those deals one or more women were present in the board of the acquirer at the time of the deal. These percentages correspond to 642 firms with a founding-family board representative and 946 firms with a female board representative. % Relative represents the percentage of founding family members active in the board of a firm. % Woman represents the percentage of women active in the board of a firm. Total assets represent the total number of assets owned by the acquirer, prior to the deal. The value of the assets and deal value have been normalized by taking the natural logarithm of those values.

Variable Obs. Mean Std. Dev. Min Max

Relative 4,490 0.1430 0.3501 0 1 Woman board 4,490 0.2107 0.4078 0 1 % Relative 4,462 0.0599 0.0343 0 0.6014 % Woman 4,462 0.1489 0.0879 0 0.6250 Total assets 3,370 10,464 37268 12 795,337 Deal value 4,490 545 2608 10 89,168 Book / market 2,239 0.3902 0.2998 0.0043 2.4455 Total assets(ln) 3,370 7.7914 1.6872 2.4825 13.5865 Deal value(ln) 4,490 4.7674 1.5217 2.3026 11.3983

(23)

23

6. Results

The results of the simple regressions on the effects of a founding-family board member and a female board member are presented in table 2 and table 3 respectively. In table 4 and 5 control variables have been added to the regressions. Table 6, 7 and 8 contain regressions with the percentage of founding-family and female board members, including control variables. In tables 9 and 10 the sample is divided in two groups, one group contains all firms with fewer than 15% women in the board, the other contains all firms with at least 15% women in the board. Tables that are not in the text can be found in the appendix. 6.1 Interpretation of results.

The coefficients for the effect of female board members on long term takeover performance are significant at a one percent level. They indicate a 19.9% and 30.9% increased performance in respectively a three and five-year period after the takeover. These effects are reduced in level and significance by the addition of the control variables, but remain significant at a ten percent level. Also, the effect of female board representation on announcement returns is significant at a five percent level after the addition of control variables. The results provide evidence in favor of the first hypothesis; that firms with a woman or multiple women on the board of directors perform better in takeovers than firms with an all-male board. This forms suggestive evidence for the theory that diversified boards make better decisions (Carter et al., 2003) (Adams & Ferreira, 2009). Also, it concurs with the results of Levi, Li and Zhang (2014) that women take less risk in takeovers.

The results remain robust when the dummy variable for the presence of a woman in the board is replaced by the variable that measures the percentage of women in the board. The results in table 6 show a positive effect that is significant at a ten percent level. The coefficients correspond to an average increase of 0.81% and 1.58% in takeover returns respectively for three and five years, for a 1% increase of women in the board. The coefficients remain nearly unchanged after the addition of the percentage of family members in the board, which can be seen in table 8 in the appendix. Overall, the results confirm the first hypothesis; firm with at least one woman on the board outperform firms with an all-male board in takeovers. Hypothesis two is discussed in the next paragraph.

(24)

24 Table 4: gender regressions with controls.

Woman_board is a dummy variable that equals 1 when one or more women are active in the acquiring firm’s board and 0 otherwise. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Total assets represent the natural logarithm value of the total assets of the acquirer. Book / market represents the book-to-market ratio of a firm.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

woman_board 0.007** 0.004 0.170* 0.275* Deal value 0.0003 -0.002 -0.097*** -0.164*** Total assets -0.003*** -0.0006 0.109*** 0.213*** Book/market 0.010** 0.016 1.077*** 1.696*** N 2239 2239 2239 2239 adj. R-sq 0.008 0.001 0.032 0.033 * p<0.10, ** p<0.05, *** p<0.01.

Table 6: gender percentage regressions with controls.

% Woman represents the percentage of women active in the acquiring firm’s board. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Total assets represent the natural logarithm value of the total assets of the acquirer. Book / market represents the book-to-market ratio of a firm.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

% Woman -0.006 0.014 0.813* 1.583* Deal value 0.000 -0.002 -0.094*** -0.158*** Total assets -0.002** -0.001 0.110*** 0.208*** Book/market 0.009 0.008 1.132*** 1.752*** N 2211 2211 2211 2211 adj. R-sq 0.005 0 0.029 0.029 * p<0.10, ** p<0.05, *** p<0.01.

For the simple regressions, no significant results have been found for the presence of a family member on the board. Hence, there is no evidence for a significant difference in family firm and non-family firm takeover performance. After the addition of controls, the effect of a family board member on short term performance becomes significant at a ten percent level. The coefficient indicates a 3,3% decrease in takeover returns for firms with a founding-family board member in the short-run. All other regressions on family board representation yield insignificant results. Therefore, the third hypothesis is rejected, there is

(25)

25 no evidence that firms with founding-family board members perform better in takeovers than firms without founding-family board members. On the contrary, there is weak suggestive evidence that firms with a family board member underperform in takeovers. This is in line with the results of Burkart et al. (2003) that family directors perform worse on average than hired professionals, because they are less competent.

Table 5: relative regressions with controls.

Relative is a dummy variable that equals 1 when a founding family is active in the acquiring firm’s board and 0 otherwise. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Total assets represent the natural logarithm value of the total assets of the acquirer. Book / market represents the book-to-market ratio of a firm.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

Relative 0.004 -0.010* -0.099 -0.072 Deal value 0.000 -0.002 -0.098*** -0.165*** Total assets -0.003*** -0.000 0.123*** 0.237*** Book/market 0.009* 0.015 1.058*** 1.662*** N 2239 2239 2239 2239 adj. R-sq 0.007 0.002 0.031 0.032 * p<0.10, ** p<0.05, *** p<0.01.

The percentage of women board members in the sample is 14.89%, which concurs with the research of Deloitte and MSCI who find 12% and 15% women in board of directors. Since the percentage of women in the board is low in the sample and the population, the regressions in tables 3, 4, 6 and 8 show the effects for female board representation in a setting with a relatively small number of women in the board. Therefore, the positive result of adding more women in the board on takeover returns might only hold for firms with a low number of women on the board. In other words, there is possibly a diminishing positive effect of female board representation on takeover returns.

This theory is put to the test in tables 9 and 10. The tables show a significant positive effect of an increased percentage of women in the board, when firms with fewer than 15% female board representation are considered. Moreover, the significance of the effect is increased to a five percent level for the regressions on announcement and five year abnormal returns. On the other hand, when a firm has over fifteen percent female board representation, none of the regressions yields significant results. In fact, the coefficients

(26)

26 become negative. Consequently, tables 9 and 10 indicate that there is indeed a diminishing effect of an increased number of women in the board on takeover gains. The results suggest that the second hypothesis holds for firms with under 15% women in the board. Yet, it does not hold for firms that have at least 15% women in the board. Therefore, the second hypothesis; ‘the higher the percentage of women on the board the better a firm performs in takeovers’, is rejected.

These results can explain the results Ahern and Dittmar (2012) found in their Norwegian sample. The law that enforced public firms to have 40% female board representation, caused a decline in terms of Tobin’s Q as the number of women in the board rose (Ahern & Dittmar, 2012). Although other factors could have influenced the observed effect on Tobin’s Q, the findings of Ahern and Dittmar (2012) indicate that the initially positive effect of adding more women to the board could become negative when female board representation is already high.

Summing up the results in terms of the hypotheses; the results confirm the first hypothesis that firms with at least one woman in the board of directors perform better in takeovers than firms with an all-male board. The second hypothesis holds partially, as a diminishing positive effect for the percentage of women in the board is found. The third and fourth hypothesis are rejected on basis of the results.

Table 9: regressions with condition <15% woman.

The table shows values for firms with fewer than 15% women on the board. % Woman represents the percentage of women active in the acquiring firm’s board. % Relative represents the percentage of family members active in the acquiring firm’s board Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Total assets represent the natural logarithm value of the total assets of the acquirer. Book / market represents the book-to-market ratio of a firm.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

% Woman 0.080* 0.120 2.469* 4.223** % Relative 0.061 -0.073 -2.287 -2.726 Deal value 0.002* -0.005* -0.197*** -0.336*** Total assets -0.007*** -0.006* 0.139* 0.269** Book/market 0.010 0.016 1.038*** 1.365*** N 1070 1070 1070 1070 adj. R-sq 0.014 0.009 0.035 0.032 * p<0.10, ** p<0.05, *** p<0.01.

(27)

27 Table 10: regressions with condition >15% woman and.

The table shows values for firms with over 15% women on the board. % Woman represents the percentage of women active in the acquiring firm’s board. % Relative represents the percentage of family members active in the acquiring firm’s board. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Total assets represent the natural logarithm value of the total assets of the acquirer. Book / market represents the book-to-market ratio of a firm.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

% Woman -0.025 -0.029 -0.427 -0.705 % Relative -0.048 0.000 -0.886 -1.661 Deal value -0.001 0.000 -0.032 -0.055 Total assets 0.000 0.001 0.080** 0.159*** Book/market 0.01 0.000 1.209*** 2.154*** N 1141 1141 1141 1141 adj. R-sq 0 0.000 0.024 0.031 * p<0.10, ** p<0.05, *** p<0.01.

6.2 Interpretation of results for controls

The effects of deal value, total assets and book-to-market ratio are all significant at a one percent level for the long-term sample periods. This is additional evidence for the claim of previous literature that these factors are important indicators of long-run takeover performance (Fama & French, 1993) (Moeller et al, 2004) (Loughran & Vijh, 1997).

Moeller et al. (2004) find that small firms experience two percent higher announcement returns. While these findings are confirmed by the results of this paper, the opposite effect seems to hold for the long term abnormal returns. Large firms outperform small firms after three years and five years by respectively 10,5% and 12,4%. These results can be explained by the work of Loughran and Vijh (1997), who investigate long-run acquisition performance. They find that the post-acquisition abnormal returns decrease as the relative size of the target to the acquiring firm increases (Loughran & Vijh, 1997). In other words, the bigger the acquiring firm relative to the target firm, the higher the abnormal returns. Since large firms don’t necessarily make larger acquisitions (Moeller et al., 2004), the results of Loughran and Vijh are in line with the results in this paper.

A low book-to-market ratio suggests that a firm is overvalued by the market, while a high book-to-market ratio suggests the opposite (Shleifer & Vishny, 2003). Overvalued firms are more likely to engage in takeovers, as they can exchange their stocks at a better rate

(28)

28 than undervalued firms (Rau & Vermaelen, 1998). However, Rau and Vermaelen find that overvalued firms, with a low book-to-market ratio, perform worse three years after an acquisition than firms with a higher book-to-market ratio. That statement is consistent with the results in this paper, where the coefficients for book-to-market ratio are positive and significant at a one percent level. These results indicate that relatively undervalued firms perform significantly better than relatively overvalued firms in the post-takeover years. Which can be explained by the market revaluating the acquiring firm, hence partly correcting the over- or undervaluation (Rau & Vermaelen, 1998).

(29)

29

7. Robustness

Robustness checks are done in table 11 through 16. Industry fixed effects are added to check for possible industry bias. The results in table 11 and 12 show that the addition of industry fixed effects, lead to increased levels of significance for the dummy and percentage variable of female board representation. The coefficients are now significant at a one percent level for the five year sample period and the three year period for the dummy regression. Also, the coefficient for the three year sample period is positive and significant at a five percent level. These findings increase the robustness of the results that are reported in the previous section, as the results remain significant after the addition of the industry fixed effects. In fact, they become significant at a higher level.

After adding the logarithm of total assets to the industry fixed effects regressions, the significance of the dummy and percentage variable for female board representation are reduced in significance. The long-term coefficients remain significant at a ten percent level, as was the case in most of the previous regressions. Also, the announcement coefficient for the dummy variable is significant at a five percent level. Yet, the total assets variable does seem to reduce the explanatory power of the variables for the number of women in the board. Therefore, it seems like the total assets of a firm and the number of women in the board are correlated. This is put to the test in the correlation matrix in table 15, the results indeed show that the number of women in the board and total assets are strongly related. The coefficients in the table that represents the correlation between the women variables and the total assets variable are negative However, this represents a positive relation between the variables. Since the reported correlation shows that the coefficients diminish each other in terms of effect on cumulative abnormal return, hence they run in the same direction. Consequently, there is some multicollinearity in the regressions where a variable for women in the board and total assets are present. This makes it harder to interpret the weight and significance of the coefficients, hence it reduces the robustness of the results found in previous tables.

In table 16 the amount of takeovers per five years are presented. In table 16 the year fixed effects are shown. Adding year fixed effects to the regressions increases the significance of the variable for the number of women in the board. The coefficients are now

(30)

30 positive and significance at a five percent level. The fact that the results hold after the addition of year fixed effects provides additional evidence for the first hypothesis, therefore increases the robustness of the results found in the previous section.

Robustness checks on the variables for family board members yielded no significant results. Therefore, they have not been reported in this section. Both industry and year fixed effects have been added to the regressions in table 5 and 7, but no significant results were found in any of those regressions. Hence, hypothesis three and four do not hold after the addition of industry and year fixed effects. These outcomes provide further proof that there is indeed no difference in takeover performance between firms that have founding-family board members and firms that don’t. Thereby adding to the robustness of the findings in the results section.

Table 12: industry fixed effects and woman percentage.

% Woman represents the percentage of women active in the acquiring firm’s board. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Book / market represents the book-to-market ratio of a firm. Industry 1 represents the agriculture, forestry and fishing Industry. Industry2 represents the mining industry. Industry3 represents the construction industry. Industry4 represents the manufacturing industry. Industry 5 represents the transport and energy industry. Industry 6 represents the wholesale industry. Industry 7 represents the retail industry. Industry 8 represents the services industry.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

% Woman -0.026* 0.001 1.448** 2.646*** Deal value 0.000 -0.002 -0.0527* -0.077* Book/market 0.011** 0.010 1.103*** 1.697*** Industry1 0.002 -0.095*** -0.792 -0.594 Industry2 -0.009 -0.047* -0.310 -0.404 Industry3 -0.020 -0.085** -1.209 -1.918* Industry4 -0.003 -0.038 -0.240 -0.015 Industry5 -0.006 -0.041 -0.272 0.038 Industry6 -0.009 -0.052* -0.225 0.006 Industry7 -0.003 -0.051* -0.278 0.031 Industry8 -0.003 -0.047* -0.339 -0.147 N 2211 2211 2211 2211 adj. R-sq 0.001 0.001 0.023 0.023 * p<0.10, ** p<0.05, *** p<0.01.

(31)

31 Table 14: industry fixed effects and woman relative.

% Woman represents the percentage of women active in the acquiring firm’s board. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Total assets represent the natural logarithm value of the total assets of the acquirer. Book / market represents the book-to-market ratio of a firm. Industry 1 represents the agriculture, forestry and fishing Industry. Industry2 represents the mining industry. Industry3 represents the construction industry. Industry4 represents the manufacturing industry. Industry 5 represents the transport and energy industry. Industry 6 represents the wholesale industry. Industry 7 represents the retail industry. Industry 8 represents the services industry.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

% Woman -0.009 0.012 0.943* 1.585* Deal value 0.000 -0.002 -0.090*** -0.149*** Total assets -0.002** -0.001 0.110*** 0.213*** Book/market 0.010* 0.010 1.161*** 1.809*** Industry1 0.004 -0.094*** -0.873 -0.751 Industry2 -0.006 -0.045 -0.469 -0.711 Industry3 -0.019 -0.084** -1.288 -2.070* Industry4 -0.001 -0.036 -0.360 -0.248 Industry5 -0.003 -0.039 -0.427 -0.264 Industry6 -0.006 -0.051* -0.346 -0.229 Industry7 -0.002 -0.050* -0.348 -0.104 Industry8 -0.002 -0.047* -0.383 -0.234 N 2211 2211 2211 2211 adj. R-sq 0.004 0.001 0.028 0.031 * p<0.10, ** p<0.05, *** p<0.01. Table 15: correlation coefficients.

e(V) % Woman woman_board Deal value Total assets Book/market

% Woman 1.000

woman_board 0.006 1.000

Deal value 0.008 -0.122 1.000

Total assets -0.223 -0.016 -0.399 1.000

(32)

32 Table 16: year fixed effects.

% Woman represents the percentage of women active in the acquiring firm’s board. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Book / market represents the book-to-market ratio of a firm.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

% Woman -0.003 0.021 1.021** 1.995** Deal value 0.000 -0.002 -0.087*** -0.145*** Total assets -0.002** -0.001 0.100*** 0.190*** Book/market 0.008 0.004 1.055*** 1.594*** 1986-1990 0.001 -0.002 -0.064 -0.132 1991-1995 -0.003 0.005 -0.162 -0.253 1996-2000 -0.005 -0.016* -0.283** -0.603** 2001-2005 -0.002 0.002 -0.138 -0.198 2006-2010 0.004 -0.003 -0.089 -0.186 N 2211 2211 2211 2211 adj. R-sq 0.006 0.004 0.037 0.043 * p<0.10, ** p<0.05, *** p<0.01.

(33)

33

8. Discussion

The coefficients in the regressions on female board representation and takeover performance show a positive significant result in all long-run regressions and some short-run regressions. The direction of causality is not identified, however. In this paper it is assumed that an increasing number of women in the board has a positive effect on takeover returns. It could be that firms with better takeover performance can allow having more women on the board, thereby implying that causality runs in the opposite direction. Even though the latter seems farfetched, it can’t be ruled out by the findings of this paper. Future research should focus on identifying causality, via Granger causality tests for example.

The results show that women work in larger firms and those larger firms perform better in takeovers in the long-run. The multicollinearity of these variables decreases the validity of the results. Future research should attempt to isolate the effect of women in the board from the size of firms. Interaction terms could also be introduced to help deal with multicollinearity.

The regressions on the relation between founding-family board members and takeover performance are almost entirely insignificant. However, the dynamics in firms that have founding-family members on the board and firms that don’t can be significantly different. Many of those possible differences are reported in the literature section. In other words, the insignificant results can be driven by: a) the family and non-family board members do not differ significantly or b) the positive and negative differences between family and non-family board members offset each other. Neither of the two options can be ruled out by the findings in this paper. To discover which of the two options holds in reality, more detailed information on the firms is required. For example a sample with family ownership and corporate governance characteristics would help.

(34)

34

9. Conclusion

The results in this paper show that there is a positive relation between the number of women in the board of directors and takeover returns. Therefore, it opposes the tokenism argument by Farrel and Hersch (2005) that women are merely assigned for gender instead of skill. By showing that the presence of women on the board has a positive significant effect on takeover performance, this paper complements the findings of Levi, Li and Zhang (2014), which show that boards with women pay lower premiums and make fewer bids. It also concurs with the findings of Carter et al. (2003) and Adams and Ferreira (2009) that a broader perspective and better monitoring provided by women can lead to increased performance.

A diminishing positive effect on takeover returns for the number of women in the board of directors on takeover performance is found. The relation is positive and significant for firms with fewer than 15% female directors, but becomes insignificant when considering firms that have more than 15% women on the board of directors. This forms an interesting addition to the work of Ahern and Dittmar (2012), who analyze a sample that is constrained by a law, enforcing a minimum of 40% women in the board. Ahern and Dittmar (2012) find a decrease in terms of Tobin’s Q related to an increase in the number of women on the board. Admittedly, the findings of Ahern and Dittmar (2012) are likely to be influenced by other factors than gender. Yet, investigating whether the initially positive effect of more women in the board turns negative if the number of women is already high, forms an interesting question for future research.

Family firms performance does not differ significantly from non-family firms in a takeover. There are different explanations for results. It is clear that the reduction in agency problems in a family firm (Villalonga & Amit, 2006), does not create value in a takeover setting. Firstly, the reduction in the agency motive can be offset by the possible lower competence of family directors (Burkart, Panunzi & Shleifer, 2003). Secondly, other agency problems can arise in a family firm (Villalonga & Amit, 2006). These agency problems can cancel out the positive effect of the reduced classical agency problem. Thirdly, the isolated effect of a reduction in the agency motive in family-firms can simply be insignificant. Overall,

(35)

35 the results lead to the conclusion that family firms do not differ from non-family firms in takeovers.

9.1 Advice on board diversity regulation

The results in this paper yield important implications for policy makers and corporate governance. It is shown that a higher number of women in the board has a positive impact on takeover returns, hence it is in all probability that women positively impact overall firm performance. Although indications have been found that the positive effect eventually diminishes, it is undisputed that gender diverse boards outperform all male boards.

I suggest following the example of Sweden. Sweden has implemented a voluntary policy for firms to increase female board representation to 25% or above, warning firms that laws will follow if the target is not met. This policy encourages firms to increase board diversity, without disrupting the process of assembling boards as rudely as in Norway, where a mandatory 40% representation led to the destruction of firm value (Ahern & Dittmar, 2012). The example of Norway shows that a voluntary transition period is important in aiding firm value. Furthermore I suggest a minimum of one woman per board, as it is unclear what the firm value implications of the 25% ratio will be. I recommend enforcing firms to have at least one woman in the board via laws, after a transition period of one year. The optimal ratio of men and women in the board forms an interesting topic for future research.

(36)

36

Appendix

Table 2: Simple regressions with relative.

Relative is a dummy variable that equals 1 when a founding family is active in the acquiring firm’s board and 0 otherwise.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

Relative 0.003 0.000 -0.038 -0.036

(1.32) (0.09) (-0.46) (-0.27)

N 4490 4490 4490 4490

adj. R-sq 0.000 0.000 0.000 0.000

t-statistics in parentheses: * p<0.10, ** p<0.05, *** p<0.01. Table 3: Simple regressions on gender.

Woman_board is a dummy variable that equals 1 if one or more women are active in the acquirers’ board and 0 otherwise.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year woman_board 0.002 0.001 0.199*** 0.309***

(1) (0.39) (3.23) (3.08)

N 4490 4490 4490 4490

adj. R-sq 0.000 0.000 0.002 0.001

t-statistics in parentheses: * p<0.10, ** p<0.05, *** p<0.01. Table 7: relative percentage regressions with controls.

% Relative represents the percentage of family members active in the acquiring firm’s board. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Total assets represent the natural logarithm value of the total assets of the acquirer. Book / market is the book-to-market ratio of a firm.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

% Relative 0.033 -0.033 -1.555 -1.930 Deal value 0.000 -0.002 -0.093*** -0.157*** Total Assets -0.002** -0.000 0.125*** 0.240*** Book/market 0.009* 0.007 1.112*** 1.708*** N 2211 2211 2211 2211 adj. R-sq 0.006 0 0.028 0.028 * p<0.10, ** p<0.05, *** p<0.01.

(37)

37 Table 8: relative and gender percentage regressions with controls.

% Relative represents the percentage of family members active in the acquiring firm’s board. % Woman represents the percentage of women active in the acquiring firm’s board. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Total assets represent the natural logarithm value of the total assets of the acquirer. Book / market represents the book-to-market ratio of a firm.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

% Relative 0.033 -0.034 -1.589 -1.996 % Woman -0.006 0.014 0.825* 1.598* Deal value 0.000 -0.002 -0.094*** -0.159*** Total Assets -0.002** -0.001 0.107*** 0.205*** Book/market 0.009 0.008 1.140*** 1.762*** N 2211 2211 2211 2211 adj. R-sq 0.005 0.001 0.029 0.029 * p<0.10, ** p<0.05, *** p<0.01.

Table 11: industry fixed effects and woman dummy.

Woman_board is a dummy variable that equals 1 when there is at least one woman on the board and 0 otherwise. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Book / market represents the book-to-market ratio of a firm. Industry 1 represents the agriculture, forestry and fishing Industry. Industry2 represents the mining industry. Industry3 represents the construction industry. Industry4 represents the manufacturing industry. Industry 5 represents the transport and energy industry. Industry 6 represents the wholesale industry. Industry 7 represents the retail industry. Industry 8 represents the services industry.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

woman_board 0.003 0.003 0.292*** 0.509*** Book/market 0.012*** 0.018 1.052*** 1.643*** Deal value -0.001 -0.002 -0.054** -0.075* Industry1 0.005 -0.095*** -0.747 -0.570 Industry2 -0.005 -0.047 -0.169 -0.225 Industry3 -0.017 -0.086** -1.107 -1.802 Industry4 -0.002 -0.037 -0.066 0.242 Industry5 -0.007 -0.039 -0.144 0.205 Industry6 -0.008 -0.052 0.016 0.377 Industry7 -0.002 -0.051 -0.067 0.353 Industry8 -0.002 -0.047 -0.175 0.094 N 2239 2239 2239 2239 adj. R-sq 0.001 0.002 0.027 0.027 * p<0.10, ** p<0.05, *** p<0.01.

(38)

38 Table 13: industry fixed effects and woman dummy.

Woman_board is a dummy variable that equals 1 when there is at least one woman on the board and 0 otherwise. Deal value is computed by taking the natural logarithm value of the takeover’s deal value. Total assets represent the natural logarithm value of the total assets of the acquirer. Book / market represents the book-to-market ratio of a firm. Industry 1 represents the agriculture, forestry and fishing Industry. Industry2 represents the mining industry. Industry3 represents the construction industry. Industry4 represents the manufacturing industry. Industry 5 represents the transport and energy industry. Industry 6 represents the wholesale industry. Industry 7 represents the retail industry. Industry 8 represents the services industry.

( 1 ) ( 2 ) ( 3 ) ( 4 )

announcement short term three year five year

woman_board 0.007** 0.005 0.176* 0.273* Deal value 0.000 -0.002 -0.094*** -0.155*** Total assets -0.003*** -0.001 0.105*** 0.208*** Book/market 0.011** 0.017 1.099*** 1.736*** Industry1 0.005 -0.095*** -0.747 -0.569 Industry2 -0.002 -0.046 -0.263 -0.412 Industry3 -0.017 -0.086** -1.125 -1.839 Industry4 0.001 -0.035 -0.168 0.040 Industry5 -0.003 -0.038 -0.285 -0.074 Industry6 -0.004 -0.050 -0.119 0.110 Industry7 -0.000 -0.050 -0.146 0.198 Industry8 -0.001 -0.047 -0.201 0.043 N 2239 2239 2239 2239 adj. R-sq 0.006 0.002 0.032 0.035 * p<0.10, ** p<0.05, *** p<0.01.

Table 17: CAR for different periods.

Variable Obs Mean

CAR Announcement 4,490 0.0017 CAR Short term 4,490 -0.0036 CAR Three year 4,490 -0.2478 CAR Five year 4,490 -0.4570

Table 18: Amount of takeovers per period.

Period # Takeovers 1986-1990 293 1991-1995 580 1996-2000 1417 2001-2005 1491 2006-2010 709

(39)

39 Table 19: Amount of takeovers per industry

Industry Takeovers Agriculture (1) 14 Mining (2) 267 Construction (3) 52 Manufacturing (4) 2841 Transport Energy (5) 316 Wholesale (6) 145 Retail (7) 154 Services (8) 681

Referenties

GERELATEERDE DOCUMENTEN

Om tot het gewenste rijgedrag te komen is het vervolgens niet alleen belangrijk dat wegen van elkaar te onderscheiden zijn (waarvoor het wegontwerp binnen wegtypen uniform moet

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers).. Please check the document version of

Bruijn, de, E. Determination of the underivatized antineoplastic drugs cyclophosphamide and 5-fluorouracil and some of their metabolites by capillary gas chromatography combined

Er kon worden vastgesteld dat de bodem tot op vrij grote diepte uit opgevoerd materiaal

Longmans, Parallel Series Parallel with &#34;Longmans' Leesboek voor Verenigd Zuid - Afrika&#34;. Longmans' Union South African

Provincies zullen deze ‘reactieve’ inzet van instrumenten onder de oude WRO opvangen met inzet van achtereenvol - gens nieuwe instrumenten: eerst overleg, bij geen opvolging van

This research offers a more complex view of the relationship between the observable (i.e. age, gender, nationality) and underlying (i.e. functional background,

Board independence equals the number of outside directors over the number of total board members, board size equals the total number of board members, gender diversity equals the