Women on
European Boards:
Director Gender and the M&A Game
Nga Nguyen
5989523
University of Amsterdam
Supervisor: Dr. J. Ligterink
“Girls can wear jeans and cut their hair short and wear shirts and boots because it’s okay to be a boy; for girls it’s like promotion. But for a boy to look like a girl is degrading, according to you, because
secretly you believe that being a girl is degrading.”
Statement of originality
Statement of Originality
This document is written by Student Nga Nguyen 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.
Acknowledgement
At the end of this long journey, I would like to thank everyone who have made this thesis possible and to all who have made this an unforgettable experience for me. First of all, I acknowledge gratitude to my supervisor Dr. Jeroen Ligterink at the Business University of Amsterdam for his patient and advise throughout the thesis process. I thank him for the thoughtful guidance, critical comments and corrections enabling me to deliver an academic paper.
I cannot finish without thanking my family and my boyfriend Dennis, without their support and encouragement, I could not have finished this work. I sense a profound gratitude and appreciation for their continuous feedback, taking the time to proofread my thesis and providing my with constructive comments.
Contents
1 Introduction 4
2 Literature review 7
2.1 Gender behavioral differences . . . 7
2.1.1 Overconfidence . . . 8
2.2 Gender diversity in corporate boards . . . 9
2.2.1 Agency theory . . . 9
2.2.2 Resource dependence theory . . . 10
2.2.3 Human capital theory . . . 11
2.2.4 Social impact theory . . . 11
2.2.5 Board gender diversity and the role of firm’s cash holding . . . 11
2.3 Empirical literature . . . 12
2.4 Contribution . . . 13
3 Methodology 14 4 Data and summary statistics 18 4.1 The bid initiation sample . . . 18
4.1.1 Independent variables: the extent of female directors . . . 18
4.1.2 The instrument: male director connectedness . . . 18
4.1.3 Control variables . . . 19
4.1.4 Cash reserves variables . . . 20
4.2 Descriptive statistics . . . 20
5 Results 23 5.1 First part of the research . . . 23
5.2 Second part of the research . . . 30
6 Conclusions 32
7 Bibliography 34
1
Introduction
According to the World Economic Forum, gender inequality has been identified in 2015 as one of the key challenges of our time. Although gender equality, particularly in the work force, has been on the rise over the past forty years, there is still plenty of room for improvement (World Bank, 2014). For example, in Table I we see that the number of female directors has been steadily rising, although very modest. In Grant Thornton’s newly released International Business Report, 6.600 private companies in 45 countries were surveyed to determine where the most women are in senior management positions. Results point to women covering less than 25% of executive positions globally. The evidence is more striking for the Netherlands, which ranks in the bottom with a mere 10% of female executives in middle management, dropping 1% compared to last year. European firms are under increasing pressure to appoint directors with different backgrounds and bases of expertise, and while it is true that gender equality in managerial and professional occupations is picking up pace, deep challenges remain when it comes to gender diversity in the boardroom (European Commission, 2014). Yet, in this respect, progress is being made, with traditional views on the composition of board of directors changing. Board directors have long been regarded as a homogeneous group of elites having similar socio-economic backgrounds, similar degrees, and, as a result, having similar views about business practices (Grant Thornton, 2014). However, a shift in beliefs of traditional boards is taking place, as in today’s world, organizations are facing more uncertain, complex and volatile market environments. Therefore, firms are increasingly recognizing the importance of attaining the highest standards of corporate governance, with a diverse board of directors acting as an effective internal governance mechanism. (Hermalin & Weisbach, 2003)
The importance of diversity in the board room has been widely acknowledged, not the least by the European Committee, whom in 2012 proposed a governance reform with the aim of attaining 40% objective of female directors in corporate boards in publicly listed firms. The idea behind this legislative initiative is that more women on boards could positively impact corporate governance and ultimately firm performance. This effect of board diversity, in particular director gender and its impact on firm performance, has been extensively researched by academics (Carter et al., 2010; Adams & Ferreira, 2004; Adams & Ferreira, 2009; Farrell & Hersch, 2005; Terjesen & Singh, 2008; Ahern & Dittmar, 2012). However, limited research has been performed on board gender diversity in the M&A domain. Prior literature seems to suggests that in many cases there is few to little wealth creation for bidding firms, evidenced by negative abnormal returns on bidder shares (Servaes, 1991; Sudarsanam & Mahate, 2003; Sudarsanam and Mahate, 2006). If M&A deals on average do not enhance firm value creation, or worse destroy firm value, shareholders are better of without the transaction. Moreover, the accumulation of large cash balances is an invitation for firms to engage in wealth-destroying acquisitions. As Jensen (1986) argues, the presence of excess cash has the potential to go hand in hand with agency problems, as managers may have incentives to grow their corporate empires at the expense of shareholders. The agency cost of excess cash increases in the extent of information asymmetry between managers and the external environment, with a subset of information asymmetry being the overconfidence effect (Malmendier & Tate, 2008). Theories related to gender risk preferences and gender behavioral characteristics suggest that gender may influence the M&A process, as women are hypothesized to be less risk prone and to exhibit less overconfidence in their decision-making (Croson & Gneezy, 2009; Levi et al., 2013). Following in the same line of reasoning, the inclusion of more females on corporate boards could lead to less M&A activity and better M&A deals. This hypothesis lays at the heart of this research, and provides a rationale for examining how the presence of female directors is associated with M&A performance and
the potential role of excess cash holdings in bidding activity. The primary goal of this study is to asses whether adding more female directors to the corporate board has any influence on the decision of the firm to undertake acquisitions and if the results remain robust to the presence of excess cash. This thesis will draw on findings from social psychological, behavioral finance and cash management literature showing female directors to be less overconfident in their decision-making than male directors. Thus, the following research question is formulated:
Does board gender diversity and excess cash holdings in European firms matter for
M&A activity?
The first part of this research will examine the relation between director gender and M&A activity. Psychological and experimental economics literature assert that the set of beliefs and risk preferences of women compared to men may differ when making certain choices, especially when it comes to undertaking difficult tasks or tasks that involve a sheer amount of risk. Evidence from empirical research find that women, generally, are associated with more risk-aversion in their decision-making than men, particularly when faced with complex tasks. As a result, it can be hypothesized that female directors are less inclined in undertaking M&A deals, which is characterized by intransparent and complex information.
Firms that hoard cash are more likely to deploy their resources to bad acquisitions. The relation between excess cash and M&A activity becomes even stronger when managerial overconfidence comes into play. Managers that are overconfident overestimate the returns to their investments, and are therefore less likely to resort to external funds. Malmendier & Tate (2006) find empirical evidence that when CEO’s fall into the overconfidence trap, they perform more ‘value-destroying’ M&A deals, and are more likely to engage in M&A activities when excess cash is available to fund the investments. While both men and women are prone to overconfident behavior, men are generally more likely to do so than women (Fellner & Maciejovsky, 2007; Niederle & Vesterlund, 2007). Therefore, if director gender matters for the amount of acquisitions made by a firm, then how does the presence of excess cash influence the merger decision? In the second part of this research we attempt to find an answer to this question by analyzing the impact of excess cash on director gender and merger frequency.
This thesis will answer the research question by an in-depth literature review on board gender diversity in chapter 2, detailed explanation of methodology and sample data in chapter 3 and 4, respectively, analysis of results in chapter 5, and concluding with conclusion and discussion in chapter 6.
ble
I
vi ew fe m ale d ir ec to rs o n c or po rat e bo ar ds fo r 2 0 Eu ro pe an c ou ntr ie s du ri ng th e pe riod 2 00 0 to 20 12 . F ig ur es b elo w ar e av er ag e p er ce nt ag e f em ale d ir ec to rs o n por at e boa rds a nd h av e be en ca lc ul ate d ov er su bs am pl es fr om th e f ul l s amp le (N =2 8, 36 5) e xt ra cte d f rom th e Boa rdE x da ta ba se . * Oth er s i s c omp ile d of th e f ol low in g pe an c ou nt ri es : B ul ga ri a, C ro at ia, C yp ru s, C ze ch R ep ub lic , G ib ralt ar , H un gar y, Ic elan d, Is le o f M an , Je rs ey an d Ro m an ia.20
00
20
01
20
02
20
03
20
04
20
05
20
06
20
07
20
08
20
09
20
10
20
11
20
12
st
ria
1.
7%
1.
3%
2.
7%
1.
8%
1.
2%
1.
2%
1.
3%
3.
1%
5.
9%
7.
2%
8.
1%
3.
9%
4.
2%
lg
ium
n/
a
2.
7%
4.
5%
6.
9%
6.
0%
6.
7%
7.
0%
6.
4%
6.
0%
6.
5%
8.
5%
11
.1
%
10
.6
%
m
ar
k
8.
6%
6.
6%
14
.9
%
13
.1
%
12
.3
%
12
.8
%
12
.5
%
10
.8
%
12
.0
%
12
.2
%
14
.3
%
14
.9
%
11
.8
%
and
14
.1
%
13
.7
%
15
.6
%
14
.8
%
12
.4
%
13
.4
%
12
.9
%
14
.5
%
13
.2
%
16
.2
%
16
.3
%
17
.8
%
14
.3
%
e
4.
5%
5.
3%
5.
2%
5.
5%
6.
2%
7.
4%
6.
7%
6.
4%
6.
0%
6.
9%
8.
6%
10
.8
%
11
.1
%
m
an
y
2.
6%
3.
8%
4.
2%
4.
8%
5.
7%
5.
5%
4.
9%
5.
4%
6.
1%
5.
9%
5.
3%
5.
3%
5.
7%
e
1.
1%
3
.3
%
3.
2%
2.
6%
2.
2%
3.
9%
5.
3%
4.
0%
5.
9%
4.
6%
9.
5%
9.
6%
7.
8%
an
d
0.
4%
1.
6%
2.
4%
2.
2%
3.
7%
4.
1%
4.
1%
3.
7%
4.
4%
4.
6%
5.
7%
7.
1%
6.
7%
ly
3.
4%
2.
4%
2.
3%
1.
5%
1.
8%
2.
3%
2.
3%
3.
3%
4.
9%
5.
1%
5.
7%
4.
6%
3.
0%
em
bour
g
2.
0%
1.
4%
0.
5%
4.
2%
6.
0%
4.
8%
7.
3%
7.
2%
6.
0%
5.
3%
3.
3%
6.
2%
9.
2%
her
lan
ds
3.
0%
4.
7%
4.
8%
3.
4%
3.
3%
3.
6%
4.
0%
4.
7%
5.
8%
6.
8%
6.
5%
8.
4%
8.
0%
w
ay
8.
3%
12
.7
%
12
.5
%
19
.0
%
21
.1
%
24
.3
%
33
.7
%
40
.0
%
40
.8
%
37
.0
%
39
.1
%
40
.6
%
39
.3
%
la
nd
n/
a
n/
a
14
.3
%
3.
1%
3.
9%
9.
5%
12
.8
%
13
.0
%
8.
9%
8.
3%
7.
6%
11
.0
%
12
.3
%
tug
al
n/
a
n/
a
n/
a
n/
a
7.
3%
8.
1%
7.
0%
5.
3%
5.
3%
2.
9%
2.
5%
3.
9%
7.
1%
in
1.
3%
2.
2%
1.
7%
2.
6%
3.
3%
3.
7%
5.
0%
5.
6%
7.
6%
7.
8%
7.
7%
8.
4%
9.
2%
ed
en
7.
8%
7.
9%
8.
3%
15
.2
%
19
.7
%
18
.3
%
19
.8
%
18
.5
%
18
.5
%
17
.9
%
17
.9
%
19
.2
%
17
.2
%
itz
er
la
nd
5.
3%
5.
0%
2.
4%
1.
7%
1.
8%
1.
7%
3.
2%
3.
1%
4.
4%
5.
7%
5.
8%
5.
2%
4.
2%
rk
ey
n/
a
n/
a
n/
a
n/
a
7.
2%
6.
6%
2.
4%
13
.3
%
14
.2
%
6.
6%
5.
1%
12
.2
%
14
.3
%
te
d K
ing
do
m
7.
8%
7.
2%
6.
8%
7.
3%
6.
1%
7.
1%
6.
3%
7.
2%
6.
5%
7.
3%
8.
8%
9.
8%
13
.4
%
ers
*
8.
3%
1.
3%
2.
4%
4.
4%
2.
8%
4.
0%
5.
3%
5.
2%
5.
7%
5.
5%
7.
6%
7.
1%
8.
0%
ta
l (
fu
ll s
am
pl
e)
5.
1%
5.
5%
5.
7%
6.
7%
6.
8%
7.
3%
7.
5%
7.
9%
8.
0%
8.
3%
9.
3%
10
.1
%
11
.2
%
52
Literature review
There has been considerable research on the impact of gender diversity on business. Most studies
focus on the key question whether having more female executives in corporate leadership roles improve performance. Unfortunately, there is not one easy answer to this question, and while it is difficult to demonstrate definitive proof, many European firms already have witnessed high rates of change in female representation. For example, in Norway a 40% quota for female directors in listed firms was enforced in 2008, the first European country to apply such a quota. Since then, gender quota’s have been enforced in the Netherlands, Spain, Belgium, Germany, Italy and Iceland, although less strict. Despite this positive trend, women at the board level generally remain very low. There is a significant body of literature describing the impact of gender diversity in corporate boards on firm performance, generally, yet few have looked at the economic consequences of gender-diverse boards in mergers and acquisitions setting. Moreover, many studies on board gender diversity focus on the impact of decisions made by sub committees of the board of directors (Xie et al., 2003; Bilimoria and Piderit, 1994; Carter et al., 2010; Peterson & Philpot, 2007). On the contrary, this paper will look at a setting where board directors as a whole control the M&A process, and as such measures first order effects of decisions (Fama & Jensen, 1983a; Bugeja et al., 2012).
To understand why gender diversity matter for corporate performance, we first need to take a step back and ask ourselves whether gender-related differences exist in the decision-making process, and more importantly, to what extend the patterns of behavior carry over into a financial context. As our research interest lies within a M&A setting, an environment characterized by high levels of complexity and ambiguity, we specifically will look at gender differences in decision-making under risk and uncertainty. This section will start by elaborating on differences in general risk preferences between men and women. Following, the theoretical framework for gender diversity within the boardroom, and related to this, the role of cash holdings in M&A setting will be fleshed out. To conclude the literature research, a comparison and discussion of empirical results will be discussed.
2.1
Gender behavioral differences
Basic differences between men and women that make them behave in distinctive ways are noncontroversial. Scientist have extensively debated whether behavioral differences in gender can be attributed to social construct or biological imperative, i.e. the nature v.s. nurture debate. Biological theories outline sex differences, and differences in genes and hormones specifically, as the underlying basis for the behavioral differences between men and women. From this perspective, observed sex differences are understood as an evolved adaptation that benefited the reproduction and consequently the continuity of human species (Geary, 1998; Knight, 2002). For women, a low-risk behavior in parenting efforts yields a higher return, whilst for men a higher expected payoff is obtained by taking a high-risk behavior in reproduction efforts. In contrast, sociological and psychological theories stress the importance of sex-specific socialization in explaining the difference in behavior between men and women. Proponents of these theories emphasize social and cultural causes, defining feminine and masculine characteristics, fundamental in explaining gender behavioral differences. Accordingly, these differences are more pronounced in some societies than others (Butler, 1999; Twenge, 1997; Rose, 1982; Levy, 2008). For this research, however, we will not question the fundamental differences in behavior between men and women. We take a neutral stance in the nature-nurture debate, and do not ascribe biological or social causes to gender-related differences
in decision making. That is, we are interested in the how and not why men and women take different decisions.
Particularly, gender differences in decision-marking under risk and uncertainty has been of interest among researchers (Powell and Ansic, 1997; Loewenstein et al., 2001; Daruvala, 2007; Meier-Pesti & Penz,
2008). Many of the decisions people make involve risk1, and gender differences in decision making may be
the result of differences in underlying risk preferences. Croson and Gneezy (2009) reviewed experimental economics literature examining gender differences in risk preferences using objective probability lotteries and found that men are more risk-prone than women in a vast majority of environments and tasks. For example, Eckel and Grossman (2002) conducted a laboratory experiment in which subjects participated in five alternative gambles with substantial financial stakes. They found that women are consistently more risk averse, on average, than men.
2.1.1 Overconfidence
In modern economic models one of the assumptions made is that humans behave rationally. However, in reality people do not always make rational decisions, and their deviations from rationality are often systematic (Barber and Odeon, 2001). Behavioral Finance relaxes this assumption by incorporating observable human behavior deviating from rationality into financial models, including overconfidence. Confidence is a double edged sword, motivating people to undertake risky and challenging endeavors, but at the same time can lead to so-called overconfident behavior. Croson and Gneezy (2009) point out that, although both men and women can be overconfident in their decision-making, it is men who in general are more overconfident in their success than women. Nevertheless, gender differences in overconfident
behavior seems to be highly task dependent (Lundeberg, Fox, and Pun¸cochar, 1994). Psychological
research find evidence that men are more prone to overconfidence than women in male-dominated environments such as finance. Barber and Odeon (2001) elaborate on specific conditions, as overconfidence is greatest for difficult tasks, for forecast with low predictability, and for endeavors that require clear and frequent feedback. Undertaking a merger or acquisition is a complex task, predictability of deal value is low, and feedback is very noisy. Thus, M&A activity is the type of task for which people are often overconfident. Unfortunately, there is no universal definition for overconfidence. For this research we will follow Croson and Gneezy (2009) and Levi et al. (2014), whom refer to overconfidence as the precision of the estimation of knowledge. More specifically, this implies that analysts overestimate the precision of his or her knowledge about the expected value of a merger or acquisition. Consequently, overconfident people overestimate the probability that their assessments of the deal’s value are more accurate than the assessments of others. Malmendier & Tate (2008) demonstrate that overconfident CEO’s, measured as personal over-investment in their company and press portrayal, overestimate their ability to generate returns. They find that overconfident CEO’s overpay for target firms and engage in mergers which are value-destroying. Barber and Odeon (2001) find evidence for overconfidence as an explanation for high levels of trading in the financial market. Their findings document that men trade 45% more than women, in line with theories predicting male investors to trade more excessively than their female counterparts.
1Throughout this paper risk and uncertainty is used interchangeably. As under uncertainty one still can assign subjective
probabilities to outcomes, Knight’s (1921) distinction between risk, in which one knows the probabilities, and uncertainty, in which this randomness cannot be expressed, will not be used.
2.2
Gender diversity in corporate boards
For most organizations putting a board of directors in place is often one of the legal requirements. But why do they exist? According to Fama & Jensen (1983a), the board of directors, as representatives of shareholders, play a key role within the corporate governance framework of large corporate entities. Their principal role as directors is to hire, fire and compensate senior management, and to assess the overall direction and strategy of the business. Carter et al. (2010), on the other hand, believe board directors fulfill at least four important functions: monitoring manager, counsel management, monitoring compliance and linking the organization to the external environment. Although there is no clear-cut definition, all theories point out that board directors are the highest control mechanism in a firm. Consequently, the existence of a corporate board can potentially enhance the corporate governance structure, and ultimately overall economic performance. However, corporate boards do not function properly when they do not have the ‘right people’ in place. This has motivated a lot of research on this topic. Recent literature has focused on whether certain board characteristics matter to the effectiveness of boards and their decision-making (Hermalin and Weisbach, 1998).
For this research, the main interest lies within the effect of board gender diversity on corporate policies. Why is gender diversity important within the board of directors? Before we can dive deeper into this matter, it is needed to explain what role is assigned to corporate boards in organizational context. The board of directors are believed to fulfill at least four important functions: monitoring managers, counsel management, monitoring compliance and linking the corporation to the external environment (Cartel et al., 2010). Basically, top management has to justify all it’s decisions, including the decision to undertake acquisitions, to the board of directors and is therefore seen as the highest control mechanism in a firm. Hermalin and Weisbach (2003) point out that empirical literature on corporate boards has focused on how certain board characteristics such as composition influence the effectiveness of board decision-making and firm performance. For example, a number of papers have found correlations between firm performance and outside directors (Hermalin and Weisbach, 1991; Mehran, 1995; Klein, 1998; Bhagat and Black, 2000), director ethnicity (Carter et al., 2003) or size of the board (Yermack, 1996; Eisenberg et al., 1998; Wu, 2000). Another subset of board composition is director gender and its relationship with firm performance is of growing interest among researchers. Although there is no single theory that directly predicts the nature of the relationship between board diversity and firm performance, theories from several disciplines provide valuable insight into this matter. This thesis will adopt an interdisciplinary approach and draws from theories taken from behavioral finance, corporate governance, sociology and social psychology to provide the basis of the theoretical framework for the hypotheses tested. The first hypothesis is postulated as following:
H1: Ceteris paribus, the extent of female directors is negatively associated with the propensity to initiate
acquisition bids.
2.2.1 Agency theory
The agency theory describes the relation between the principal and the agent. There are conflicting interests between shareholders and managers of the firm. Boards monitor and discipline management as a means to resolve the conflicts and align the interest of managers with that of shareholders, which is a fundamental concept of agency theory (Jensen & Meckling, 1976). An implication of the agency theory is that diverse boards are better monitors of managers, because board diversity increases board independence. When board directors act independently, they are less likely to be captured by the CEO
or top management. However, the agency theory does not provide a clear prediction of this link. That is because other factors, such as ownership concentration, may have a more powerful influence on board monitoring than board independence. Carter et al. (2003) examine the percentage of female directors on Fortune 100 corporate boards and find evidence that female directors are associated with outside directors and consequently lead to higher firm performance. Jensen (1993) on the other hand finds that the composition of financial compensation of directors is a more important factor in increasing the willingness of directors to monitor management than board independence. Jensen (1993) argues that when directors have relatively high equity ownership, they have more incentives to monitor since pay for performance increases. Thus, the agency theory does not provide strong support for board diversity as a means to improve financial performance. However, it does also not rule out the possibility that board heterogeneity can be advantageous for firms to enhance their performance.
2.2.2 Resource dependence theory
In contrary to the agency theory, the resource dependence perspective is better able to pinpoint the financial benefits of board diversity. The resource dependence theory approach view firms as open systems which are continuously exchanging and acquiring certain resources in order to survive, establishing a dependency between the firm and external entities. Within the corporate governance literature, the resource dependency approach present corporate boards as a means to provide human capital and social capital that are essential to the firm. While organizations depend on external resources for their survival, Pfeffer and Salancik (1978) highlight the importance of the board of directors whom are able to link the organization to its external environment to gain important resources. The authors suggest that there are four benefits for the external board linkages: provision of resources, provision of commitments of important external organizations, communication channels that would otherwise not have been installed, and legitimacy in the external environment. Hillman et al. (2000) translate these four benefits to director types and state that different types of directors provide various resources to the firm: business experts, insiders, community influentials etc. Recent scholars have extended the resource dependence theory suggesting that different types of directors provide different kind of resources that are beneficial to the firm. As a result, board gender diversity increases the breadth of resources a firm can tap on, and hence will likely lead to better firm performance. Does the type of diversity matter? According to empirical literature this seems to appear an important factor. For example, Booth and Deli (1999) find there exists a positive relation between commercial bankers on corporate boards and the total debt of a corporation. Their finding is consistent with the resource dependence view, as commercial bankers are linked to the debt market and as such can provide valuable expertise on capital structure and financing. Agrawal and Knoeber (2001) examine outside directors that have political and legal backgrounds. They find that these type of directors are more likely to have a seat on a board of a firm that sells to the government or face government regulation. In the same line of reasoning, one can argue that the inclusion of more females to the board of directors increases the valuable benefits and resources the firm can capitalize on. Resource dependence theory provides the basis for some of the most convincing arguments for firms to pursue diversity in their boards (Carter et al, 2010). A diverse board improves the quality of information that is provided to the management to aid in their decision-making. In addition, diverse boards bring diverse perspectives and nontraditional solutions to problems. According to Carter et al (2010) this is because diverse directors are less likely to be inside directors or business experts. Moreover, board diversity sends a positive signal to the labor market. As a result, if a firm incorporates board diversity into their policy it is more likely that they have more access to talent.
2.2.3 Human capital theory
According Terjesen et al. (2009) the human capital theory is derived from the paper by Becker (1964). Becker (1964) addresses a person’s stock of competencies, education, skills and experience as attributes embodied in the ability to perform labor and thus benefit an organization. Terjesen et al. (2009) moreover state that gender differences in corporate boards results in directors having unique human capital. In essence, human capital theory complements some concepts from the resource dependence theory. If we believe that the gender of directors influence human capital, then a reasonable hypothesis is that human capital increases in board diversity. In this respect, it is often claimed that women lack the ’right’ human capital to fulfill a position in a board (Terjesen et al, 2005). Evidence pointed out by Terjesen et al (2009) reveals that women are just as qualified as men for directorships, the crux is hidden in the fact that women tend to have less relevant working experience as business experts. Thus, female directors are expected to bring less human capital to corporate boards. Concluding, human capital theory predicts that board diversity negatively affects the performance of a board and subsequently firm performance.
2.2.4 Social impact theory
The presence of minorities on boards are often perceived as favorable by corporate stakeholders. However, the academic literature of social psychology takes the opposing view, which is more pessimistic about the extent to which minorities within corporate boards can effectively influence group decisions. Westphal and Milton (200) suggest that a central finding in social psychology literature is that there is a inverse
relationship between demographic differences and social cohesion within and between groups. The
implication here is that social barriers prevent the minority group from influencing group decisions. According to Westphal and Milton, this concept is derived from the social impact theory, which state that individuals who have majority status have the potential to exert the highest social impact on decisions made by the group. Thus, following the social impact theory, we can hypothesize that board diversity will not or negatively affect board performance as a result of internal group dynamics within the board.
2.2.5 Board gender diversity and the role of firm’s cash holding
Roll (1986) was the first to explain failed mergers by linking takeover contest to the winner’s curse. His proposed ’hubris’ hypothesis on corporate takeovers implies that acquiring firms simply pay too much for estimated synergies in a takeover due to managerial optimism, overconfidence or competitive bidding. However, Malmendier & Tate (2008) argue that, in this regard, the implications of overconfidence for mergers are more subtle than mere overbidding. Malmendier & Tate (2005a) find evidence for corporate investment decisions made by overconfident CEO’s to be related to internal resources. Their findings suggest that rational CEO’s are less likely than overconfident CEO’s to finance mergers with (excess) cash. The underlying rationale stems from the fact that external financing is more expensive than internal financing and as a result, along the lines of the pecking theory postulated by Myers and Majluf (1984), investors prefer seeking internal funds rather than external funds for investments, and are more likely to raise debt over equity. This is consistent with Harford’s (1999) findings that cash-rich firms are more likely to engage in acquisitions than cash-poor firms. If investors are more likely to undertake acquisitions when excess cash is available, then this effect is expected to be more pronounced for investors exhibiting overconfident behavior. As overconfident investors overestimate their own returns compared to rational investors, overconfident investors are more reluctant to raise external capital. They believe
that the firm value is undervalued by the financial market and the cost of external financing is overpriced. As a result, overconfident investors are more inclined to undertake mergers only when there is abundant internal capital available. Thus, overconfidence is a potential explanation for excessive merger frequency exacerbated by the availability of internal resources. Only in firms with large cash holdings or sufficient riskless debt capacity are overconfident managers more likely to undertake mergers. Although the CEO is responsible for the day-to-day running of the company, corporate boards retain control over big decisions, including M&A deals. As we hypothesize that male directors are more likely to overestimate the precision of their own knowledge, we expect male-dominated board of directors more frequently to perform M&A deals than corporate boards with female directors when internal resources are abundant. Thus, the second hypothesis is postulated as following:
H2: Ceteris paribus, conditional on the level of firm’s excess cash, the extent of female directors is
negatively associated with the propensity to initiate acquisition bids
2.3
Empirical literature
All the theories that have been elaborated in the above sections support the view that board diversity affects the financial performance of a corporation. The agency theory and resource dependence theory in particular are in favor of the hypothesis that a positive relation exists between the number of female directors and firm performance. Do these theories still hold when empirically tested? Empirical studies on the relation between board gender diversity and firm performance show mixed results. Smith, Smith and Verner (2006), Catalyst (2007) and Singh et al. (2001) find a positive relation between the fraction of female directors and firm performance, whereas Ahern and Dittmar (2012), Adams and Ferreira (2009) and Farrekk and Hersch (2005) find no such relation or even a negative relation. The inconsistencies in the empirical work on board diversity and firm performance may stem from the fact that these studies are plagued by the usual problem of joint endogeneity. As Hermalin and Weisbach (2008) point out there on the one hand a potential sorting of directors to firms and on the other hand firms preferences to choose certain types of directors.
What if we take a step back and look at the relationship between board gender diversity and M&A activities? M&A is a very important corporate decision that can make or break a firm. Although M&A performance is not a direct proxy for firm performance, given the amount of money spend on deals and the high uncertainty involved around the transaction, it is more than likely that it influences firm performance. As we have already cited, past research has shown that men and women behave differently in terms of how overconfident they are. This is especially true in events with great uncertainty (Levi et al., 2013). When looking at unknowns or when feedback is delayed or uncertain, women exhibit less overconfidence in their decision-making than men. In this regard, a M&A setting amplifies the differences between male and female directors and how they respond to uncertainty with overconfidence, and is therefore the ideal setting to study. Levi et al. (2013) examine whether female directors on US boards influence CEO empire building and the bid premium paid for target firms. Their results show that female directors make fewer and better acquisitions. They conclude that female directors create shareholder value through their influence on M&A activities. Dowling and Ali Aribi (2013) find similar results as Levi et al (2013) where they use a novel dataset covering large acquisitions by the FTSE 100 companies. They show that the presence, and proportion, of female directors leads to fewer acquisitions in the UK. Harford et al. (2008) explore the link between corporate governance and cash holdings positions of US firm. Their finding suggests that managers in poorly governed firms spend cash quickly on acquisitions,
rather than accumulate it. Malmendier and Tate (2005) find that overconfident managers are more likely to acquire firms, especially when there are abundant internal resources. The overconfidence of CEO leads them to overpay for target companies and undertake mergers and acquisitions that are value-destroying.
2.4
Contribution
This thesis will follow Levi et al. (2013) by examining the impact of female directors in European firms on M&A performance. In addition, this thesis is the first in the field on board gender diversity that will explore the potential role of cash holdings and the method of payment in the effect of female directors on M&A. In doing so, I will use the model of cash reserves by Harford et al. (2008) and the intuition by the paper of Malmendier and Tate (2006). Moreover, this thesis will incorporate evidence of the effects of gender on behavior in general and financial performance in particular. Furthermore, this study contributes in offering new insights in why and how M&A activities take place and the possible consequences of social and political pressure of mandating gender quotas for directors.
3
Methodology
The first part of this study will address the relationship between board gender diversity and firm’s bidder activity. The first hypothesis is postulated as follows:
H1: Ceteris paribus, the extent of female directors is negatively associated with the propensity to initiate
acquisition bids.
The use of panel data allows us to use firm-fixed and time-fixed effects to address the problem of omitted variables by accounting for unobserved heterogeneity that are time invariant and firm specific.However, there are still endogeneity concerns about the direction of causality between acquisition bids and the extent of female directors. Reversed causality may be of an issue because of potential sorting of male and female directors to firms based on M&A performance. Levi et al. (2013) argue that the choice for nominating women on the board might depend on the management style of the entrenched CEO. If the CEO prefers a ’quiet life’, it might favor female directors whom they believe will be less acquisitive. In contrast, CEOs engaging in empire-building might be less appealed to female board members exactly for the same reasons. The endogeneity concerns regarding the extent of female directors are shared by Hermalin and Weisbach (2003), and Adams and Ferreira (2009), of whom the latter believes that both director compensation and firm performance play a significant role in the incentives of women to join firms and the incentives of firms to hire female directors. The problem of reversed causality can be solved by estimating the models using the instrumental variable approach (IV). An instrument requires to fulfill the conditions of being correlated with the extent of female board directors, but uncorrelated with the number of acquisition bids. Model (1) is referred to as the first stage OLS regression estimating the fitted values of the fraction of female board directors by using male director connectedness as an instrument, various control variables, and year fixed effects and clustered standard errors for the panel regression.
(1) F EM ALESit= β0+ β1DIRECT ORCON N ECTit+ β2CON T ROLSit+ ci+ vt+ it
Following Adams and Ferreira (2006) and Levi et al. (2013) the instrument measure that will be used is the number of male directors that have board connections to female directors sitting on other corporate boards. They argue that one foremost reason why women are absent on boards is due to the fact that they lack the social connections. Medland (2004) extends on this notion by arguing that boards remain stale, pale and male because the social network linking board directors primarily consists of men. If male directors have more women in their social network, it is hypothesized that more female directors should be observed. Although the informal social connections between board directors cannot be observed, networks that occur because male directors have multiple seats on boards can be accurately measured. It is expected that this measure is positively correlated to board gender diversity. The larger the fraction of male director connectedness, the more likely female directors are observed in corporate boards.
Having done that, we will explore the effect of board gender diversity on a firm’s acquisitiveness in two ways: by a nonlinear regression model that classifies firms into two groups and a linear regression model that accounts for over-dispersed outcome variable. First, we use the Probit regression, which is specifically designed for binary dependent variables, forcing the predicted values to be between 0 and 1. The classification criteria is based on the number of bid initiations performed in a given year, with 0 being
firms absent of bidding activities, and 1 being firms with one or more bid initiations performed within a
fiscal year. Second, to account for the over-dispersion2of the dependent variable number of bid initiations
(see Table II), we use the Negative Binomial Regression that takes into account the over-dispersion in count outcome variables. Model specification (2a) and (2b) are referred to as the second stage of the IV regression, where the dependent variable will be regressed upon the independent variable and controls by a Probit regression and Negative Binomial regression, respectively.
(2a) Pr(BID IN IT IAT IONit= 1|X = x) = Φ(x0α + ci+ vt+ it)
(2b) LOG(BID IN IT IAT IONit) = δ0+ δ1F EM ALESit+ δ2CON T ROLSit+ ci+ vt+ it
Concerning model specification (2a): i is an index for firmid, t is an index for time, Pr denotes the probability, BID INITIATION ∈ (0,1) measures the acquisitiveness of a firm and is a binary outcome which is equal to 1 if the firm initiates one or more bids in a given year and 0 otherwise, Φ is the cumulative distribution function of the standard normal distribution, α is a vector of parameters consisting of x
coefficient vectors for constants, x coefficient vectors for F EM ALES and F EM ALES, which is thed
extent of female directors and the predicted TSLS estimators for female directors obtained from model specification (1), x coefficient vectors for CON T ROLS, which are control variables for firm-,
board-and country characteristics, ci is a firm-fixed effect and vtis a time-fixed effect. Hence, the probability
of initiating a merger bid, given certain values x for explanatory variables, is a function of these values multiplied by a set of these parameters.
Concerning model specification (2b): i is an index for firmid, t is an index for time, BID INITIATION
measures the number of acquisition bids initiated in a given year, δ0 is a constant, δ1 is a coefficient
vector for FEMALES, which is the extent of female directors and the the predicted TSLS estimator,
δ2 is a coefficient vector for CONTROLS, which are control variables for firm-, board- and country
characteristics, ci is a firm-fixed effect and vt is a time-fixed effect. See Appendix A for an overview of
variables used in this study and their corresponding definitions.
The second part of this study will look at the role of excess cash and its effect on the relationship between board gender diversity and M&A activity. The hypothesis that will be tested is the following:
H2a: Ceteris paribus, given the level of firm’s excess cash, the extent of female directors is negatively
associated with the propensity to initiate acquisition bids
The aforementioned sample will be replicated to test the second hypothesis, which means that the same set of firms will be used. The only difference is that the sample will be divided in two subsamples dependent on the level of excess cash held in a given year. Following Harford (1999), we will use a
cash management model to determine the level of excess cash3, which is defined as the residual from a
regression of cash holdings4on firm-specific variables, see model specification .
2Data becomes over-dispersed when the conditional variance exceeds the conditional mean.
3According to Harford (1999), excess cash is the unexplained portion of the firm’s cash holdings.
4Harford et al. (2008) view cash as a means to support working capital needs of a firm, which is closely related to its
sales. However, since obtained data on sales figures are often incomplete we will compute cash holdings as cash and cash equivalents to total assets
(3a) CASH HOLDIN Git= γ0+γ1T OBIN0SQi,t−1+γ2F IRM SIZEi,t−1+γ3N ET CF/T OT ALASSET Sit+
γ4CF V olatilityi+ γ54 N ET CF/T OT ALASSET Si,t+1+ γ64 N ET CF/T OT ALASSET Si,t+2+ it
Where i is an index for firmid, t is an index for time, CASH HOLDING is a firm’s cash reserve at
the beginning of the year, measured as cash and cash equivalents to total assets, γ0 is a constant, γ1 is
a proxy for Tobin’s Q, γ2 is a coefficient vector for the market value of the firm, measured by a firm’s
market capitalization defined as the number of outstanding stock multiplied by share price, γ3, γ5and γ6
are coefficient vectors for the level, the t to t + 1 and the t + 1 to t + 2 changes in net cash flow to assets,
defined as operating cash flow minus capital expenditures over total assets, γ4 is a coefficient vector for
cash flow volatility, measured by the coefficient variation of firm-specific operating cash flows and is the error term.
After defining the baseline cash holdings for each firm, the following step is to compute the subsample
of cash-rich firm years, in which a firm’s cash reserves is at least 1.5 standard deviation5above the value
predicted by the cash-management model. The standard deviation used is the firm-specific standard
deviation of cash reserves. Then, in turn, model specification (3b) will be regressed to obtain the
relevant results that allow us to test hypothesis 2. The variables in this model are similar to the ones explained in model specification 2, with the addition of the presence of firm’s excess cash.
(3b)Pr(BID IN IT IAT IONit= 1|X = x)|CASHRICH = 1
= Φ(x0α + c
i+ vt+ it)
ble
II
vie w me th odol og y a nd va ria bl es u se d i n p re viou s e mp ir ic al stu di es .Fa
rr
el
l a
nd
H
er
sc
h (
20
05
)
Ma
lm
en
di
er
an
d T
at
e
(2
00
8)
Car
ter
et
al
.
(2
00
7)
Ad
am
s a
nd
Fe
rr
ei
ra
(2
00
9)
Car
ter
et
al
.
(2
01
0)
Ah
er
n a
nd
Di
ttm
ar
(2
01
2)
Lev
i et
al.
(2
01
3)
ard
-lev
el
ch
ar
ac
ter
ist
ic
s
ber
o
f f
em
ale
ct
or
s
x
cen
tag
e o
f fe
m
ale
ct
or
s
x
x
x
x
x
le
d
ir
ect
or
in
di
ca
to
r
x
ar
d s
iz
e
x
x
x
x
x
x
cen
tag
e i
nd
ep
en
den
t
ct
or
s
x
x
x
b
ei
ng
ch
ai
rm
an
o
f
ar
d
x
x
x
-le
ve
l ch
ar
act
er
ist
ics
er
o
f b
id
in
itia
tio
ns
x
bin
’s Q
x
x
x
x
x
x
x
x
x
ok
lev
er
ag
e
x
x
h
ol
di
ng
s
x
x
ket
cap
ital
iz
at
io
n (
ln
)
x
x
x
e f
ix
ed
eff
ec
ts
x
x
x
x
x
x
fi
xed
eff
ec
ts
x
x
x
x
x
x
x
tr
um
ent
al
va
ri
abl
es
x
x
x
x
164
Data and summary statistics
4.1
The bid initiation sample
The dataset is comprised of the largest listed firms from 20 European countries covering the period 2000 to 2012. Board characteristics are collected from the BoardEx database. The BoardEx database provides profiles of private and public firms executives and board directors globally. Detailed information on the personnel level of board directors contain information such as board size, the number of outside directors, and the director’s gender. BoardEx does not provide standardized information on whether the CEO holds a position as chairman of the board of directors. Yet, from the director role description we were able to compute CEO’s role in the corporate board by means of an algorithm that is able to identify all directors that have role descriptions containing the words ’CEO and Chairman’ or ’President and Chairman’ and other variations. All observations from the BoardEx database are uniquely identified by their companyid. Sadly, this identifier does not match with identifiers employed by other databases, such as isin, sedol, cusip etc. Therefore, company’s ISIN together with the company’s ticker symbol are used to match data from other databases with the observations from the BoardEx sample. Next, data requirements on deal characteristics are retrieved from Thomson ONE Banker. The M&A module in Thomson ONE Banker contains information on over 400,000 deals worldwide. From the M&A League Table all acquisition bids are filtered by sample period (2000-2012), region (Europe) and company type (listed firms), including only bids where the firm’s toehold before the deal announcement was less than 50% and the sum of the toehold and the percentage of shares sought in the bid is greater than 50%. We obtained the 6-digit sedol and ticker symbol provided by Thomson ONE Banker and used Datastream to convert the identifier to isin in order to match the data on bid initiations with board characteristics from the BoardEx sample. Having done that, firm-level data are retrieved from Worldscope through Datastream. Lastly, to control for country differences the annual percentage of female labor participation for each country is retrieved from the World Bank database. The definitions of variables used for this research and their sources are found in Appendix A.
4.1.1 Independent variables: the extent of female directors
For this research we use three measures for board gender diversity: the number of female directors, the percentage of female directors and an indicator variable that equals one when a corporate board has one or more female directors and zero otherwise. These measures are extensively used in other empirical studies on board gender diversity (see Table I), thus testing them all together increases the robustness of our results. Data regarding female directors are retrieved from BoardEx, which provides available data concerning the gender of board directors.
4.1.2 The instrument: male director connectedness
The main problem with studies on gender diversity, as evidently pointed out by Hermalin and Weisbach (2003), is that the extent of female directors on corporate boards and takeover activity can be joint-endogenously determined, which has implications for testing hypotheses and interpreting empirical results. Nonetheless, using instruments that correlate with female directors and that are uncorrelated to acquisition bids, except through variables that are controlled for, is a way to circumvent the econometric problem of endogeneity. For this study the instrument that will be used is the number of male directors
that are linked to female directors sitting on other corporate boards. The rationale behind this instrument is that females often lack the social connections giving them access to directorship positions (Medland, 2004). The recruitment of a board director position often happens through an informal process, in which potential candidates are drawn from an informal social network that consists primarily out of men. As a result, the ‘old boys club’ may exclude ‘potential’ female talent from entering the board of directors, which explains why there are still so few female directors seen in corporate boards globally. This finding is confirmed by the pioneering work of Granovetter (1973), who highlights the importance of informal social networks for career advancement. Going forward, a person who sits on a corporate board may sit on several other corporate boards, or may have been a board director there at a previous time, with the director building its own network at each firm it is being employed. Because this elite network structure of directors is of particular value to the recruitment of directors, we can argue that the extent of females on corporate boards significantly increases with the social connections of directors sitting on other boards with one or more female directors. Available data on male director connectedness is obtained from the BoardEx database and is computed as the percentage of directors that are linked to female directors sitting on other corporate boards. With the use of this instrument, we follow Adams and Ferreira (2009) and Levi et al. (2013).
4.1.3 Control variables
To avoid omitted variables bias we control for variables relating to firm-specific and country-specific characteristics (see Appendix A for an overview). The first set of controls used in this study are linked to board characteristics and have been obtained from BoardEx. These controls consist of the following: board size, percentage of independent directors and CEO holding a chairman position in the board of directors. Following Hermalin and Weisbach (2003), whom argue that boardsize, measured by the number of members on the board directors, appears to affect firm performance. They suggest that large boards are less effective than small boards because of the increase in agency problems. When boards become too big, they become less part of the management process and suffer more from director free-riding, resulting in less monitoring and disciplining board (Jensen, 1993; Lorsch, 1992). Fraction of independent directors is calculated as the number of non-executive directors divided by board size. As evidenced by Adams and Ferreira (2009), independent directors is associated with improved governance, and consequently less M&A activity. Therefore, a negative relationship is anticipated between independent directors and the number of bid announcements made. CEOs who also serve as chairs are expected to exert more influence over directors, and thereby increase their ability to affect the M&A process (Shivdasani and Yermack, 1999). The second set of control variables are related to firm-level characteristics and are drawn from the Worldscope database, consisting of two firm performance measures (tobin’s q and roa), book leverage, cash holdings and a proxy for firm size (market capitalization). Lastly, we control for differences between countries measured by the percentage of female labor participation within European countries. Views on the division of labor in households and gender equality on the labor market may differ from country to country. Wright and Rogers (2010) present evidence of how the increase in female labor force participation is accompanied by a significant change in the economic opportunities of women in terms of the occupations they fill. When more females actively participate on the labor market, more females acquire experience and skills necessary to fulfill top corporate positions and consequently gain accession to the board of directors. Therefore, the percentage of female labor participation is expected to exhibit a positive relationship with the extent of female directors.
4.1.4 Cash reserves variables
Firm characteristics that drive the value of cash reserves are often considered to be related to a firm’s operating cash flows, cash flow volatility, and the degree of information asymmetry between managers
and external capital providers. In accordance with Harford’s cash-management model, we use the
market-to-book ratio of equity as a proxy for the information asymmetry faced by managers in a given industry. Future cash-flows net-of-investment outlays are a proxy for operating cash flows over time, and is calculated as operating cash flows minus capital expenditures. Cash flow volatility used is the time series standard deviation of the firm’s operating cash flows. Furthermore, market capitalization will be used as a proxy for firm size.
4.2
Descriptive statistics
Table III shows the descriptive statistics of all relevant variables used in this research. It reveals that all board, firm and country characteristics have 12,226 observations. What is striking is that on average firms have only 0.330 female board directors, which is remarkably low. The same holds for the average percentage of female board directors and the average percentage of firms having one or more female board directors, which is 8% and 23.6% respectively. The average board size is equal to 7.9, the average percentage of independent directors in a given firm is equal to 75.6%, and the average percentage of independent CEO’s equals 23.8%. Lastly, the average female labor participation equals 53.4%.
Table IV exhibits the pair wise correlation between all variables. The bid initiation indicator is positively correlated to the number of female directors and the female director indicator, being only significant in the latter, while the bid initiation indicator is negatively but insignificantly correlated with the percentage of female directors. Kennedy (2008) explains that correlation coefficients above 0.7 may be an indication of multicollinearity in the regression analysis. Following this reasoning, examination of the table reveals little problem of multicollinearity, apart from the three measures for the extent of female board directors. However, univariate correlations can hide the true relations between the variables when there is omitted variable bias present (Levi et al., 2013). The next step is to employ multiple regressions to present a more accurate picture of the variables associated with corporate acquisitions.
Table III
Descriptive statistics of the sample for European board, firm and country characteristics. The sample consists of 12,226 firm-year observations during the period 2000 to 2012. Available data on board characteristics are retrieved from BoardEx, which have been matched with deal characteristics from Thomson One Banker and firm characteristics from Datastream and/or WorldScope. Lastly, data on female labor participation per country has been retrieved from the World Bank database. All firm-level variables are measured at the end of the fiscal year prior to the date of the bid announcement. See Appendix I for more details on variables definitions and corresponding sources.
Panel A: descriptive statistics
N
Mean Std. Dev.
Min
Max
Board-level characteristics
Number of female directors
12,226
0.330
0.701
0
6
Percentage of female
directors
12,226
0.080
0.192
0
1
Female director indicator
12,226
0.236
0.424
0
1
Board size
12,226
7.948
3.612
1
19
Independent directors (%)
12,226
0.756
0.325
0
1
CEO being chairman of board
12,226
0.238
0.426
0
1
Firm-level characteristics
Number of bid initiations
12,226
0.665
1.273
0
20
Tobin’s Q
12,226
1.899
1.850
0.191
6.256
ROA
12,226
0.012
0.874
-1
1.321
Book leverage
12,226
0.503
0.460
0.001
0.956
Cash holdings
12,226
0.151
0.164
0
0.999
Market capitalization (ln)
12,226
13.184
2.322
5.170
20.986
Country-level characteristics
ble
IV
w is e c or re la tion ma tr ix . S up er sc rip ts ** a nd * r ep re se nt th e s ta tis tic al sig nif ic an ce a t th e 1 % a nd 5 % le ve l, re sp ec tiv el y. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) N umb er of le dir ec to rs 1. 000 P er cen ta ge o f le dir ec to rs 0. 694 ** 1. 000 In dic ator le dir ec to r 0. 847 ** 0. 747 ** 1. 000 al e d ir ec to r ec ted nes s 0. 264 ** 0. 110 ** 0. 306 ** 1. 000 B oa rd siz e 0. 430 ** 0. 022 * 0. 369 ** 0. 215 ** 1. 000 P er cen ta ge en den t d ir ec to rs 0. 121 ** 0. 084 ** 0. 106 ** 0. 100 ** 0. 049 ** 1. 000 C EO b ei ng ma n of b oa rd 0. 027 ** ‒0. 07 1 ** 0. 040 * 0. 021 ** 0. 297 ** ‒0. 36 1 ** 1. 000 In dic ator itia tion 0. 009 ‒0. 01 6 0. 007 ** 0. 064 ** 0. 050 0. 024 ‒0. 00 4 1. 000 T ob in’s Q ‒0. 01 0 ‒0. 00 6 ‒0. 01 0 ‒0. 00 6 ‒0. 01 7 ‒0. 00 8 ‒0. 00 5 ‒0. 00 6 1. 000 R OA 0. 003 ‒0. 00 2 ‒0. 00 3 0. 012 0. 018 * ‒0. 00 4 ‒0. 01 8 0. 025 ** ‒0. 10 2 ** 1. 000 B oo k le ve rag e 0. 034 ** 0. 018 * 0. 040 ** 0. 017 0. 044 ** 0. 009 0. 020 * 0. 018 0. 091 ** ‒0. 58 4 ** 1. 000 C as h h ol di ngs ‒0. 05 6 ** ‒0. 03 5 ** ‒0. 05 8 ** ‒0. 04 0 ** ‒0. 09 5 ** 0. 005 0. 020 * ‒0. 08 4 ** 0. 074 ** ‒0. 02 5 ** ‒0. 15 7 ** 1. 000 M ar ket liz ation (l n) 0. 058 ** 0. 037 ** 0. 071 ** 0. 050 ** 0. 081 ** 0. 039 ** ‒0. 02 4 ** 0. 088 ** 0. 005 0. 017 ** ‒0. 02 9 ‒0. 04 2 ** 1. 000 Fem ale lab or ip atio n 0. 166 ** 0. 141 ** 0. 140 ** 0. 082 ** ‒0. 10 5 ** 0. 126 ** ‒0. 04 3 ** 0. 027 0. 010 ‒0. 03 0 ** ‒0. 01 3 0. 063 ** ‒0. 00 8 1. 000 215
Results
This study consists of two parts, both related to board gender diversity within a M&A framework. The first part looks at the relation between director gender and acquisition bids in Europe. To test this relationship we first perform a probit model that estimates the effect of female directors on the
likelihood of a firm to initiate a bid acquisition6. The problem with board composition, and director
gender specifically, is that there are potential endogeneity concerns present that could ultimately bias the results (Hermalin and Weisbach, 2002). Therefore, an instrumental variables regression is used to circumvent the potential endogeneity issues related to the presence of women on corporate boards, ensuring robust results obtained from the regression analyses. As a follow-up on the first part of this thesis, the second part explores the firm’s residual cash variation with respect to board gender diversity and corporate acquisitiveness. By dividing the sample in subsamples of cash-rich and cash-poor firms, we are able to determine whether the presence of excess cash has any significant effect on the link between board gender diversity and the likelihood of a firm to initiate bids. In doing so, we will adopt the instrumental variables approach to test the potential role of excess cash, and more fundamentally, if the instrumental variables results obtained from the first part of this study remains robust when accounting for excess cash holdings.
5.1
First part of the research
Table Va and Vb show the results of the Probit and Negative Binomial estimations, where the likelihood of bid initiation and the natural logarithm of number of bid initiations is regressed upon three different measures for female directors and control variables, respectively. In columns 1, 2 and 3 the dependent variable is binary and corresponds to the likelihood of a firm becoming a bidder, whereas in columns 4, 5 and 6 the dependent variable is discrete and corresponds to the annual number of bid initiations. The standard errors for the Probit and Negative Binomial estimations are corrected for group correlation within firms and heteroskedasticity. In addition, entity fixed effects are included to absorb the influences of all omitted variables that varies across firms but are constant over time (Stock and Watson, 2012). However, all measures for female directors are considered to be endogenous variables, as the relation between a firm’s M&A activity and the extent of females on corporate boards may suffer from simultaneous causality. The endogeneity concerns surrounding the independent variables are not addressed in Table Va and Vb, and will be looked at in the following Tables. The estimated values of the coefficients in a probit regression are difficult to interpret, because they affect the probability of initiating a bid via the z-value. but their signs and significance are not. Hence, the results in Table V, columns 1 to 3 reveal that all measures for female directors are negatively related to the probability of a firm becoming a bidder and that this relationship is statistically significant at the 1% level. This suggests that a firm is less likely to become a bidder when it has more female members on its board of directors. With respect to the number of bid initiations, the results from columns 4 to 6 in Table V show negative estimates for all three measures of female directors and that this relationship is statistically significant at the 1% level for percentage of female directors and number of females directors, and statistically significant at the 5% level for female indicator. The results indicate that a firm is likely to perform less bid initiations when
6Alternatively, and in the same manner as Levi et al. (2013), the three measures for female board directors have been
regressed on the number of bid initiations in a given year. However, the effect of director gender on acquisition bids is not robust to the IV method addressing endogeneity, and has therefore less explanatory power than its likelihood equivalent variable. For reasons of brevity, the results are reported in Appendix B.
