Women on corporate boards: Do they advance corporate social responsibility?

Corporate boards and gender diversity

To date, research on the organizational impacts of gender diversity has been inconclusive and inconsistent (Harrison and Klein, 2007; Joshi and Roh, 2009; van Dijk et al., 2012). For instance, some studies have found that women’s integration into leadership ranks improves a number of organizational metrics related to equity, including reduced segregation, increased pay equity and improved career mobility for women (Bilimoria, 2006; Cohen and Huffman, 2007; Konrad et al., 2008; Skaggs et al., 2012). In some contexts, gender-diverse groups outperform all-male groups, particularly when group members are in frequent contact and must collaborate on a shared objective (Knippenberg et al., 2010; Zhang and Hou, 2012). Research specific to corporate boards finds that gender-diverse boards tend to be more transparent, to engage in more effective communication practices, and to be more likely to pursue innovation (Bear et al., 2010; Brown et al., 2002; Torchia et al., 2011). Other studies, however, find that gender diversity can impede board functioning by increasing intragroup conflict and competition, reducing cooperation and morale, and ultimately weakening performance (Hogg et al., 2012; Pelled et al., 1999; Williams and O’Reilly, 1998).

Although extant research has analyzed the impact of gender diversity on intragroup processes and firm-level performance outcomes, much less has sought to identify the impact of gender diversity on strategy (Adams and Ferriera, 2009). This omission is important because there is evidence that female directors may be less likely than male directors to prioritize financial performance over non-performance-related priorities. For instance, in some contexts female leaders are more likely than men to emphasize innovation, equity and fairness – long-term goals that may not translate into short-term performance gains (Brown et al., 2002; van Dijk et al., 2012). The current study seeks to advance this debate by considering the relationship between leadership diversity – in this case gender diversity on boards of directors (BODs) – and outcomes associated with CSR. These relationships have been under-explored, and this study aims to contribute to our current understanding.

Corporate boards and corporate social responsibility

The growing focus on board composition reflects mounting awareness regarding the role boards play in shaping corporate policy and practice. Traditionally, the role of boards was assumed to be limited to appointing top executives, reviewing the performance and compensation of top executives, and approving annual budgets. Yet recent research has underscored the vital role boards play in shaping corporate strategy and driving strategic change (e.g. Oehmichen et al., 2017).

Scholars increasingly recognize that board directors provide firms with a vital source of market-wide information and intelligence that often shapes firms’ policies and practices (McDonald and Westphal, 2013; Shropshire, 2010). Furthermore, scholars have increasingly documented the impact of corporate boards on shaping a firm’s commitment to CSR (Hung, 2011). In part, a response to shareholder activism with regard to issues such as labor rights, human rights and environmental protections, boards have increasingly been charged with proposing and monitoring strategic decisions related to CSR (Altschuller, 2011). A survey of firms in the S&P 100 found that a significant majority of firms has designated board committees dedicated to CSR concerns (Calvert, 2010). And indeed, extant research suggests that gender diversity on the board is associated with stronger environmental records and greater support for LGBT-friendly policies (Cook and Glass, 2016; Glass et al., 2015).

Gender diversity and corporate social responsibility

As noted, several scholars have found that gender diversity among leaders significantly affects the functioning of an organization (Bilimoria, 2006; Post et al., 2011; Zhang and Hou, 2012). Organizational scholars have identified at least three mechanisms that may lead female directors to display a stronger commitment to CSR: gender role expectations, gender-based functional differences, and gender discrimination. Each of these perspectives situates gender differences within organizational contexts. Rather than indications of essential or natural differences between men and women, gender differences emerge out of the structural and cultural limitations and biases that exist within organizations and that are rooted in gender stereotypes and norms (Acker, 1990). Here, we seek to understand how these pressures may influence the outlooks of female directors in ways that influence their leadership priorities.

First, organizational scholars have identified a range of gender-specific pressures and expectations that limit the scope of acceptable ways women can enact leadership (e.g. Eagly et al., 2012). Gender role expectations posit that women are kind, friendly and community-oriented. Yet in leadership roles, those characteristics often undermine women’s ability to be viewed as competent and capable (e.g. Schein, 2001; Schein and Davidson, 1993); conversely, if women violate gender expectations by demonstrating competitiveness, toughness or assertiveness, they are often evaluated negatively (Catalyst, 2007; O’Neill and O’Reilly, 2011). Several scholars have posited that the pressures arising from this ‘double-bind dilemma’ lead to differences in the priorities, behaviors and styles women and men bring to leadership roles (Adams and Funk, 2009; Eagly et al., 2003; Jaffee and Hyde, 2000). For instance, there is evidence that female leaders are more likely than men to demonstrate a commitment to equity, fairness and collaboration (Eagly et al., 2003; Eagly and Johannesen-Schmidt, 2001; Konrad et al., 2006). There is also evidence that gender roles and expectations lead women to place a greater emphasis on meeting the needs of diverse stakeholders, including employees and community members (Adams and Funk, 2009; Bilimoria and Wheeler, 2000). Thus, although female leaders are expected and encouraged to participate in arenas such as corporate monitoring and public affairs, they are less likely to serve in areas such as finance and operations (Adams and Ferriera, 2009; Dalton and Dalton, 2010; Peterson and Philpot, 2007). To the extent that CSR reflects leaders’ commitment to equity and stakeholder concerns and exposure to relevant strategic areas, gender role expectations may lead to a greater demonstration of CSR commitment among women than among men.

Second, owing to organizational constraints on their mobility, female directors have different functional backgrounds and career trajectories than male directors. For example, because of higher standards and greater levels of scrutiny, women are much more likely to need stronger credentials to advance into leadership roles (Weyer, 2007). As a result, female corporate leaders are more likely than their male peers to have Master of Business Administration (MBA) and doctorate degrees, which are associated with greater exposure to innovative or cutting-edge business practices, including those associated with CSR (Hillman et al., 2002; Singh, 2008; Zweigenhaft and Domhoff, 2011). In fact, corporate leaders with MBA degrees are more likely to prioritize transparency and to enjoy greater influence over corporate practice (Walls and Berrone, 2017, forthcoming). Female leaders’ mobility patterns are also more circuitous than men’s (Eagly and Carli, 2007), resulting in female leaders’ greater exposure to a range of organizational contexts. For example, female leaders are more likely than men to have career experience in non-profit, philanthropic and community organizations. In turn, that experience increases their awareness of and concern for a broad range of stakeholders (Hillman et al., 2002; Zweigenhaft and Domhoff, 2011). In fact, relevant experience in such organizations is associated with a stronger commitment to sustainable environmental practices and, at least potentially, broader stakeholder concerns (Walls et al., 2012). Finally, compared with men, female directors are more likely to hold directorships on multiple firms’ boards, which is known to increase awareness of innovative practices within and without the focal industry (Shropshire, 2010). Taken together, women’s educational credentials, greater exposure to community organizations, and status on multiple boards is likely to increase their commitment to policies associated with CSR.

Finally, a great deal of evidence suggests that the advancement into leadership ranks within work organizations is highly gendered (Eagly and Carli, 2007; Konrad and Cannings, 1997). Female leaders’ career trajectories tend to be characterized by scrutiny, bias, negative stereotypes and discrimination (Eagly and Karau, 2002; Kanter, 1977). Not only is the competence and work commitment of female leaders challenged, but they also face scrutiny of their bodies, appearance and dress (Konrad and Cannings, 1997; Mavin and Grancy, 2016). In fact, women are particularly likely to report higher levels of harassment, bias and discrimination in male-dominated jobs, suggesting that the higher women advance in corporate firms, the more bias they are likely to experience (Konrad et al., 2010). Female leaders’ greater exposure to bias and discrimination may endow them with a greater commitment to fairness, equity and transparency (Eagly and Carli, 2007; Raeburn, 2004). Indeed, there is evidence that female leaders, as compared with men, are more likely to prioritize fairness and equity with regard to organizational policies and practices (Adams and Funk, 2009; Barron, 2003; Skaggs, 2008; Torchia et al., 2011). To the extent that a greater commitment to egalitarianism is associated with greater investments in policies aimed at ethical practices and addressing the needs of vulnerable populations (McCabe et al., 2006; Siegal et al., 2011), female leaders may be more likely to champion CSR-relevant practices.

Taken together, these factors – gender role expectations, functional differences and experience of discrimination – suggest that female leaders may bring different views, perspectives and priorities than men to their roles as board directors. In particular, these factors suggest that female directors may be more likely to prioritize CSR policies and practices and, as a result, the presence of female directors will improve a firm’s CSR record:

Token theory

One of most influential theoretical paradigms in organizational demography, token theory, predicts that women’s efficacy in organizations is limited by their numerical representation. Originally formulated by Kanter (1977), this theory defines solos as single members of a social group in a work setting, and tokens as members of a social group that are significantly underrepresented. By virtue of their numerical rarity, solos and tokens experience heightened visibility, performance pressures and negative evaluation bias. Their presence often induces the dominant group to heighten the boundaries between the in-group and the out-group, leading to the exaggeration of differences among groups, and exclusionary practices and behaviors targeted at members of the out-group (Kanter, 1977: 201).

According to Kanter, ‘tokens are, ironically, both highly visible as people who are different and yet not permitted the individuality of their own unique, non-stereotypical characteristics’ (Kanter, 1977: 211). These constraints impede the ability of solos and tokens to fully contribute to their organizations. First, heightened scrutiny and performance pressures often inspire reticence and restraint in solos and tokens, who may fear reprisal for violating existing norms. Fear of retaliation for being a ‘troublemaker’ can limit solos’ and tokens’ efforts to advance organizational change (Kanter, 1977). Second, bias often limits the support and resources solos and tokens can mobilize to advance their strategic vision. Members of the dominant group often overlook the achievements of solos and tokens (a phenomenon Kanter terms the ‘token eclipse’). Tokens often respond through assimilation with the dominant group or through social invisibility – two strategies that limit the efficacy of their unique voice, vision and values. Finally, solos and tokens are often isolated within their work settings, limiting their ability to build coalitions of support for their proposals. Isolation takes the form of exclusion from key networks, formal and informal socializing activities, and mentoring opportunities.

Empirical research on the pernicious effects of tokenism within work organizations provides support for token theory. There is evidence, for instance, that women token leaders experience lower status and less influence than their male peers, and that women are more likely to be denied key organizational resources necessary to lead effectively (Eagly and Karau, 2002). There is also evidence that female directors are less likely than men to receive mentoring about the prevailing norms and expectations of corporate executives, which leads to weaker performance, limits their ability to serve on multiple boards and, by extension, limits their influence over firm or industry policies and practices (McDonald and Westphal, 2013).

While Kanter predicted that the pressures facing solos are most extreme, she also posited that two members of a minority group are insufficient for overcoming the pressures associated with token status. According to Kanter (1977: 987), ‘two … is not always a large enough number to overcome the problems of tokenism and develop supportive alliances’. Thus, drawing on token theory, solo and token female directors will likely be limited in their ability to impact organizational practice so as to advance CSR goals.

Critical mass theory

Critical mass theory seeks to identify the representational thresholds necessary to overcome the limitations associated with solo and token status (see Dahlerup, 2006, for a review). Whereas token theory focuses on the limitations to women’s agency when they represent a numerical minority, critical mass theory seeks to identify the representational threshold necessary to overcome those limitations.

Token theory posits that as women’s representation moves from a skewed to a tilted group (i.e. from a group where women are solos or tokens to a group where women enjoy sufficient though not equal representation), women will be less subject to isolation, scrutiny and negative bias, and more capable of engaging in non-conforming pursuits without fear of reprisal (Kanter, 1977: 966). As a result, when women are numerically represented beyond solo or token status, groups are more likely to benefit from the presence of women’s perspectives generally and from women’s agency specifically.

Others have built on this perspective, arguing that not only are numerical minorities more efficacious in tilted groups compared with skewed groups, but also that members of the minority and majority groups are better able to cooperate on shared tasks in more balanced settings (Bear et al., 2010). In more balanced settings, numerical minorities are more likely to be viewed as individuals rather than members of a group, are more likely to enjoy the trust and support of members of the majority, and are more likely to exert influence over group outcomes (Bear et al., 2010; Erkut et al., 2008; Etzkowitz et al., 1994).

Critical mass theory has received some support in analyses of board composition and firm performance outcomes. For instance, Torchia et al. (2011) find that the presence of three or more women on the board increases firm innovation. And in their analysis of German firms, Joecks et al. (2013) found that the presence of three or more women on the board increases returns on equity. While these studies suggest that gender diversity on boards can enhance performance, we seek to advance this field by analyzing board composition across a range of outcomes associated with CSR, including community outreach, governance, product development and environmental sustainability. Specifically, we test the following hypothesis:

Procedure

Our analysis relies on two merged datasets: a unique, author-constructed dataset and the Kinder, Lydenberg, Domini, Inc. (KLD) dataset. The author-constructed dataset includes all CEOs and BODs for Fortune 500 firms for the years 2001–2010. The list of Fortune 500 firms over the 10-year period was collected from CNN’s money website. ² CEOs’ and BODs’ biographical information was collected using several reference websites, such as edgar.sec.gov, investing.businessweek.com and people.forbes.com, among others. Company websites were also used to gather this information. Firm data of return on assets (ROA) and total number of employees were collected using Computstat. Computstat is a research database available through Wharton Research Data Services (WRDS).

The other dataset, KLD, is also available through WRDS, and provides CSR data of publicly traded companies such as governance, community, product and environmental issues. Researchers have described KLD as ‘the de facto research standard at the moment’ for social research data (Waddock, 2003: 369) and the ‘largest multidimensional CSP [corporate social performance] database available to the public’ (Deckop et al., 2006: 334). Walls et al. (2012) suggest that KLD data minimize social desirability bias in that the data are not based on an organization’s own self-report but rather on objective measures that are gathered for and used by investors. KLD raters use internal sources of the firm such as annual reports, regulatory filings and proxy statements, among others (Waddock and Graves, 1997). External sources such as articles from Fortune magazine, Wall Street Journal, Business Week magazine, and other press materials are also used by the KLD raters to assess particular attributes to the firm (Waddock and Graves, 1997). Although the exact items used to determine the attribute level for each firm is proprietary to KLD, the validity and reliability of those attributes have been substantiated by various researchers (Hart and Sharfman, 2015; Mattingly and Berman, 2006; Sharfman, 1996; Waddock and Graves, 1997).

As noted by Waddock (2003: 372), ‘social research of all kinds is a human activity undertaken by a community of researchers and scholars interested in achieving the best measures possible for the underlying constructs’. The concept of CSR has evolved over time from being simply the discretionary actions of a firm’s responsibility to a more broadly accepted view reflecting a wide range of responsibilities that affect both the stakeholder and the environment (Waddock, 2003). KLD data have shifted over time to accurately reflect the multidimensional view of CSR (Waddock, 2003). Indeed, KLD measures are aligned with best practices for CSR. This is evidenced in Business Ethics Magazine’s list of ‘100 Best Corporate Citizens’. This ranking has become recognized nationally as a signal of best practices in the realm of CSR (CSR Wire, 2005) and is derived from the KLD measures.

In our examination, we use the KLD measures of community strengths, corporate governance strengths, product strengths and environment strengths as our dependent variables. In the Business Ethics Magazine list, it uses the four measures mentioned as well as employee relations, human rights and diversity. Because there were substantial amounts of missing data, we were not able to use employee relations or human rights strengths in our analysis. And the diversity measure was omitted given the conflation of it with our predictor variable of women on the BOD.

The predictor and control variables come from the author-constructed dataset. Our control and predictor variables have been lagged in order to mitigate possible endogeneity issues of simultaneity. In other words, our control and predictor variables are from the year prior to the assessment of the outcome variables. Thus, the control and predictor variables used are from 2001 to 2009 and our outcome variables are from 2002 to 2010. Our panel data provide for a longitudinal analysis of the same firms measuring several points in time (Finkel, 1995). Although our panel is an unbalanced panel given not all firms were on the F500 list each year, the majority of the firms are represented for all nine years of analysis. Only 12% of the firms have fewer than four years represented, and only two companies have just one year in the analysis.

Though we have a longitudinal design with our panel data, we are suggesting associations and not causation. Many firms remained constant with regard to the number of women on their board and, as such, causation cannot be determined. Although we caution against claiming causality, there are several examples that are suggestive of a causal effect of women being placed on the board and the corresponding shift in one or more of the examined outcome variables. For example, Caterpillar went from zero women on its board to one woman, and the following year its environmental strengths score increased. Honeywell went from one woman on its board to no women on its board, and the following year its community strengths score decreased. Raytheon went from one to two women on its board, and its environmental strengths increased. FMC Technologies went from zero women to one woman to two women. And not until the second woman’s inclusion did its environmental strengths increase. Texas Instruments went from one to two women on its board and then to four women. Once the board reached four women, both the corporate governance and the community strengths measures increased. And Hormel Foods started with one woman and increased over the nine years up to five women. Not until it had all five were the outcomes influenced. But at that point, community strengths jumped from zero to two, corporate governance from zero to one, and environment from zero to three. Those are just a few examples out of the dataset that illustrate the connection between women on the board and a firm being recognized for positive CSR.

Dependent variables

Community strengths. Community strengths is an index measure that includes the following: charitable giving and giving that supports nonprofit organizations, charitable giving outside the US, housing and education support, volunteer programs and other positive community activities. Each of the sub-items is coded as 1 for yes, and the index is a summation of those items. For further detail regarding the dependent variables, refer to Getting Started with KLD Stats and KLD’s Ratings Definitions (KLD, 2006).

Corporate governance strengths. Corporate governance strengths is an index item that includes the following: assessments regarding a positive corporate culture, leadership on public policy issues, social and environmental reporting measures, effectiveness in its communication processes, ownership of another company that has social strength, and limiting compensation for its top management or BOD. Each of the sub-items is coded as 1 for yes, and the index is a summation of those items.

Product strengths. Product strengths is an index item that includes the following: a recognized long-term quality program, leader for research and development, known for bringing innovative goods or services to the marketplace, mission to provide some goods or services to the disadvantaged, and its goods or services have recognizable social benefits. Each of the sub-items is coded as 1 for yes, and the index is a summation of those items.

Environmental strengths. Environmental strengths is an index measure that includes the following: known for bringing goods or services to the marketplace that have unquestionable environmental benefits, recognized for strong pollution prevention programs, user of clean energy, major player in recycling either as a facility or a user of recycled materials, and communicates environmental best practices within the organization. Each of the sub-items is coded as 1 for yes, and the index is a summation of those items.

Independent variables

Number of women on the BOD. Given our large dataset, we were able to compare groupings of organizations based on the number of women on the BOD. The categories examined are an all-male board compared with a board with women, one or two women on the board compared with an all-male board and compared with a board with three or more women, and an analysis examining firm outcomes as the number of women on the board increases. Nineteen percent of our firm observations had no women on the board; 37% had one woman; 29% had two women; 10% had three women; three% had four women; 1% had five women; .3 % had six women; and one firm, TIAA Cref, for two years had eight women on its board.

Control variables

The control variables used are the size of the firm as determined by the log of the number of employees, current financial performance as determined by return on assets (ROA), the total number of board members, average age of the board members, CEO age and CEO/Chair duality. It is important to control firm size because it is directly linked to firm visibility and accountability (Arthur and Cook, 2009). The firm’s financial performance is also important to control as shareholder pressures may differ during lucrative versus unprofitable times (Cook and Glass, 2011). Total board size is important to control as it may be representative of the overall importance attributed to the board. Given societal shifts toward some of the outcome variables, controlling for average board age is warranted as well as the age of the CEO (Cook and Glass, 2016). CEO/Chair duality suggests potential for greater influence and power of the CEO, and as such should be controlled (Finkelstein and D’Aveni, 1994).