Abstract
The need for greater diversity among organizational leaders and directors remains a challenge for organizations within the third sector, and beyond. This study examines diversity through a critical mass lens; that is, we examine an alternative approach to understanding the relationship between the ethno-racial composition of boards of directors and their perceived ability to engage stakeholders, improve organizational responsiveness, and effectively manage fiduciary responsibilities. Our study, drawing on a survey of 247 boards, clarifies the need for a critical mass approach to leadership diversity by highlighting the uneven impact of diversity on performance demonstrated by periods of accelerating and decelerating effect. We find that boards achieving a critical mass of ethno-racial diversity improved board performance among three governance activities—fiduciary performance, stakeholder engagement, and organizational responsiveness—with our critical mass approach illustrating the uneven impact of diversity on performance for each governance activity.
Organizations in the third sector, and beyond, have recognized and continue to wrestle with the significant implications of the current diversity gap, which threatens to undermine the relevance, the legitimacy, and the ability of organizations to fulfill their mandates (Moore, 2000). Persistently, the nonprofit governance and leadership literature has focused on the relationship between board diversity and performance outcomes by examining how intervening processes (Fredette, Bradshaw, & Krause, 2016), practices (W. A. Brown, 2002; Buse, Bernstein, & Bilimoria, 2016), or contextual characteristics (Bradshaw & Fredette, 2013) might shape these relationships. Advocacy for greater leadership diversity is built on the premise that a broader and more diversely composed group, team, or organization improves the richness of decision-making outcomes and the ability to engage otherwise out-of-reach stakeholders and enhances understanding of market opportunities (W. A. Brown, 2005; Jaskyte, 2012). However, activating these benefits, we suggest, is an uncertain proposition as is the pacing under which these diversity-derived performance benefits are received. We argue that there is a pressing need to further explore the impact of diversity in the boardroom and to examine the complexities and nuances that diversity brings to governing groups.
Our work seeks to disentangle and clarify the relationship between board diversity and governance outcomes by testing a critical mass or tipping point explanation for the observed variability of this relationship. Our intuition, drawing on our work in the field, is that boards often struggle with their initial attempts at increasing the diversity of their members. That is, moving from a board composed solely of one racial or ethnic group (i.e., a homogeneous board) to one that includes a first diverse or demographically different member sometimes results in performance decline, perhaps due to the slowing of governing processes or extending debate and decision-making activities, with boards finding that they do not achieve the anticipated performance improvement often prescribed by proponents of diversity. However, in our experience, boards that persist in efforts to transform their composition to better reflect their constituents and community members do frequently achieve a breakthrough, reaching a tipping point or critical mass threshold that seems to unlock performance improvements that seemingly exceed previous expectation, suggesting that diverse boards perform beyond the sum of their parts. Oliver, Marwell, and Teixeira (1985), exploring the nature and dynamics of social change, suggest that collective action “entails the development of a critical mass—a small segment of the population that chooses to make big contributions to the collective action while the majority do little or nothing” (p. 524). The presence of a critical mass lies at the heart of our thinking as an important determinant in understanding how and when diversity influences board performance, as well as a significant factor in explaining the inconsistent impact of diversity. To test this idea, we examine three important aspects—fiduciary performance, stakeholder engagement, and organizational responsiveness—of effective board performance, testing for the impact of increasing board diversity on these relationships, and whether effects of increasing board diversity are linearly distributed (equal to the sum of their parts), or whether these effects are nonlinear, demonstrating uneven accelerating and decelerating benefit (exceeding and failing to meet sum of their parts estimates) as threshold theories predict, thereby illustrating when and why ethno-racially diverse boards under- and overwhelm performance expectations.
We examine the impact of board diversity, more specifically defined here as the percentage of ethno-racial diversity within the membership of a board of directors, on board performance to shed new light on the complicated effect of diversity on governance capabilities. We suspect that diversity is neither constant nor linear in its impact on collective decision-making outcomes, such as board-level resource allocation choices or approaches to performance monitoring or in developing stakeholder engagement strategies. We anticipate that moving from homogeneity to low levels of ethno-racial diversity will likely not generate anticipated gains and may, in some cases, generate adverse impact relative to what a linear approach would predict, just as the change in composition from near-complete to complete heterogeneity will likely do the same (Kanter, 1977). For this, and the reasons expressed above, we explore the association of ethno-racial diversity and board performance, examining cubic (S-shaped) models of the pattern of relationships, as well as underlying first-order linear (straight-line) and second-order quadratic (U- or inverted U-shaped) patterns. Our contributions lie in identifying whether a critical mass logic better specifies the board diversity–performance relationship than conventional approaches and whether a critical mass explanation can add nuance to the leadership and governance diversity literature.
Understanding the Business Case for Ethno-racial Diversity and Board Performance
Boards of directors in the nonprofit and for-profit sectors are increasingly challenged to improve their representative diversity, as increasing boardroom ethno-racial diversity has been suggested to enhance information-sharing and decision-making (W. A. Brown, 2005), innovation and change (Jaskyte, 2012), and organizational performance (Ali, Ng, & Kulik, 2014). Pressure for increased diversity often stems from external sources, including government priorities and regulations, changing demographic characteristics of volunteer and service communities such as clients, members, or customers, interorganizational, and competitive pressures, as well as the expectations of funders or donors (Bradshaw & Fredette, 2013). Internal organizational goals and strategic directives motivate the push for greater diversity as well. These can be observed in the desire to increase the representativeness of boards and leadership teams (W. A. Brown, 2002; Cornforth, 2003), the desire to generate innovative approaches and programming (Ely & Thomas, 2001), or the willingness to demonstrate the organization’s commitment to diversity and therefore its broader relevance and legitimacy (Siciliano, 1996).
Tapping the benefits that group diversity—by combining and recombining varied perspectives, experiences, understandings, and beliefs—brings to bear on the complex problems facing today’s boards and organizations forms the basis of what has been called the business case for diversity (Jayne & Dipboye, 2004; Litvin, 2006). To us, the challenge posed by governing is best reflected in the fundamental need for an organization to achieve its goals, for which equifinality reigns in the absence of singular or clear paths forward. These are strategic challenges embodying dilemma, contradiction, uncertainty, and paradox, in which a plurality of directions, approaches, choices, and opinions are equally (although uncertainly) deployable. These are not the incremental, day-to-day, routine managerial decisions faced frequently and resolved by habit, but rather those which exemplify the functional aspects of governance, including the building and sustaining of stakeholder relationships, the development and maintenance of the board’s strategic direction, the protection and mitigation of the organization’s financial risk profile, and the decision to deploy, withdraw, and redeploy organizational resources and assets (W. A. Brown, 2005; Cornforth, 2012; Gazley & Nicholson-Crotty, 2017). Diversity, and its embodied varieties of intellectual tradition and experiential understanding, contribute meaningfully to functional decision-making central to fiduciary performance (Fredette et al., 2016), as well as cultural familiarity needed for responsiveness (Buse et al., 2016).
One pillar of the business case for diversity stems from research that links increased group diversity to constructive conflict (Jehn, Northcraft, & Neale, 1999), which tends to slow decision-making processes and afford opportunities to discuss taken-for-granted practices and informal norms, all of which are believed to improve problem-solving as compared with more homogeneously composed groups. Over time, research has highlighted inconsistencies in conventional treatments of the business case perspective, demonstrating that the benefits of work-group diversity remain an enigma, sometimes supporting and sometimes undermining performance outcomes (Harris, 2014; Kochan et al., 2003; van Knippenberg & Schippers, 2007). While increasing ethno-racial diversity in the boardroom has been suggested to enhance information, knowledge, perspectives, and organizational performance (Ali et al., 2014; Ely & Thomas, 2001), whether these benefits can be achieved and whether the downside of increasing group heterogeneity mitigate them remains contested (Williams & O’Reilly, 1998). In the face of competing results and conclusions, we are left to wonder what might be confounding our understanding of the relationship between board diversity and board performance outcomes?
This research returns to examine an early intuition in the diversity literature, suggesting that group dynamics and decision-making are a result of differing degrees of majority–minority composition and that the balance of influence is subject to change based on the underlying composition of the group, or alternatively stated, outcomes are in part a product of the compositional differences of the decision-making group. In this regard, examinations of the role of gender diversity on corporate boards using various performance measures have been inconsistent (e.g., Joecks, Pull, & Vetter, 2013). Positive associations were found by Bear, Rahman, and Post (2010), Campbell and Mínguez-Vera (2008), and Mahadeo, Soobaroyen, and Hanuman (2012), while negative associations were found in studies by Bøhren and Strøm (2010), Dobbin and Jung (2011), and Haslam, Ryan, Kulich, Trojanowski, and Atkins (2010). In contrast, other studies found no significant relationship at all (Carter, D’Souza, Simkins, & Simpson, 2010; Wang & Clift, 2009), with Kochan et al. (2003) finding racial and gender diversity to have neither a positive nor a negative effect on performance or group processes. With that in mind, we examine the effect of ethno-racial diversity on three aspects of governance activity for which boards of directors are responsible—fiduciary performance, stakeholder engagement, and organizational responsiveness—which are important indicators of organizational health, to examine whether our data support a critical mass pattern in which increased board diversity coincides with increased board performance in each of our constructs of interest.
Fiduciary Performance
Studies have highlighted the benefits to board performance, demonstrating financial and social implications, of greater diversity in the presence of inclusion-focused policies and practices (Ali et al., 2014; Torchia, Calabro, & Huse, 2011). Frequently, fiduciary activities—including the oversight of legal and financial decision-making, adherence to governing bylaws and procedures, as well as the evaluation of senior organizational executives—are cited as among the principal responsibilities of nonprofit boards of directors (Bradshaw, Murray, & Wolpin, 1992; Fredette & Bradshaw, 2012). A primary reason for this is that boards act as a mechanism for collecting and integrating information to reduce uncertainty (Miller & del Carmen Triana, 2009), bringing together multidisciplinary perspectives (e.g., legal, financial, strategic, and contextual) to reconcile competing interests that impact where and how the organization secures, controls, and deploys its resources (W. A. Brown, 2005; Gazley & Nicholson-Crotty, 2017). These challenges are urgently encountered among the members of today’s nonprofit boards of directors, and in this regard, diversity of thought, insight, and perspective are impactful to the people tasked with collectively addressing these activities (Jehn et al., 1999). Indeed, this is the domain in which the promise of diversity holds the greatest potential, particularly because inclusion in these processes tends to slow the pace of activity and trigger additional consideration of past practices (Buse et al., 2016; Fredette et al., 2016).
Stakeholder Engagement
The capacity to engage diverse and traditionally marginalized community members and constituents for the purposes of fundraising (Ostrower, 2007), or expanding relevance and legitimacy through the enlargement of the organization’s client-base (Abzug & Galaskiewicz, 2001), are domains in which greater board diversity have shown value. Bradshaw and Fredette (2013) found that the demographic diversity of the community in which the organization is situated is a significant driver of board diversity, but did not find significant linkages to other aspects of the external environment. Indeed, the application of stakeholder theories to the governance of nonprofit organizations has served to further highlight the important signaling effect of leadership diversity as being representative of community interests (W. A. Brown, 2002; Cornforth, 2003), but also the ability to attract resources and demonstrate a tacit awareness market opportunities and limitations.
Organizational Responsiveness
Adapting to changing needs and updating programming to better serve the interests of current and potential clients improves an organization’s fitness and long-term sustainability (Voss, Sirdeshmukh, & Voss, 2008). Organizations attempt to do so by updating, refining, revising, reinventing, or innovating current programming and offerings to better meet the needs and interests of current and potential client communities. In this regard, many communities may have noticed football and baseball fields repurposed to accommodate growing interests in soccer (also known as football) and cricket leagues, or perhaps as urban gardens in some cases, which are largely driven by changing tastes often brought on by, but not exclusive to, greater concentrations of demographic diversity in larger urban communities. Identifying and responding to external or market-driven pressures is aided by greater familiarity with an organization’s constituents, which allows for a better understanding and incorporation of the needs and interests of diverse and traditionally marginalized community members which are often tacitly-held and who would otherwise remain underserved (Foldy, 2004).
Applying a Critical Mass Lens to the Diversity–Performance Relationship
We believe that prior specifications of the diversity–performance relationship, including our own, have been limited in their approach to understanding the implications of interdependence among board members in decision-making, particularly as it relates to the nonsequential nature of governance work, which is subject to the exercise of majority–minority voting control. Indeed, as we note above, prior studies have produced varied outcomes across a variety of contexts, leaving the field to question which pattern of results are reliable and which are noise. We question whether the noise is random, that is, whether smoothing diversity–performance relationships by plotting the least objectionable straight line through the data appropriately captures what in practice is often a very messy affair. Our sense is that the pattern of inconsistent results is not random and that the effects of group or board diversity are not smooth or constant, but rather they reflect a series of setbacks and lunges forward resulting in predictable patterns of under- and overestimation when compared with linear predictions.
One of the critical factors in understanding how and when board diversity impacts performance is a matter of the relative degree or proportion of diversity within a given composition as it is frequently the minority—whether in terms of diverse identity or in terms of voting proportion—who alter the conditions of the majority (Oliver & Marwell, 1993); that is, numbers matter in understanding how and when diversity generates negative, neutral, or positive outcomes. Blalock (1967), for example, contended that while the minority group made gains, the majority group members suffered a commensurate loss, because the majority group perceived the threat by the proportionate size of the minority group as undermining their established positions. In contrast, Kanter (1977) argued that both the minority and majority benefit from the resources the minority members bring to the organization, but that these gains became realized once there were enough minority group members to overcome dynamics of tokenism and dismissiveness. Building a “critical mass” is an important determinant in understanding how and when diversity influences board performance, as well as a significant factor in explaining the uneven impact of diversity. The intuition is that the benefits of diversity exist within a range of group variety, outside of which, too little or too much variety undershoots linear estimates. That is, that the first addition of diversity into a group does little to benefit group performance due to the limited influence in vote and marginalization of voice (Kanter, 1977). However, with the addition of each subsequent ethno-racially diverse member, these restraints diminish, and vote and voice become more influential, building to a critical mass, tipping the group or board from inertia to action, providing an accelerating benefit for performance that is underestimated by conventional modeling (Oliver & Marwell, 1993). Some have suggested that a change in the nature of the relationship occurs at three members (Konrad, Kramer, & Erkut, 2008), other say at compositions above 35% (Kanter, 1977).
Intersectionality research largely suggests that members of ethno-racially diverse communities struggle with the same dynamics of tokenism that occur in the study of gender, illustrated by the presence of compounded penalties (Anthias, 2013). We argue that ethnic and racial facets of diversity are likely to respond similarly to the dynamics of critical mass logic (Oliver et al., 1985), suggesting an uneven distribution of influence on the problems plaguing board effectiveness. We differ from more conventional treatments in organizational thinking (cf. Jia & Zhang, 2013; Konrad et al., 2008; Torchia et al., 2011), in terms of how we conceive of critical mass—not as a categorical step function, but rather as a continuous, albeit nonlinear, S-shaped function with periods of accelerating and decelerating effect (Oliver et al., 1985). The important differences here are more than a matter of nuance, they reflect the difference between an on–off dichotomy, suggesting a single quadratic threshold or tipping point effect, and a relationship for which both lower and upper inflection points bound an area of accelerating effect. We believe that there is both an early period of deceleration in the diversity–performance relationship and a subsequent period of acceleration, in which the impact of critical mass diversity is realized. We add to this, the belief that the acceleration of the board diversity–performance relationship has an upper bound in which saturation occurs and diminishing returns from additional diversity precede a final period of deceleration when the minority become a dominant majority, reverting to a new homogeneity or alternatively to fragmentation, where the difficulties of integration exceed the benefits of diversity.
D. A. H. Brown, Brown, and Anastasopoulos (2002) determined that Canadian boards in the private, public, and nonprofit sectors with three or more women are different from all-male boards when it comes to identifying criteria for measuring strategy, demonstrating that 94% of boards with a critical mass of gender monitor the implementation of their corporate strategy. Torchia and colleagues (2011), drawing on the work of others (Konrad et al., 2008), tested boards with one, two, and three women directors to determine whether reaching a critical mass enhanced firm innovation, finding that increasing from one to two women did not have an impact on the level of firm innovation. However, when the number of female directors was increased to a critical mass of three women, they saw a positive and significant increase in the level of firm organizational innovation. The value in achieving a critical mass of women on corporate boards is the ability to provide alternative perspectives, expand board discussions, and enable a greater stakeholder focus, raising more difficult questions about tough issues, and positively influencing board processes that enhance the overall quality of governance (Konrad et al., 2008). Other studies show that women impact corporate ethical, strategic, and social-responsibility practices (Nielsen & Huse, 2010), tending to be responsive to a broader spectrum of stakeholder concerns.
More recently, Joecks and colleagues (2013) found evidence that gender diversity initially negatively affects performance and then, only after a critical mass of about 30% women (or in their study, about 3 women) has been reached, does firm performance increase, suggesting a U-shaped relationship between gender diversity and firm performance; finding, in accordance with critical mass logic, more diverse boards outperformed uniform and less diverse boards. In studies examining representative bureaucracy in the public sector, relationships among composition diversity of public servant leadership (i.e., teachers and educational administrators) and performance outcomes of student and subordinate constituent groups give evidence of critical mass relationships, demonstrating nonlinear associations (Meier, 1993; Meier, Wrinkle, & Polinard, 1999). While other research focused less-explicitly on critical mass, they too have noted the importance of coalition building (Konrad et al., 2008; Lortie-Lussier & Rinfret, 2002), the role of momentum (Ali et al., 2014; Kanter, 1977), and the impact of structural solutions (Kogut, Colomer, & Belinky, 2014; Torchia et al., 2011). Here, we hypothesize that the impact of board diversity on board performance is amplified by the presence of a critical mass of ethno-racially diverse members.
Research Design, Method, and Analysis
We test our hypotheses with hierarchical regression techniques using survey data collected as part of an outreach initiative exploring ethno-racial diversity among boards of directors in organizations spanning the arts, sports, and environment. The Greater Toronto Area (GTA) is composed of slightly more than 6.5 million residents, with approximately 51.5% self-identifying as a member of a visible minority community (Statistics Canada, 2016). The term “visible minority” is defined in Canadian employment law, used in the collection of census information, and is a term in common usage and understanding among our respondents. Visible minority members are defined as “persons, other than Aboriginal peoples, who are non-Caucasian in race or non-white in colour” distinguishing among ethno-racial categories including Aboriginal, Arab, Black, Chinese, Filipino, Japanese, Korean, Latin American, South Asian, Southeast Asian, West Asian, and White (Statistics Canada, 2016). Respondents identified only 16.2% of the board members included in this study as visible minority members, slightly better than found in other research conducted in this regional context (Cukier, Yap, Aspevig, & Lejasisaks, 2011), but lower than what would be anticipated relative to the underlying demographics of the surrounding population. This is indicative of a gap between organization leadership and the constituencies these organizations serve.
Data were collected using a brief online survey focused on developing an understanding of the participation and perceived contribution of visible minority members to the leadership and governance of nonprofit organizations engaged in the sectors of arts, sport, and the environment. Our approach to sampling was purposive in the selection of arts, sport, and environmental dimensions of the nonprofit sector, domains in which we had limited understanding of the state of leadership diversity (Kalton, 1983). A leading civic rights and a social justice foundation aided by identifying and contacting 903 organizations, with 269 organizations responding, of which 247 provided enough data to be considered complete (response rate = 27.4%). Research has suggested that web-based survey response rates often differ from more conventional survey approaches (Cook, Heath, & Thompson, 2000; Shih & Fan, 2008). While lower than we would have preferred, the sample provides reasonable representativeness, given our sample size and target population (Baruch & Holtom, 2008; Cook et al., 2000). Introductory and follow-up communication with the targeted organizations has been demonstrated to enhance the quantity and quality of collected data (Shih & Fan, 2008). With the assistance of two well-regarded umbrella organizations, we were able to engage in such communication. Further efforts were taken to protect survey access control, by providing respondents with a direct link to our survey and using organization names and/or respondent contact information to confirm that there was no redundancy. In prior research, these efforts have been shown to ensure data quality and integrity (Fan & Yan, 2010).
Participants and Procedures
For this study, 247 boards in the Greater Toronto Area were surveyed, representing 2,789 board members, 83.8% of whom respondents identified as White. Of the nonprofits, 55.5% had at least one visible minority member, but only 25.5% of the boards surveyed had at least three. The United States has a similar majority–minority board member divide with 16% of board seats filled by those who identified as racial or ethnic minorities; however, 27% of boards identify as having all White members (BoardSource, 2016). Organizations represented in this survey were primarily well-established with an average budget of CAD$1,437,122 (SD = CAD$3,633,148) and an average number of 12.5 full-time staff. The organizations had an average age of 30 years (SD = 25 years). Of the three broad organization types, 57.1% were involved in arts and culture, 26.3% sports and recreation, 6.5% environmental, and 10.1% fell into other sector categories. Only senior leaders in the organizations, which included Chief Executives, Executive Directors, Board Chairpersons, and Directors, participated in the survey. These respondents identified as being with their organizations for an average of 10.4 years and having served in their current roles for an average of 7 years. The respondents were 56.7% White and 59.3% female.
Variables and Measurement
Diversity
Diversity was derived from a board-level measure of the membership composition based on Canadian census categories of ethnic origin and visible minority status. The proportion of majority–minority members was reported as a percentage score (the number of visible minorities as a percentage of the total number of board members). This percentage score forms the basis for nonlinear quadratic and cubic testing (board diversity-squared and board diversity-cubed). The composition of board membership was dominantly White, with boards having an average of 1.79 ethno-racial groups represented (SD = 1.53). According to Ostrower and Stone (2005), contemporary nonprofit organizations continue to find ethno-racial diversity a challenge. In our view, the leadership and governance of nonprofit organizations is also an important societal linkage, providing communities with unique opportunities to share decision-making power and participate in the shaping and expression of community values and priorities.
Fiduciary performance
We examine the capacity of the board to execute its fiduciary responsibilities, a core aspect of governance which centers around ensuring the organization’s legal, financial, and competitive health into the future (Green & Griesinger, 1996; Ostrower & Stone, 2005). We operationalized effectiveness by asking the informants to assess the contribution of diversity to board performance in terms of their perception of the effectiveness of the board’s ability to improve governance processes, understand and address legal obligations, evaluate senior executives, and monitor financial health. Fiduciary performance was measured with four items on a 5-point Likert-type scale (Cronbach’s α = .922).
Stakeholder engagement
Our interest in stakeholder engagement largely stems from opportunities and concerns related to the acquisition of resources and funding. We operationalized stakeholder engagement by asking respondents to assess the contribution of diversity to their board’s ability to shape strategic direction, enhance fundraising efforts, and contribute to the management of stakeholder relationships. The stakeholder engagement scale was comprised of the three items on a 5-point Likert-type scale (Cronbach’s α = .906).
Organizational responsiveness
Identifying and meeting the evolving programming needs to better serve the interests of current and potential clients improves an organization’s ability to achieve long-term viability (Voss et al., 2008). We assessed the contribution of diversity by asking informants to assess their board’s capacity to advance and improve service levels. Measurement items focused on areas such as community responsiveness, changing service design and delivery methods, ideation and decision-making, and progress toward a more inclusive organization. Organizational responsiveness was measured with six items on a 5-point Likert-type scale (Cronbach’s α = .962).
Controls
This sample was drawn from nonprofit organizations in the well-documented ethno-racially diverse Greater Toronto Area (Statistics Canada, 2016). We included in our analysis three organizational characteristics considered significant to our analysis as control variables. The first control variable draws from the type of nonprofit organizations surveyed, coding the sector of activity in terms of artistic, sport, or environmental. The second variable measures organization age by number of years in existence. The age of an organization has been shown to influence how organizations approach diversity (Bradshaw & Fredette, 2013; Bradshaw, Murray, & Wolpin, 1996). The final control variable we included is the organization’s budget, as budget size might well (de)prioritize diversity in light of other challenges. Boards that are predominantly White have been found to be most frequently local or regional in geographic scope, and smaller in terms of budget and size (Ostrower, 2007). By including these control variables, we aim to mitigate confounding factors while improving our models’ accuracy and practical significance.
Data Analysis and Hypothesis Testing
The means, standard deviations, and zero-order correlations among the variables are presented in Table 1. The correlations demonstrate that moderate, significant correlations were found between our dependent variables fiduciary performance, stakeholder engagement, and organizational responsiveness. Correlations are also present between organization age and board diversity as well as between organization age and budget. All variables apart from organization annual budget had skew statistics smaller than±2, and all are within acceptable levels for multiple regressions (not surprisingly, squared, and cubed terms presented skew statistics of 2.188 and 2.543, slightly above the ±2 threshold). In our modeling, the variables most skewed are included as predictors and controls, where higher levels of nonnormality are expected for regression, with our errors meeting the assumption of normality (McClelland & Judd, 1993).
Descriptive Statistics and Correlations of Measures.
Correlation is significant at the .05 level (two-tailed). **Correlation is significant at the .01 level (two-tailed).
Results
To examine our hypotheses, we first entered the control variables, followed sequentially by our main effects, producing a series of hierarchical multiple regression models in what Cohen, Cohen, Aiken, and West (2003, p. 210) describe as a build-up approach to modeling curvilinear relationships. Recognizing that we would be creating higher-order terms from our predictor variable and the meaningfulness of interpreting a zero or “no board diversity” condition, we chose not to mean-center our diversity predictor, recognizing the potential for multicollinearity (Cohen et al., 2003). To test the linear, quadratic, and cubic effects of board diversity on each of our dependent variables, quadratic and cubic terms derived from the diversity predictor were introduced sequentially to our modeling with the final model including each of the lower-order variables to partial out variance and improve interpretability of our analysis (Cohen et al., 2003, p. 200). Each step was repeated for each of the board-level dependent variables as Cohen and colleagues suggest (Table 2), with relationships illustrated in Figure 1. All three S-shaped curves demonstrated some consistency with respect to inflection points, with critical points occurring with minima of approximately 9% and maxima of approximately 70% diversity (8.95% and 70.46% for fiduciary performance, 8.65% and 69.87% for stakeholder engagement, 11.86% and 72.38% for organizational responsiveness, respectively). This means that on average, the low point in the independent variable/dependent variable relationships occur around the addition of the first ethno-racially diverse board member (assuming a 12-member board), with the peak occurring just after the addition of the eighth ethno-racially diverse member, but of course, board size varies somewhat among organizations.
Results of Hierarchical Multiple Regressions.
Relationships among ethno-racial diversity and dependent variables.
We test H1 using a four-step build-up model with the fourth model containing our hypothesized critical mass or third-order cubic term and each subordinate lower-order term. We began by introducing our control variables in Step 1 of our modeling (Table 2), following which we examined our linear predictor, the percentage of ethno-racial board diversity, for a positive association with (i) fiduciary performance, (ii) stakeholder engagement, and (iii) organizational responsiveness in Step 2. We find partial evidence supporting a linear relationship with (i) fiduciary performance, as neither of the other models reached statistical significance and are therefore uninterpretable. In Step 3, the quadratic or second-order predictor was introduced. Here, statistically significant relationships were revealed for (i) fiduciary performance and (ii) stakeholder engagement (model reaches marginal significance with p = .057), with (iii) organizational responsiveness remaining nonsignificant and therefore uninterpretable. In Step 4, we test the hypothesized critical mass relationships among the criterion variables (i) fiduciary performance, (ii) stakeholder engagement, and (iii) organizational responsiveness and the percentage of board diversity, such that each relationship reflects a third-order or cubic function with an S-shaped curve accelerating between the inflection points. Here, we find statistically significant results supporting the predicted cubic relationship between board diversity and each of our criterion variables, H1: (i) B = −13.740, p = .007; (ii) B = −11.550, p = .029; (iii) B = −11.568, p = .033, although the model for (iii) reaches marginal significance with p = .098. Consistent with the interpretation of interaction terms, higher-order effects override the interpretability of lower-order effects (Cohen et al., 2003), illustrating the predicted patterns of acceleration between the inflection points as well as a waning or deceleration at very high levels of board diversity. The presence of significant higher-order terms, even in the absence of linear ones, is particularly interesting as they reveal otherwise unobserved or overlooked associations in the data, as we discuss below.
Discussion of Findings
Our findings highlight a number of important dynamics associated with increasing relative board diversity as it pertains to governing toward solutions in the nonprofit sector. First, governance is not simply one set of activities requiring one set of capabilities (Gazley & Nicholson-Crotty, 2017); it is instead rather open-ended, at times paradoxical, and frequently facing crises which are both abrupt and discontinuous. In our study, having greater ethno-racial diversity—with its commensurate variety of experience, thought, and perspective—was imminently helpful in some domains (fiduciary performance), while less-obviously so in others (stakeholder engagement and organizational responsiveness). This is somewhat surprising, in that stakeholder-based arguments tend to suggest that diversity should help organizations better reach otherwise unrepresented constituencies and to better adapt to the interests and concerns of these clients (W. A. Brown, 2002; Cornforth, 2003). Equally curious, however, are the strength and dynamics of building toward a critical mass of board diversity, even in the absence of first-order effects. Noteworthy is the regularity with which the criterion–predictor relationships reach inflection points relative to changes in board diversity, demonstrating a change in the strength and direction of the criterion–predictor association.
Our results speak to the difficulties in recognizing when diversity might improve governing outcomes, and how much organizations should invest in the development of their own business case for diversity to avoid the potential pitfalls and activate the potential benefits that leadership diversity offers. We aimed to confront some of the resistance impeding greater participation of members of diverse and traditionally marginalized communities in the leadership and governance of nonprofit organizations. Indeed, even in a major North American metropolis where visible minority groups compose nearly half of the population, only about 16% of the board members in our study were of visible minority status. Our intuition is that achieving a critical mass is a precursor to creating more inclusive and ultimately more equitable boardrooms, because it affords opportunities to exercise collective influence and motivate change. Our findings are not simply about the business case for diversity and the role of difference in assumption or experience but speak to shifts in control that stem from the balance of power held in check by majority–minority composition. Understanding boards as centers of power provides insight into the benefits of greater diversity beyond what other perspectives of the business case would suggest across considerable portions of the diversity–performance curve.
This examination of visible minority members of nonprofit boards allows us to recognize the need for greater numbers of participants, bringing into the foreground the impact that increases in board diversity hold for governance effectiveness, which to date has largely been thought of in terms of discontinuous steps or as categorical-dichotomies in their impact (Kanter, 1977; Konrad et al., 2008). Our approach to understanding the impact of board diversity on governance outcomes offers insight by demonstrating the bandwidth of effect that increasing diversity to a critical mass provides (Oliver & Marwell, 1993). In our modeling, early performance losses give way to a period of accelerated gains, which later decelerate with an apparent diminishing return from further diversity. This study holds potential for understanding the need for a critical mass of visible minority members to serve on nonprofit boards to positively impact board governance and performance. Achieving critical mass may offer the opportunity for transformational change, legitimization of critical mass leadership, and may be seen as critical to nonprofit sector success. Achieving these benefits might begin with increasing visible minority representation in nonprofit leadership positions but depends on evolving into an organization that values leading and governing together, an outcome which might well emerge as boards of directors accelerate toward critical mass diversity.
While previous research has illustrated that the risk of performance decline and destructive conflict tend to increase in association with diversity (van Knippenberg & Schippers, 2007; Williams & O’Reilly, 1998), our results call attention to the pattern of effects of critical mass diversity as an important mechanism in reshaping how we understand the diversity–performance relationship within nonprofit boards of directors. Prior research has demonstrated the importance of policy (Bradshaw & Fredette, 2013), practice (Buse et al., 2016), and process in meaningfully engaging members of diverse and traditionally marginalized communities in the social and functional aspects of governance (Fredette et al., 2016). Of serious concern are the potential risks of generating conflict among board members arising from greater inclusion in the decision-making and governance of the organization, as these appear to be impactful on the performance and viability of the board. At the same time, the risks to relevance and legitimacy from failing to recruit, include, and engage talented visible minority members are also quite real, and intervening approaches have been shown to attenuate, and in many cases, mitigate the presumed risk of performance decline, thereby avoiding the need to trade off effectiveness in exchange for building more diverse composition.
Limitations
Our study is prone to the concerns found in many cross-sectional surveys, including single method and perceptual biases in which a single informant per organization, at a single point in time, responds to questions that are subject to desirability biases such as self-rating performance. For example, respondents may perceive more or less diverse boards as contributing to higher or lower performing based on beliefs about their composition. Alternatively, informants may make judgments about performance outcomes and attribute them to the composition of the group-based implicit or explicit biases. In addition, our measure of board diversity is an imperfect one because it could be seen to assume equivalence among visible minority board members and may be sensitive to the effects of board size. Similarly, it does not distinguish between compositions that are multicultural and it fails to capture the intersectional effects of multiple facets of diversity.
Some may comment that our study is limited to narrow subsectors of nonprofits and grounded in contextual and temporal influences that may limit the generalizability of our findings. As we have noted, our sample was drawn from a significantly multicultural community, and while cities across North America tend to be more diverse than rural communities, the GTA has frequently been noted for its acceptance of multiculturalism, which may be different than in other jurisdictions.
Directions for the Future
Our sense is that there is a need for greater exploration of the underlying dynamics of change through the lens of “critical mass” thinking, and with respect to leadership diversity and inclusion, we need to better understand the mechanisms underlying change as a function of critical mass. Is it simply a numbers game with respect to majority voting or is there an accompanying cultural change component? Further investigation is warranted and revisiting past results would be a welcome first step, given the “messier reality” of diversity (Klein & Harrison, 2007). Attempts to replicate our results across a variety of settings and groups, inside the boardroom setting and beyond, would clarify the generalizability of our findings, both as it relates to our measurement approach (i.e., the third-order cubic specification), and to our pattern of results within and across sectors of the nonprofit community. We encourage others to return to the deeply held assumptions in our field to push our understanding forward.
Achieving critical mass diversity may offer the opportunity for transformational change, impact board governance and performance, and aid in nonprofit sector success. These benefits begin with increasing visible minority representation in nonprofit leadership positions but depend on evolving into an inclusive organization that values leading and governing together.
Footnotes
Acknowledgements
We gratefully acknowledge the support of The Maytree Foundation, CivicAction, and the Social Science and Humanities Research Council of Canada. Thanks also to the many thoughtful participants of an early morning session of the Governance section of ARNOVA 2017, to Susan Phillips, and Nonprofit and Voluntary Sector Quarterly’s (NVSQ) anonymous reviewers for skillfully guiding us to our best work, and to Julie Ann Israel for insightful comments and feedback.
Authors’ Note
This research is a collaborative effort resulting from the equal contribution of both authors.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
