Board composition and corporate philanthropy in companies from the United Kingdom

0

Master Thesis

Board composition and corporate

philanthropy in companies from the

United Kingdom

Student number: S3497925 Emal: g.hristova.1@student.rug.nl

Supervisor: dr. Viacheslav Iurkov (v.iurkov@rug.nl) Co-assessor: dr. Olof Lindahl (olof.lindahl@fek.uu.se)

Word count: 14 868 (excluding References)

18th _{of June, 2018}

MSc International Business & Management 2017-2018 University of Groningen

1

Table of Contents

Abstract ... 4

Introduction ... 5

Literature review ... 7

Corporate Social Responsibility ... 7

Corporate philanthropy ... 9

Corporate philanthropy in the UK ... 13

Board of directors ... 15

Hypothesis development ... 17

Outside directors on the board ... 17

Women on the board of directors... 18

The moderating effect of financial slack ... 20

Research Methodology ... 22

Sample ... 22

Data collection ... 23

Measures ... 24

Data analysis ... 29

Empirical analyses and results ... 31

Preliminary Analysis ... 31

Descriptive statistics and correlation test ... 33

Hypotheses testing ... 34

Robustness checks ... 37

Discussion ... 39

Implications and contributions ... 42

Limitations and future research ... 43

Conclusion ... 44

Appendices ... 45

2

List of Figures and Tables

Figures

Figure 1. Carroll’s Pyramid of Corporate Social Responsibility 9

Figure 2. Total pre-tax earnings and charitable donations made by the largest 100 UK companies 14 Figure 3. Conceptual model 22

Tables

Table 2. Means, standard deviations and correlations among variables 34

Table 3. Results from the regression analyses for Corporate philanthropy 36

Table 4. An aggregated overview of the regression outcomes and tested hypotheses 37

3

List of abbreviations

BoD - Board of directors CEO - Chief executive officer CP - Corporate philanthropy

CSP - Corporate social performance CSR - Corporate social responsibility DV- Dependent variable

IV - Independent variable LSE - London Stock Exchange MNE - Multinational enterprises OLS - Ordinary least squares PLC - Publicly listed company ROA - Return on assets SM - Senior management TM - Top management UK - United Kingdom

4

Abstract

Philanthropic activities have generated considerable interest among researchers and practitioners. In relation to this, the following paper, presents an empirical test of the determinants of corporate philanthropy. More specifically, it interlaces two important features of board composition - independent directors and female directors and consequently examines their effect on the amount of charitable giving. A moderation effect of slack resources is further implemented in the model. Drawing on a sample size of 278 publicly listed companies from the UK, multiple regression analysis has been performed in order to assess the relations between all included variables. Consistent with the first two hypotheses, the results of this study show that both outside and female representation on the board has a strong positive impact on the amount given to charity. Furthermore, financial slack is found to have an insignificant moderating effect over the abovementioned relationships. Implications for the research and business fields are given at the end of the paper along with some main contributions of this study.

5

Introduction

The modern concept of corporate philanthropy (CP) was set in 1953 in the United States (Cochran, 2007). Following the decision of the Supreme Court in New Jersey, the A.P. Smith Manufacturing Company was allowed to make a donation of $1500 to Princeton University without violating the interests of its shareholders (Burlingame, 2004). Thereafter, as a demonstration of social responsibility, CP has received considerable attention in a range of academic, business and community circles (Kabongo, Chang, & Li, 2013). Previous research has been mainly focused on the organization and the management of corporate donations (Campbell, Gulas, & Gruca, 1999; Siegfried, McElroy, & Biernot-Fawkes, 1983), the effect of tax adjustments on CP (e.g. Schwartz, 1968) or very often the relationship between firm’s financial performance and corporate charitable giving (Wang & Qian, 2011; Zulfiqar, 2016). Another line of research considers the strategic aspect of CP, showing that if handled strategically, corporate giving can positively affect the reputation of the company, as well as its competitive advantages and economic performance (Porter & Kramer, 2002; Saiia, Carroll, & Bucholtz, 2003; Wang & Qian, 2011).

However, strategic or not, the decision over the distribution of resources for social causes often lies in the hands of the top management (TM) (Brammer & Millington, 2006). Therefore, the relationship between corporate governance and CP has also been among the most frequently examined in the literature (Buchholtz, Amason, & Rutherford, 1999; Wang & Coffey, 1992; Navarro, 1988). In their paper, Wang and Coffey (1992) have found a positive relationship between corporate giving and the ratio of insiders to outsiders, the proportion of female and minority directors, and the percentage of shares owned by insiders. Ten years later, Bartkus, Morris, and Seifert (2002) have found positive relation between companies defined as “big givers” and larger corporate boards and no significant relationship between philanthropy and board composition. In his study of Fortune 500 companies, Williams (2003) has examined the relationship between the proportion of females on the company’s board and the extent to which those companies engage in philanthropy. Consequently, he has found strong support for the positive relation between higher proportion of women and corporate philanthropic activities.

6 behind this notion. First, CP is considered to be the most plausible and obvious demonstration of CSR towards societies (Harrow, 2013). Secondly, CP differs from the other forms of corporate responsibility as it is neither stimulated by the need for legal or regulatory compliance, nor is it mandatory for the firm (Brammer & Millington, 2005). Third, challenges related to the promptly changing corporate environment have led to the assumption that both companies’ and societies’ well-being may increase precisely as a result of CP (Coffey & Wang, 1998).

The importance of outside directors, female directors and slack resources is not less significant for the business and practice fields. Expanding the existing research on the role of directors in different firm outcomes is important because board executives may have strong influence on those outcomes (Daily, 1995). Moreover, at the end of the day boards are held responsible for the firm’s performance (Vance, 1983). Given the fact that nowadays, most of the publicly listed companies (PLCs) have more than 75% outside representation on their boards (Lorsch & MacIver, 1990), the necessity to develop better understanding of the role of independent board members is even greater than before. The UK represents both interesting and challenging setting for this research, since the variation between outside and inside directors is greater in comparison to the United States (Peasnell, Pope, & Young, 1999).

Gender diversity, as an important aspect of board composition, is also highly debated and significant issue (Rao & Tilt, 2015). Female directors are assumed to have a strong favorable impact over the CSR strategies (Ayuso & Argandona, 2007). Their increasing role in the board (Vinnicombe, Singh, Bilimoria, & Huse, 2008) attracts significant attention, especially in relation to socially responsible practices. Female directors are gradually gaining on their male counterparts in terms of CP, and as their board representation continues to grow (Williams, 2003), the importance of acquiring better knowledge on the topic grows in parallel.

By incorporating the moderating effect of slack resources, this paper adds insights to the scant empirical literature on the relationship between corporate giving and available resources (Seifert, Morris, & Bartkus, 2003). Financial slack is an important factor in this research, since it has been assumed as one of the main determinants for the variance between the giving practices of small, medium and large companies (Seifert, Morris, & Bartkus, 2004).

7 factors relative to CP is a proper way to narrow the literature gap in which those factors have been studied separately.

More specifically, the objective of this study is to provide answers to the following research questions:

How does the composition of the board influence the amount of corporate charitable giving?

 How does the proportion of outside directors affect the amount of charitable giving?

 How does the representation of women on the board affect the amount of charitable giving?

 How do slack resources moderate the relationship between 1) higher outside representation on the board and CP and 2) the number of women on the board and CP?

Literature review

Corporate Social Responsibility

In recent years, the concept of CSR has attracted substantial attention in the academia, business, media and even in the daily life of people. The proliferation of international trade conducted by MNEs and the globalization process have led to an increased demand for socially responsible practices in all industries, at almost every point on the globe (Kercher, 2007). Nowadays, CSR represents a global concept and an extremely essential part of the modern business agenda, since the expectations of stakeholders towards MNEs and smaller companies are constantly growing in way that merely making profits is not enough anymore (Carroll, 2015).

8 improvement of the CSR notion has been made by Walton (1967), as he introduces the voluntary aspect of the concept. More specifically, CSR very often involves expenditures of firm resources without guarantees for direct financial gains (Schwartz et al., 2009).

A critique response to the mentioned above CSR views was made by Milton Friedman in 1970 with his rebuttal paper ‘The Social Responsibility of Business is to Increase its Profits”. In his work, Friedman criticizes the rise of CSR and contradicts the growing importance of social activities exercised by enterprises, claiming that the sole responsibility of the company is to use its resources for profit maximization while conforming to general rules of society (Friedman, 1970). Furthermore, he argues that executives must conduct all business activities only in accordance with the desires of the shareholders. According to his work, all other social matters are of no concern to businessmen. Moreover, involvement in suchlike CSR practices represents wrong exploitation of corporate resources that could be better spent on internal firm projects with actual financial returns for the shareholders (McWilliams et al., 2006).

9 Figure 1. Carroll’s Pyramid of Corporate Social Responsibility

Corporate philanthropy

10 which they reside (Berman, Wicks, & Jones, 1999), today CP is a widespread practice across MNEs and small and medium-sized firms. It is also extensively used as a strategic approach for gaining competitive advantages (Gautier & Pache, 2015).

However, there is an ongoing debate both in the literature and in the practice field on whether corporations should undertake charitable activities (Buchholtz et al., 1999). The opponents of CP argue that such practices depict a misuse of firm resources that rightfully belong to the shareholders and can be better spend on more efficient projects (Buchholtz et al., 1999). Much of the controversial research is based on Friedman’s argumentation that corporations justify all their responsibilities by merely being profitable while creating wealth for the shareholders (Friedman, 1970). Moreover, he argues that instead of giving away their own money, executives spend the shareholders’ money on causes that may not be supported by the owners, hence CP is meaningless unless it brings back any financial value (Buchholtz et al. 1999). Donating resources to social causes generates additional costs for the company, which could interfere with the production process and eventually become detrimental for the whole organization (Gautier & Pache, 2015). From this point of view, the economic gains from CP are largely outweighed by the potential costs associated with it (Choi & Wang, 2007).

On the other hand, the expanding influence of corporations as a powerful structures has led to an increased interest towards the environmental and social impacts of their business operations (Paine, 2002; Rosen, Simon, Vincent, Fox, & Thea, 2003). As a result, CP has been perceived as an absolutely legitimate and appropriate practice that even helps companies gain social legitimacy (Margolis & Walsh, 2003). According to Leisinger and Schmitt (2011), in its pure sense, CP goes beyond the legal obligations of companies as it demonstrates the values they stand for, while participating in social activities without any expectations for financial returns. Other proponents of CP have devoted their work on the possible outcomes, building their arguments upon the social duties of corporations and the public welfare that can be created (Buchholtz et al., 1999). One of those works belongs to Buhl (1996), where he points that corporations are powerful constructs with the capacity to positively influence societies in return for the opportunity to compete in an untrammeled market environment. Moreover, Dickson (2003) indicates the favorable nature of corporate contributions, as such donations reduce governmental burdens. He further states that in countries with limited governmental resources, social improvements made by corporations are highly appreciated and important for the community (Dickson, 2003).

11 been encouraged to involve in philanthropy for strategic reasons. An abundance of empirical research examines the strategic relevance of CP and its usage as a tool for both supporting social causes and improving the firm’s bottom line (Porter & Kramer, 2002; Saiia et al., 2003). In his work “The New Corporate Philanthropy”, Smith (1994) associates charitable giving with improved brand recognition, increased employee productivity, decreased research and development costs and regulatory barriers and improved relationships with investors. CP can switch the intentions of customers and boost their loyalty, hence as noted by Campbell et al. (1999), donations can have a positive effect on firm’s sales, revenues and profitability. Moreover, companies involved in improving the common good build positive image and gain sociopolitical legitimacy as they declare responsible attitude towards the stakeholders, institutions and the environment (Qian et al., 2015). From the perspective of Fombrun (1996), corporate giving can create what he calls “reputational capital” or in other words the ability of companies to arrange more favorable contracts with host governments, to attract valuable workforce, to decrease their cost of capital and charge with artificially high prices for their products. In this line of research, Godfrey (2005) further argues that high stock of reputational capital can mitigate the risk of reputational losses and secure the value of intangible assets occurred as a result of the relationship between companies and stakeholders. From these perspectives, the strategic benefit of corporate giving suggests that participation in social causes most often increases the value of the donor firm (Patten, 2008).

Corporate philanthropy through the lens of Agency theory

CP is not a legal obligation for corporations, but it is highly desirable by the stakeholders (Carroll, 2016). However, spending firm resources may not be satisfying for the interest of all stakeholders. Clarkson (1995) indicates the need of CEOs to balance the interest of customers, employees, suppliers, investors, shareholders and local communities, while the traditional economic perspective holds the corporation responsible only to its stockholders and the maximization of their wealth (Gan, 2016). Following this perspective, the corporation has no reasons to spend the money of the shareholders for merely altruistic purposes, unless those expenditures create benefits higher than the incurred costs (Gan, 2016; Seifert et al., 2003). For this reason, CP is considered as a sensitive issue that could lead to agency problems since agents (namely executives who make corporate decisions) and principals (shareholders or company owners) are likely to have conflicting interests.

12 profit-related. Principals are profit oriented since they have direct connection to the firm’s earnings, while agents are driven by other motives such as improving their individual social standing and self-interest (Wang & Coffey, 1992). Several studies (Boatsman & Gupta, 1996; Galaskiewicz, 1997) present evidence of the significant impact of managers and directors on the philanthropic activities in a company. Agency theory proposes that, if not properly monitored, executives might spend corporate resources on charities in order to improve their own image and achieve better appreciation from the society (Galaskiewicz, 1985). CP represents an agency cost because it enhances the agent’s utility on the account of the operating losses that occur for the owners (Brown 2006). Consequently, problems of agency type arise as a result of goal divergence and information asymmetries between the agents and the principals, followed by the potential for opportunistic behavior exerted by the agent (Fama, 1980; Fama & Jensen, 1983; Jensen & Meckling, 1976).

Corporate philanthropy through the lens of Stakeholder theory

A different theoretical perspective considers the importance of the external environment and the preferences of stakeholders. As an important underpinning of the CSR research, stakeholder theory has received significant attention in the literature, as it calls for change in the value system of modern corporations from shareholder perspective to a stakeholder perspective (Wang & Dewhirst, 1992). One of the advocates of the stakeholder theory, Freeman (1984), claims that companies operate in complex environments, comprised of diverse members where each of them can influence or is influenced by the business actions of firms. He further states that owners, investors, managers, employees, customers, governments and societies all have legitimate claims on today’s modern corporation (Freeman, 1984).

13

Corporate philanthropy through the lens of Upper echelons theory

The upper echelons perspective focuses on the impact of age, gender diversity, tenure, aptitude, experience and other individual characteristics of the senior management (SM) on the decision-making process in a company (Finkelstein, Hambrick, & Cannella, 2009). According to this theory, the TM has the capacity to make specific strategic decisions and select among range of potential options (Marquis & Lee, 2013). In their paper, Hambrick and Mason (1984) introduced a framework based on the connection of CEOs’ characteristics to strategic choices, arguing that the corporate decision-making process reflects the values and cognitive perceptions of the executives. Given that capturing the exact cognitive frames of directors is not possible, the upper echelons research has implemented observable characteristics such as race and gender as relevant proxies (You, Terjesen, & Bilimoria, 2016). In this line of research, several scholars have outlined the contribution of women’s previous experience, values and unique knowledge to the corporate pool of information and its importance for decision-making (Dutton & Duncan, 1987; Larkey, 1996; Sutcliffe, 1994; Watson, Kumar, & Michaelsen, 1993).

As a corporate practice, decisions for charitable giving are formulated at the top of the company (Marquis & Lee, 2013). Prior study has found that CP has been influenced by the individual characteristics of the executives responsible for the payment activities of the company (Galaskiewicz, 1997). In addition, Useem & Kutner (1986) have found that corporate charitable decisions have been affected by the relationship of the executives with managers of other firms and non-profit organizations.

Corporate philanthropy in the UK

14 In their paper, drawing on a sample of 550 UK companies, Brammer and Millington (2005) have found that British firms’ corporate charitable activities were mostly driven by economic incentives. From a legal perspective, companies involved in purely philanthropic activities with no direct business relations have been subject to questionable discussions in the UK (Cowton, 1987). In fact, in the British corporate law, the whole topic of in-kind gifts and charitable donations has been somehow vague and has created complex issues (Sealy, 1978). However, according to the Companies Act 1985, business organizations are now required to disclose information regarding their charitable activities and the amount of their payments (Companies Act 1985, Schedule 7, Paragraph 3 to 5). In general, companies should have sections in their annual reports and corporate accounts containing disclosed information regarding contributions donated to either social charities or political purposes (Cowton, 1987). According to the Companies Act 2006, part 14, companies are obliged to report political donations made within a 12-month period with an aggregate amount exceeding £5,000. Similar threshold is not required for donations directed at social or community causes. Limited companies form the UK pay less corporate taxes if they donate cash, products or services they make or sell, land property or shares, volunteered employee work time, or provide sponsorships payments (UK, 2017).

15

Board of directors

Every PLC - large, medium or small is required by the corporate laws of its country of operations to appoint board of directors (BoD) (Wang & Coffey, 1992). Boards exist as regulatory organs and governance mechanisms for avoiding or solving agency problems between the TM and the shareholders (Hermalin & Weisbach, 2003). The main responsibility of board executives is to ensure that the decision-making process benefits the stockholders, while protecting the interests of the stakeholders (Wang & Coffey, 1992; Rechner & Dalton, 1991). Furthermore, the board has the obligation to oversee both the CEO and the SM on behalf of the owners (Bratkus et al., 2002). Directors on the board have the jurisdiction to employ or dismiss people, to compensate the TM, to resolve agency conflicts between decision-makers and company owners (Baysinger & Butler, 1985). Very often boards represent a mixture of some of the firm’s top managers and outside (independent) directors (Byrd & Hickman, 1992). While inside directors are focused on day-to-day activities and provide information about the firm’s operations, outside directors redound external expertise and objectively monitor the decision-making process (Byrd & Hickman, 1992).

The research on BoD has generated considerable attention both in academic and business aspects and has been extensively conceptualized in the empirical realm (Johnson, Daily, & Ellstrand, 1996). However, convergence over the specific role of corporate boards is still missing (Lorsch & MacIver, 1990; Vance, 1983). One literature stream (Forbes & Milliken, 1999; Aguilera, 2005; OECD, 2003; Nicholson & Kiel, 2004) suggests that the main roles of the board are direction (strategic guidance) and control (managerial supervision). Another stream in the literature (Huse, 2005; Roberts, McNulty, & Stiles, 2005) proposes that board roles are created by the contributions of individual board members.

16 perspective holds the board responsible for the corporate leadership and the legal obligations, leaving day-to-day duties to the CEO and the SM (Zahra & Pearce, 1989). In its legal function, the board is responsible for selecting or replacing the CEO, representing and protecting the stockholders’ interests, operating as monitoring and controlling body and providing guidance to the TM (Carpenter, 1988; Ewing, 1979; Mattar & Ball, 1985; Muller, 1979; Vance, 1983).

Despite the growing importance and various positives, during the past few decades the discussion on corporate boards has also received some strong criticism from both practitioners and intellectuals (Wang & Coffey, 1992). Moreover, multiple recommendations have been made with respect to reforms in the BoD (Wang & Coffey, 1992). For example Lorsch and Maclver (1990) argue that due to the intense influence of the CEO over the rest board members, they usually do not have the motivations to be effective, thus reconsideration of the responsibilities of directors has been increasingly required. Other restructuring suggestions such as splitting the position of the CEO from the one of the Chairman of the board, abbreviating the number of directors, demanding directors to own a certain share of the stocks relative to their compensation or implementing criteria for board evaluation can be also found in the literature (Drucker, 1973; Mace, 1972; Vance, 1964, 1983).

Nevertheless, requiring boards to include more independent, non-affiliated representatives has become the most frequently suggested boardroom reform (Vance, 1964; Wiley, 1995). For example, Dunn (1987) argues that boards generally fail to monitor and control the management process, hence calling for boards almost exclusively composed by outsiders. In addition, Eisenberg (1976) suggests that the ideal board composition consists entirely of independent, non-employed directors, with the exception of the CEO. Other proponents of this idea believe that higher representation of non-aligned directors can enhance the independence of the board, provide better knowledge, greater experience and increased board efficiency (Bacon and Brown, 1973; Williams and Shapiro, 1979).

17

Hypothesis development

Outside directors on the board

Outside and inside directors differ in terms of their level of involvement in the corporate operations as well as in their CSR orientation (Zhang, Zhu, & Ding, 2013). Inside directors are current associates in the TM and are currently employed in the firm or in its subunits, while independent directors have no direct association with the firm or its subsidiaries (Pearce & Zahra, 1992). The classification of outside directors can be broadly separated in two – affiliated or non-affiliated. Affiliated outside members are not part of the present management and are not employed by the company, but still have close business relations to it (Pearce & Zahra, 1992). Directors with no relations to the company, who are not under the control of the executives are called non-affiliated or independent and are mainly employed because of their previous experience, managerial expertise or name recognition (Pearce & Zahra, 1992).

When it comes to the relationship between board composition and different CSR aspects (such as CP), the existing research is quite mixed (Zhang et al., 2013). For example, Wang and Coffey (1992) have suggested that boards with more salient inside representation will display higher levels of CP involvement since agents do not have any direct claims on the firm’s earnings and they are more likely than principals to spend corporate funds on charity (Wang & Coffey, 1992). By exploring the retailing industry, Chaganti, Mahajan, and Sharma (1985) argued that non-affiliated directors might encounter difficulties in understanding the internal complexities of the company, hence might not be fully effective when it comes to decision-making and responsible behavior.

18 to that, Sheppard (1994) contends that outside directors bring indications for strong external environment orientation, hence have the ability to better respond to changes and support the corporation through their network of external contacts (Borch & Huse, 1993). In line with the prevalence of findings in support of the outside-dominance, the following can be hypothesized:

H1a: Higher proportion of outside directors on the board will positively influence the amount of

corporate philanthropy.

Women on the board of directors

19 stakeholders and are more responsive to their needs (Galbreath, 2011), or display higher solicitude towards the environmental aspects of CSR (Post, Rahman, & Rubow, 2011).

In the context of discretionary activities, such as CP, Williams (2003) claims that women are slowly outstripping men and as the number of female directors continues to grow, their role in the decision-making of money spending grows as well. In his study, he found positive relationship between women on top positions and the discretionary support given to local communities and various cultural activities (Williams, 2003). According to him, one possible explanation for the positive relationship might be that the control over abundant firm resources allows women to use charitable giving as a source of power (Williams, 2003).

However, another stream of research gives different view over the reasons why women are more inclined to engage in corporate charities than their male fellow workers. Vast amount of the gender diversity literature argues that substantial differences exist between the values, moral perceptions and beliefs of women and men (Eagly, Makhijani, & Karau, 1995; Powell, 1990). Such differences are very often reflected in their professional life and have strong influence over their board positions (Rao & Tilt, 2015). More specifically, while agentic (related to agency theory) characteristics such as ambition, self-confidence and independence are more often related to men, women are usually more sensitive, emotional and concerned about other people’s welfare, show higher levels of empathy and affection, and are generally more helpful and gentle (Rao & Tilt, 2015). Following this viewpoint, Ibrahim and Angelidis (1991) have found that as opposed to men, female directors are more concerned about CSR and philanthropy rather than the economic performance of the company. Furthermore, Harrigan (1981) has suggested that women typically have backgrounds in law, education or various non-profit activities, which justifies their propensity towards philanthropy (Wang & Coffey, 1992). Even more, women usually occupy “soft” corporate positions in departments such as human resources, marketing or advertising (Zelechowski & Bilimoria, 2006), implying that compared to men, female leaders are more likely to display deeper concern about soft managerial issues and stakeholders’ social needs (Rao & Tilt, 2015). The abovementioned arguments therefore lead to the following hypothesis:

20

The moderating effect of financial slack

One of the most common managerial problems is the struggle for proper balance between operating efficiently and having available resources that can be used in case of unexpected situations (Daniel, Lohrke, Fornaciari, & Jr., 2004). The presumption that CP somehow depends on the slack availability is broadly accepted in the literature, yet the empirical support for this notion is still scarce (Seifert et al., 2004). Some researchers have assumed that corporate profits are predictors for greater CSP and corporate giving in particular (Seifert et al., 2004). However, those studies have resulted in inconclusive findings ranging from positive to insignificant or negative (Griffin & Mahon, 1997). In their study, Griffin and Mahon (1997) have failed in finding correlation between the financial performance of the company and its philanthropic activities. According to Buchholtz et al., (1999) the main reason for the missing correlation is that “financial performance” has different meanings for the different companies as many of them might yield high profits, but still be scarce on available resources (Seifert et al., 2004). In other words, corporate giving is likely to depend on the organization’s slack resources rather than merely on its profits (Buchholtz et al., 1999).

Cyert & March (1963) define “slack” as the difference between firm’s total resources and its total obligatory payments. Over the years, scholars have indicated several forms of slack such as spare raw materials and workforce, extra inventory or additional machine capacity. However, the most popular form of slack is financial slack or the surplus of cash (Sharfman, Wolf, Chase, & Tansick, 1988). Financial slack is the excess amount of cash, which is very flexible and can be spend for a wide range of activities (Bourgeois & Singh, 1983). Following the slack perspective, both the studies of Adams and Hardwick (1998) and Galaskiewicz (1997) have discovered that accounting returns such as income over sales or income over total assets are predictors for monetary corporate giving (Seifert et al., 2004). In 1981, Bourgeois further enriches the literature by adding that slack acts as a cushion, of potential or present resources, that allows the firm to adapt effectively to internal or external pressures to change (Bourgeois, 1981).

One of the most convincing evidence for the relationship between slack and philanthropy is the study of Buchholtz et al., (1999) consisting of 43 companies from two industries. The authors found that the perceived organizational slack (the CEO’s assessment of the available firm resources relative to the competition) affects the process of decision-making in terms of discretionary activities and subsequently found positive relationship between perceived resource availability and CP.

21 Therefore, when managers spend available resources on charitable activities, they donate funds that could be otherwise invested in profitable projects and are rightful possession of the shareholders (Seifert et al., 2004; Jensen, 1986). Hence, agency theory argues that if CP creates agency problems, then they would increase if the company has slack resources (Seifert et al., 2004).

On the other hand, after making all mandatory payments companies are left with their slack, which represents a discretionary resource, part of which can be donated to social causes (Seifert et al., 2004). Following this assumption, the supporters of the CSR literature (Carroll, 1979, 1991; Mitchell et al., 1997) suggest a positive relationship between discretionary practices and available resources. Given the fact that CP is considered as the most discretionary aspect of CSR, it can be assumed that the amount of corporate donations would rise if the amount of available resources increases. As it has already been suggested, outside directors and women on the board are more concerned about the philanthropic activities rather than the economic achievements of the company in which they operate (Ibrahim & Angelidis, 1995). As external entities, outside directors are related to the stakeholders to a larger extent than their insider colleagues. Moreover, due to their increased representation and very often higher financial expertise (Fligstein, 1991), independent directors play an important role in the slack deployment decisions, leading to the presumption that if the levels of slack are higher, probably larger amounts will be spend on charitable causes. Thus, the following can be hypothesized:

H2a: The positive relationship between proportion of outside directors and the amount of corporate

philanthropy is positively moderated by the slack resources

The same propositions can be made regarding the female representation on the board. As it has already been suggested, women on the board tend to have positive impact over the charitable practices of the companies in which they enact their careers (Williams, 2003; Wang & Coffey, 1992; Ibrahim & Angelidis, 1995). Women are also found to display higher levels of affection and empathy and to be more sensitive towards the demands of stakeholders (Rao & Tilt, 2015). The strongly expressed commitment of women to stakeholders, supported by higher cash availability, will probably give those directors an incentive to engage the firm in various charitable causes and spend larger amounts of money for CP. Therefore in view of the given arguments, the following can be hypothesized:

22

Figure 3. Conceptual model

Research Methodology

Sample

23

Data collection

For the purpose of this research, data was collected only from online-accessible secondary sources. To examine the effect of board composition on CP, data for the number of outside directors, percentage of women on the board and the board size was obtained from the annual reports of the sampled companies. Information regarding the aggregate amounts of charitable donations was also collected from the annual reports of the companies. The time period chosen for the board data collection was 2016, while the period for philanthropic data collection was 2017. The time lag of 1 year appears to be sufficient in case of corporate giving, which also allows establishing temporal causality between the IVs and the DV. Engagement in CP is not a long-term process and it is usually measured on an annual basis, therefore the influence of the appointed directors on the board during 2016 can be reflected in the donated amount of charity in 2017.

One of the main objectives of this study is to examine the effect of financial slack upon the relationship between board composition and philanthropy. For this purpose, data for the available resources for each company was collected from the Orbis database. In most cases, the current ratio has already been calculated in the database, yet for some companies the ratio between current assets-to-current liabilities had to be retrieved from the annual reports. Data for cash flows generated from operating activities was also found in the financial statements of the firms. As values in Orbis were presented in US dollars, they were re-calculated in British pounds by the given in Orbis exchange rates for the respective year. Slack data was collected for the period of 2016 under the presumption that the amount of money given to charity in 2017 depends on the amount of unabsorbed financial slack in 2016.

24

Measures

Dependent variable

Corporate philanthropy

As part of numerous studies, measures for corporate philanthropy differ in the research field. For example, Wang and Coffey (1992) measure philanthropy as the percentage of the pre-tax earnings given to charitable causes. In their study, Seifert et al., (2004) examined corporate giving in cash, while excluding all non-monetary contributions. They define cash donations as the aggregated amount of monetary gifts either directly given to charities or through the corporate charitable foundation (s) of the firm (Seifert et al., 2004). Another study (Buchholtz et al., 1999) has examined CP by adding together the in-cash and the in-kind firm contributions to get the total amount of philanthropy. As a final measure, the authors have used the natural logarithm of the combined variables. Furthermore, while controlling for firm size, Wang, Choi and Li (2008), measured philanthropy by the dollar amount given by firms over a period of 1 year, scaled by the firm sales for the same period. In their paper Campbell, Moore, and Metzger (2002) examined corporate giving in the UK, by measuring corporate “generosity” as the ratio between charitable cash donations and earnings before tax and after interest.

In line with Brammer and Millington (2006) and Marquis and Lee (2013), rather than using the ratio of philanthropy to total sales or pre-tax profits, this study implements the natural logarithm of the raw amount of CP as a relevant measure. Due to the high variance of the amounts given to charities, the values were highly skewed, hence the logarithmic function has led to more normal data distribution. Furthermore, due to the imperfect linear relationship between the DV and the IVs, logarithmic transformation of the amount of CP donations was necessary. Consistent with previous research and considering the data obtainability limitations, the variable includes only the cash donations1 made by the companies within the period of 2017. All other forms of philanthropy such

as employee volunteerism, in-kind giving in the form of firm products and services, sponsorships and various assets were excluded. 1

1 _{Cash donations= direct donations to charities made in cash+ donations to corporate-sponsored foundations made}

25

Independent variables

Outside director representation

The first independent variable in this paper represents the proportion of outside directors to the total number of directors on the board. Bhojraj and Sengupta (2003) have measured the outside representation as the percentage of directors who are not simultaneously officers in the company. However, perhaps the most frequently used measurement in the literature is the ratio of outside directors to the total number of board members (Zajac & Westphal, 1996). In line with the study of Zajac and Westphal’s, outside representation will be measured as the proportion of independent directors to the total number of directors. Both affiliated and non-affiliated outside directors are included in the research. Since BoD represents a dynamic corporate structure and its composition changes frequently, data for the amount of outside directors was collected only for the year of 2016.

Female director representation

The second independent variable in the study is related to the number of female directors on the board. Previous study has examined the number of women on the board by simply counting the female representatives (Bear et al., 2010). Other researchers have used the ratio of women on the board, by dividing the number of female members to the total board size (Harrison & Klein, 2007; Campbell & Vera, 2007; Adams & Ferreira, 2009; Galbreath, 2011; Jia & Zhang, 2012; Boulouta, 2013). Following the works of Wang and Coffey (1992) and Williams (2003), this study will implement the percentage of women on the board as a relevant measurement. Almost all companies’ annual reports contain the exact percentage of female representation on their boards. For those companies without such information, data for the variable was obtained by dividing the number of women directors by the total number of directors and the result was multiplied by 100.

Moderator

Slack resources

26 ratio of firm’s general and administrative expenditures to sales. In another paper, Seifert et al., (2004) have subtracted the interests, taxes, preferred and common dividends from the operating income before deduction.

Focusing on the unabsorbed slack, this study implements one of the most commonly used measures for financial slack - the current ratio (Zahra, 1996; Hambrick, Cho, & Chen, 1996; Iyer & Miller, 2008). Moses (1992) has claimed that the best way to determine financial slack as existing in surplus is by estimating the difference between the current assets and liabilities of a company. The current ratio represents the ability of a company to generate liquid capital in order to meet its short-term monetary responsibilities (Greenley & Oktemgil, 1998). The higher the ratio between assets and liabilities, the higher the amount of liquidity that can be used for future suppleness (Greenley & Oktemgil, 1998). Due to the high variance, data for slack was also skewed, and transformed into natural logarithm. Calculations were made according to this formula:

𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑟𝑎𝑡𝑖𝑜 = 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑎𝑠𝑠𝑒𝑡𝑠 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑙𝑖𝑎𝑏𝑖𝑙𝑖𝑡𝑖𝑒𝑠

Control variables

Board size

One of the control variables included in this study is board size, measured as the total number of directors on the board. The assumption that board size might influence CP is twofold. First, larger boards may face significant free-riding disputes and communication issues, resulting in less effective control over the members (Brown, Helland, & Smith, 2006). Second, if socially responsible companies tend to employ CSR oriented directors, then larger size of the board will imply larger philanthropic commitment and desire for charitable giving from each member on the board (Brown et al., 2006). Furthermore, Aggarwal and Nanda (2004) argued that larger boards usually have numerous objectives that go beyond the shareholders’ interests. For such reasons, this variable has been frequently used to control the sample (Goodstein, Gautam, & Boeker, 1994).

Firm size

27 (2007) have discovered that large and small companies donate more, as opposed to medium-sized firms. Possible assumptions for the positive relationship between firm size and philanthropy could be that larger firms have larger resource availability and scale of economies in terms of increasing CSP (Bowen, 2000). However, despite of its wide use in the literature, there is no exact measure for firm size and scholars use different proxies (Dang, Zhichuan, & Chen, 2018). Some studies have implemented the number of employees as a relevant measure (Cha & Abebe, 2016), while perhaps the most commonly used measure for firm size in the organizational theory has been the amount of total assets (Amato & Amato, 2007; Adams & Hardwick, 1998; Seifert et al., 2004).

Using data from Orbis and the annual reports of the companies, in this study firm size has been measured by the natural logarithm of firm total assets. The logarithmic transformation was necessary because of the high difference between the amounts of total assets among the sample.

Firm performance

Next control variable in the study is firm performance. It has been vastly used in the literature, especially as a dependent variable, still there is a lack of consensus about its specific definition (Santos & Brito, 2012). Firm performance has broad meaning as it could be representative for the firm’s social performance, profitability performance, market performance, environmental performance or customer and employee performance (Santos & Brito, 2012). For the purpose of this study, only the financial dimension of firm performance has been included. Financial performance has been measured in different ways in the literature field. For example Seifert et al., (2004) used the accounting-based unit return to shareholders, while Santos and Brito (2012) suggested other relevant measures such as return on investment and return on equity. In their study of 160 companies from Fortune 500, Stimpert and Duhaime (1997) have used the operating margin for measuring firm performance. However, perhaps the most popular measure for financial performance in the literature remains return on assets (ROA) (Berman et al., 1999; Chang & Singh, 1999; Zahra, 1996; Marquis & Lee, 2013).

Following the notion that ROA has been one of the most frequently used measures, this paper will also implement it as a proxy for performance. ROA was measured by the following formula:

𝑅𝑂𝐴 =𝑃𝑟𝑜𝑓𝑖𝑡(𝐿𝑜𝑠𝑠)𝑏𝑒𝑓𝑜𝑟𝑒 𝑡𝑎𝑥

𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠 × 100

28

Firm age

Consistent with previous studies, which have assumed that firm age has significant importance for the philanthropic activities of corporations (Logsdon, Reiner, & Burke, 1990), the proxy will be included in the set of control variables. One of the assumptions is that older firms may be more inflexible towards CSR activities. Firm age might be related to differences in the competitiveness levels of firm due to their history (Wang, Choi, & Li, 2008). On the other hand, Withisuphakorn and Jiraporn (2016) have found that firm age has a substantial impact on the firm CSR investments, as more mature companies tend to be more responsible towards the environmental problems.

Because of its importance in previous research, firm age has been also considered as a relevant control variable and has been measured by the years since incorporation. Data was obtained from Orbis and 2017 was used as an utmost year. The variable was also transformed into natural logarithm, due to the high variance between the organizational ages of the companies.

International diversification

As an additional control variable, this study will incorporate international diversification. International diversification indicates company’s expansion to other countries different from its domestic market (Kang, 2013). Internationally diversified companies increase their range of stakeholders, hence are under stronger pressure for responsible behavior (Christmann, 2004). Moreover, as the amount of stakeholders increases, the issues related to the social and environmental impact of the company’s operations increase as well (Kang, 2013). Therefore, being socially responsible becomes more of an imperative rather than an optional practice (Henriques & Sadorsky, 1996). One of the most frequently used measures for international diversification is the ratio foreign sales-to-total sales (Kang, 2013; Geringer, Beamish, & Dacosta, 1989). Other methods of measurement include the total number of subsidiaries outside the home country or the total number of countries in which the company has subunits (Strike, Gao, & Bansal, 2006).

29

TABLE 1. Variables overview

Variable Measurement

Corporate philanthropy natural logarithm (ln) of the raw amount of CP Outside directors ratio outside directors-to-total number of directors Female directors % of women on the board

Slack resources current ratio = current assets-to-current liabilities (ln) Board size total number of directors on the board

Firm size total assets (ln)

Firm performance return on assets using P/L before tax (%) Firm age number of years since incorporation (ln)

International diversification 0-no foreign subsidiaries; 1-has foreign subsidiaries;

Data analysis

The main objective of this research is to test the potential relationship between board composition and CP, while adding the moderating effect of slack resources and a set of control variables. Almost all variables included in the paper are continuous, meaning that they can obtain an infinite amount of possible values. CP, board size and firm age are discrete and countable. The logarithmic transformation of CP, slack resources, total assets and firm age was necessary as it has led to more accurate data distribution. In order to fulfill the stated objective, a multiple linear regression analysis (estimated by ordinary least squares) was executed. By implementing this approach the correlations between the proportion of outside directors and the percentage of female directors to corporate giving were investigated through testing Hypotheses 1a and 1b. Furthermore, the moderating effect of financial slack on the main relationships (H1a and H1b) was examined by performing moderated regression analysis. As a result Hypotheses 2a and 2b were statistically tested. The program used for the regression analyses was Stata SE 15.

Several models were performed in order to determine the relationships between the aforementioned variables. All models were based on the following formula:

30 Y=dependent variable corporate philanthropy

β₀= constant variable β 1x1= independent variable 𝐩𝐫𝐨𝐩𝐨𝐫𝐭𝐢𝐨𝐧 𝐨𝐟 𝐨𝐮𝐭𝐬𝐢𝐝𝐞 𝐝𝐢𝐫𝐞𝐜𝐭𝐨𝐫𝐬 β 2x2= independent variable 𝐩𝐞𝐫𝐜𝐞𝐧𝐭𝐚𝐠𝐞 𝐨𝐟 𝐟𝐞𝐦𝐚𝐥𝐞 𝐝𝐢𝐫𝐞𝐜𝐭𝐨𝐫𝐬 β ₃x₃= moderator variable 𝐬𝐥𝐚𝐜𝐤 𝐫𝐞𝐬𝐨𝐮𝐫𝐜𝐞𝐬 β4x4= control variable 𝐛𝐨𝐚𝐫𝐝 𝐬𝐢𝐳𝐞 β5x5= control variable 𝐟𝐢𝐫𝐦 𝐬𝐢𝐳𝐞 β6x6= control variable 𝐟𝐢𝐫𝐦 𝐩𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 β7x7= control variable 𝐟𝐢𝐫𝐦 𝐚𝐠𝐞 β₈x₈= control variable 𝐢𝐧𝐭𝐞𝐫𝐧𝐚𝐭𝐢𝐨𝐧𝐚𝐥 𝐝𝐢𝐯𝐞𝐫𝐬𝐢𝐟𝐢𝐜𝐚𝐭𝐢𝐨𝐧

β₉x₉= the effect of slack on the relation between proportion of outside directors and CP (outside directors ∗ slack resources)

β10x10= the effect of slack on the relation between percentage of female directors and CP (female directors ∗ slack resources)

ε= the error term

I. The first model examines the effect of the control variables on the dependent variable CP. The model has been executed by using the following formula:

Y=𝛽0+ 𝛽4𝑥4 + 𝛽5𝑥5+ 𝛽6𝑥6+ 𝛽7𝑥7+ 𝛽8𝑥8+ ε

II. The second model displays the relationship between the dependent variable, the two independent variables, the moderator and the control variables. There is no multicollinearity between the

independent variables, therefore it was not necessary to run two separate regressions. The following model was used:

Y=𝛽0+ 𝛽1𝑥1+ 𝛽2𝑥2+𝛽3𝑥3+𝛽4𝑥4+ 𝛽5𝑥5+ 𝛽6𝑥6+ 𝛽7𝑥7+ 𝛽8𝑥8+ ε

III. The final model tests the moderating effect of slack resources on the relationships between the independent variables and CP. It once again combines the two independent variables instead of testing them separately. The model is represented by the following formula:

31

In order to obtain more meaningful and straightforward for interpretation results (Williams, 2015), mean-centering of the independent variables and the moderator was performed before the moderated regression (TABLE 3., Model 3). Centering the independent variables represents a set of few steps - subtracting the mean of each observation, calculating the interaction term and estimation of the model. Empirical findings show that mean-centering can enhance conditioning by decreasing the covariance and correlations between the independent variables and the moderator (Shieh, 2011). Furthermore, it reduces the potential for multicollinearity and increase the accuracy of the results (Shieh, 2011)

In order to control for potential misspecifications and heteroscedasticity all models were executed by using robust estimates of the standard errors. The robust estimator of variance differ from the conventional one because it implements the assumption that all observations are independent form each other, thus it produces “correct” standard errors (StataCorp, 2013).

Empirical analyses and results

Preliminary Analysis

The main purpose of the preliminary analysis is to revise the data, define its key features and prepare it for testing the underlying hypotheses (Blischke, 2011). The majority of statistical tests rely on certain assumptions regarding the variables included in the analysis (Osborne & Waters, 2002). Therefore, before conducting the regression analyses, four main assumptions have to be fulfilled in order to justify the aptitude of the data (Laerd Statistics, 2013). In statistics, these assumptions are known as normality, linearity, multicollinearity and homoscedasticity. Testing for those assumptions is essential, because if they are not met, the results of the statistical regressions may be unreliable (Osborne & Waters, 2002).

Normality

32 illustration of normality testing is given in the Appendix of this paper (Appendix 1.1), but as it rather provides subjective ways of investigating normality, Shapiro-Wilkinson test has been also shown in the Appendix. The Q-Q plot shows that the data for the DV is plausible and distributed normally as it almost perfectly falls on the straight line. Slight disruption of the data distribution can be seen at the top-right end of the graph due to the larger amount of donations from a few companies in comparison to the total sample. In a similar way, the Shapiro-Wilkinson test displays high values of W (=0.98524), indicating that the residuals are normally distributed. After these results, it can be concluded that the dependent variable does not violate the assumption of normality.

Linearity

The second assumption is made regarding the linearity of the relationship between the independent variable (s) and the dependent variable (s). The standard multiple regression is able to examine the connection between the predictor variables and the outcome variables accurately only if the relationships are linear (Osborne & Waters, 2002). In case of failure of fulfilling the assumption of linearity, the results may be either under or over-estimated (Osborne & Waters, 2002). A common way of testing this assumption is by using residual plots (also known as scatterplots) that display linear or curvilinear relations (Osborne & Waters, 2002). Two scatterplots have been displayed in the Appendix - one for the relationship between CP and outside directors and one for the relationship between CP and female directors. Regarding the first scatterplot, it can be observed that the linearity is not perfectly represented, but still substantial amount of the observations falls on the line, suggesting slight linear relationship between the variables. Regarding the second scatterplot, it shows better alignment of the observations with the line (with the exception of those companies without female representatives, where scores are =0). It can be assumed that the second assumption is also met, since there are no significant disturbances in the patterns. However, it is important to address the existent non-linear relationships in order to assess the relationship between the variables in the best possible way (Osborne & Waters, 2002).

Multicollinearity

33 variables becomes problematic. However, table Appendix 3. in the Appendix shows that the assumption of multicollinearity has been fulfilled as well.

Homoscedasticity

Homoscedasticity assumes that the error variation is the same within all levels of the independent variable (s) (Osborne & Waters, 2002). In case the variance of errors is not the same at different values of the independent variables, this means that there is heteroscedasticity (Osborne & Waters, 2002). When models face heteroscedasticity the analysis becomes weaker leading to potential disruptions in the findings (Osborne & Waters, 2002). Testing the assumption might be done by using visual representations - plots displaying the standard against the unstandardized residuals (Laerd Statistics, 2013). The scatterplot Appendix 4. shows that the residuals are randomly spread around 0 (the horizontal line), meaning that the last assumption is also not violated.

Descriptive statistics and correlation test

34

Hypotheses testing

Detailed results of the regression models can be seen in Table 3. The table shows that the full model has statistical significance (F= 32.68; p= 0.0000) Furthermore, 58.33 percent of the variance in the model is evidenced by the R2_.

The first model (Model 1) tested the effect of all control variables on the DV and found positive coefficients for all pairs except CP and international diversification. Board size, total assets and ROA have found to be positively and significantly correlated with CP. The null hypothesis that all coefficients in the model are equal to 0 (because Prob > F = 0.0000) is supported at 42.70 by the F-static. The proportion of variance (R2_{) is 0.4981 indicating strong association between the variables.}

35 and international diversification remains in this model as well. R2 _{has increased its value, reaching}

0.5464, while the F-static value decreases to 30.84.

The last model (Model 3) displays the full analysis as it incorporates the moderating effect of slack resources on the relationships between the DV and the IVs. The R2 _{value of the model is}

0.5470, while the p-value associated with the F-static is 0.0000. The adjusted R2 _{value 0.5293}

indicates sensibly good fit of this particular model.

Main effects

Hypothesis 1a predicted that companies with higher proportion of independent directors will contribute higher amounts of money to charitable causes. The log-linear model (log Yi = α + βXi + εi) used for testing the relationship has shown positive and significant coefficient (β=2.083; p=0.004), meaning that one unit increase in the ratio outside directors-to-total number of directors will consequently result in expected increase of the amount given to CP of 2.08 units. These results imply that the first hypothesis has received strong statistical support.

Hypothesis 1b suggests positive relationship between higher female representation on the board and charitable giving and has displayed salient significance of the coefficient (β=3.021; p=0.001). After the inclusion of the interactions, the coefficient values for female directors remained almost the same as in Model 2. The interpretation of this log-linear model suggests that one unit increase in the percentage of female representation on the board will cause, on average, an increase in the amount of corporate donations by 3.02 units. These findings, therefore, indicate that the second hypothesis has found even greater statistical support.

Interaction effects

37

TABLE 4.

An aggregated overview of the regression outcomes and tested hypotheses

Hypothesis Description Expected

relation

Result H1a Higher proportion of outside directors will positively influence the amount of CP Positive Supported H1b Higher percentage of female directors will positively influence the amount of CP Positive Supported H2a The positive relationship in H1a will be positively moderated by slack resources Positive Not supported H2b The positive relationship in H1b will be positively moderated by slack resources Positive Not supported

Robustness checks

Robustness checks were performed in order to examine how the coefficient results of the main analysis behave after slight modification of Model 3. If the outcome of the robustness test turns out to be plausible, it can be assumed that the main model has robust structural cogency (White & Lu, 2014). The regression from the final model was modified by changing the measurement of the moderator from current ratio to net cash flows generated from operating activities. Data for the raw amount of cash flows was obtained from the annual reports of the companies, after which the values were log-transformed in order to avoid problems with skewness. Table 5. exhibits the comparison between Model 3 and the robustness test.

After Hypotheses 1a and 1b were both supported, suggesting that companies with higher representation of independent directors and female members on the board will contribute more to charitable causes, robustness checks aim to explore whether transforming the measurement of financial slack will result in the same or different interaction effects. Scarce amount of previous literature has implemented cash flows generated from operations as a measurement for financial slack. The lack of theoretical support for this measurement is among the reasons why this paper has implemented current ratio instead of the amount of cash flows. However, in a study of the relationship between international expansion and firm profitability, Liu (2010) has used cash flows generated from operations as a measurement for slack and has found that both human and financial slack can be beneficiary for the firm.

39

Discussion

This paper explored the relationship between board composition and CP. More specifically, by performing a multiple linear regression analyses, estimated by OLS, this study aimed at answering the following research questions:

 How does the proportion of outside directors affect the amount of charitable giving? (H1a)

 How does the percentage of female directors affect the amount of charitable giving? (H1b)

 How does financial slack moderate the relationship between H1a and H1b?

As it has already became clear, the empirical results have confirmed the proposed positive relation between independent directors and CP. These findings are consistent with the results of previous studies, which point out that among board members, outside directors exhibit higher interest towards the philanthropic activities of the company, rather than towards its economic goals (Ibrahim & Angelidis, 1995; Ibrahim, Howard, & Angelidis, 2003). The results of Hypothesis 1a are also in line with the findings of Johnson and Greening (1999), Zahra et al. (1993) and Zhang et al., (2013), that greater outside representation is positively related to both CSR and CSP, part of which is CP. The significant results are also in line with the research of Ibrhim and Angelidis, in which they found that outside directors pursue philanthropic practices more often than inside directors. The results of this research are however in contradiction with the study of Coffey and Wang (1992). Following the agency theory perspective, they have found that agents (inside directors) are more inclined to spend money for charity than principals (the owners and their representatives).

40 their inside colleagues, outside directors are seen to be more sensitive regarding CSR decisions, while inside directors tend to pay more attention to day-to-day economic activities (Ibrahim & Angelidis, 1995).

Finally, in contradiction to agency theory, the stakeholder perspective can be also related to the results for Hypothesis 1a. Not only CP might be beneficiary to both shareholders and stakeholders, but in some cases the company owners might have non-profit interests, thus feel the obligation to contribute to the society (Wang & Coffey, 1992). Consequently, the CSR activities might become a permanent part of that particular company’s strategy (Wang & Coffey, 1992) and outside directors, as representatives of the owners, may be even more encouraged to participate in such social acts.

Regarding Hypothesis 1b, the findings indicate positive relationship between greater female representation and CP. Consistent with the study of Ibrahim and Angelidis (1995), the positive results support the notion that women are more concerned about corporate social activities and are less business-oriented than their male counterparts. Furthermore, the results are in line with the findings of Coffey and Wang (1992), Williams (2003) and Marquis and Lee (2013) who all have found significant and positive relation between boards comprised of more women and the CP levels of companies. In addition, the findings support the idea that higher representation of women in the TM is connected to various board improvements including greater sensitivity towards CSR (Williams, 2003). However, several studies have concluded that the presence of only 1 female director might not be sufficient, as minority groups are usually considered as tokens and their voices often remain unheard of (Brewer & Kramer, 1985; Lord & Saenz, 1985). Hence, when the number of female directors increases, it becomes more likely that women’s CSR demands will result in actual social contributions towards stakeholders (Kramer, Konrad, & Erkut, 2006). The findings of this paper support the above-mentioned claims as they indicate that an increase in the percentage of women reflects in an increase of the amount of charitable giving. This further implies that the contributions for the company, brought by women, become more visible when tokenism passes into normality (Erkut, Kramer, & Konrad, 2008).