What Drives Good Governance? A Structural Equation Model of Nonprofit BoardPerformance

Abstract

The study of nonprofit governance is coming into its theoretical heyday by incorporating a sophisticated understanding of its contingent and multidimensional nature. A systems view of governance acknowledges the interplay of internal and external dynamics on board performance. But empirically, large-scale, generalizable data that can test these concepts on board performance have been scarce. This study helps to fill that gap with a structural equation analysis of a national representative survey of member-serving organizations. The results suggest that board performance is associated with complex organizational and labor dynamics, and that performance metrics themselves are multidimensional. Furthermore, not all relationships with strong boards are directly measurable. Some appear related to indirect external market dynamics or healthy internal dynamics such as learning and self-evaluation.

Keywords

governance performance associations structural equation model

Introduction

All U.S. nonprofits of any tax status require a governing board, and all depend on the ability of their boards to succeed at governance. “Good governance” practices describe board activities and expectations that are both externally imposed and self-imposed to ensure a board’s duties of care, loyalty, and obedience are met. These practices, in turn, shape the way board members are selected, trained, and deployed.

The external and sometimes coercive imposition of expectations by regulators, donors, members, and other stakeholders on boards results in normative pressures to behave similarly across the nonprofit sector (Heugens & Lander, 2009; Miller-Millesen, 2003). However, while U.S. public law has some specific mandates for boards, such as a certain level of transparency, objectivity, and due diligence, the details on implementing most legal expectations are left to boards to decide. The board’s internal discretionary power over adoption of governance practices results in substantial variance in adopted practices (BoardSource, 2012; Brown & Guo, 2010; Ostrower, 2007).

One outcome of this ambiguity is myriad books for practitioners (mostly focused on charities) often with heavily prescriptive overtones. But the challenges caused by ambiguity about governance also complicate the efforts of those researchers seeking to understand what really drives board performance, including in noncharitable contexts. The potential for a range of external and internal influences on board structure, activities, and outcomes, and the ambiguous nature of the expectations imposed on boards of directors are well-known and widely discussed in nonprofit scholarship (see, for example, Brown, 2005; Cornforth, 2012; Doherty & Hoye, 2011; Herman & Renz, 1997; Stone & Ostrower, 2007). To offer just one example, boards themselves operate on multiple dimensions: as a group of individuals, as a team, and as a legal body. Theories of human, group, and organizational behavior may then be relevant to understanding them.

Literature Review: What Influences Nonprofit Board Performance?

What observers tend to believe drives nonprofit board performance is heavily influenced by the anecdotal and subjective normativism of the field, and its narrow lens on charitable organizations. One result is a strong mimetic pressure on other boards to follow prescribed practices. This situation, in concert with an understanding that each organization’s context shapes board behavior, results in some confusion in the field, driving Salipante’s (2014) call for more finely grained analysis and much greater attention to the details of each board’s situation.

Miller-Millesen (2003) offered one of the first frameworks, suggesting that a series of environmental factors (funding, regulations), organizational factors (age, stability, and professionalization), and board composition (demographics, recruitment practices) all influence board behavior. These factors do so by influencing the board’s stewardship and boundary-spanning activities, such as the extent to which board activity is needed to reduce environmental uncertainties or to raise money. In other words, not only must a range of theories be used to understand board behavior (e.g., agency, resource dependence), but data collection must also be able to capture the contingent and dynamic aspects of the governance function.

Miller-Millesen (2003) also observed a clear pattern of advice in the prescriptive literature, suggesting that, anecdotally at least, experts agree on a common set of board desirable governance practices, albeit focused on charities. This list continues to evolve: BoardSource updated its “Ten Responsibilities” document in 2015, calling for a greater focus on advocacy. Adding other current literature, both practitioner and scholarly, boards are expected to screen members for qualifications (Lakey, Hughes, & Flynn, 2004; Van Puyvelde, Caers, Du Bois, & Jegers, 2012), recruit for diverse skills and perspectives (Brown, 2002; Erhardt, Werbel, & Shrader, 2003; Gazley, Chang, & Bingham, 2010), assure healthy turnover by operating under term limits (Bradshaw, Murray, & Wolpin, 1992), educate themselves (BoardSource, 2012; Brown, 2007; Brown & Guo, 2010; Renz, 2013), self-assess (Lichtsteiner & Lutz, 2012), operate transparently (Ostrower, 2007), avoid conflicts of interest (Association Forum of Chicagoland, 2012), fulfill their duty of care by monitoring finances and supervising staff (Panel on the Nonprofit Sector, 2007), align their activities according to the organization’s strategic priorities (Brown, 2005; Herman & Renz, 2000; Ingley & Van Der Walt, 2005), be accountable for their actions (Green & Griesinger, 1996), achieve strong interpersonal relations with stakeholders (Chait, Ryan, & Taylor, 2011), and be able to resolve or avoid conflicts with staff or among board members (Bradshaw et al., 1992; Engle, 2013).

While Millesen’s model conceptually reinforces the multi-theoretical basis for board performance, tests of its predictive strength have been infrequent. Agreement on the multidimensionality of governance is far ahead of success in measuring it. Empirical tests to understand which aspects of prescribed board behavior actually make a difference in board outcomes are also still limited in their generalizability due to recurrent sampling bias. This is not a challenge faced only by governance scholars; Sowa, Selden, and Sandfort (2004) acknowledged the broader challenge of testing contingent outcome frameworks in the more general study of nonprofit organizational performance.

But some empirical efforts do pay off. Brown (2005) not only found empirical support for the multidimensionality of board outcomes but also observed the limitations on an overly “structural” perspective on boards (e.g., an emphasis on board size and other visible characteristics). Brown also noted the relationship between interpersonal and strategic board dynamics and outcomes. Similar results were found by Iecovich (2005) although with less success at determining the logic of interrelationships between board characteristics and outcomes. Engle (2013) used a finely grained mixed methods study to conclude that the quality of board members’ strategic decision-making processes and the quality of internal culture were related to board members’ assessment that they reached better decisions. And Boesso, Cerbioni, Menini, and Parbonetti (2015) found connections between healthy board dynamics and philanthropic performance in Italian foundations.

A characteristic of these various efforts is their attempt, although not always explicit, to differentiate between board structure and board functions (Iecovich, 2005). This distinction is helpful in separating not only the more passive but also potentially more visible characteristics of boards, such as meeting frequency and board size, from more dynamic but potentially more empirically challenging board characteristics such as task differentiation and role identity. For example, some structural variables such as board size are easy to capture from the annual tax return. However, scholars now are beginning to find evidence that functional characteristics are more important than structural characteristics in explaining board quality (Brown, 2005; Cornforth & Simpson, 2002). Such findings support a long-held belief by expert practitioners that successful boards rely on successful interpersonal dynamics (Axelrod, 2015; Chait et al., 2011).

A Systems Perspective on Board Performance

As noted, institutional theory scholars also observe the contingent nature of organizational behavior as well as the dearth of empirical observation of nonprofit performance (Heugens & Lander, 2009; Lecy, Schmitz, & Swedlund, 2012; Renz, 2013). Board behavior reflects variations in organizational culture, life cycle, human dynamics, and market dynamics, in addition to the previously noted voluntary nature of a board’s governance choices. Board outcomes also depend on staffing and other measures of organizational capacity, a factor already well understood to influence overall organizational performance (Eisinger, 2002).

In such cases, a contingent or systems view of boards can be useful, defined as an approach to evaluating board performance that takes into account all of the possible influences on organizational development (Bradshaw & Toubiana, 2014; Cornforth, 2012; Miller-Millesen, 2003; Ostrower & Stone, 2010). By controlling for multiple internal and external circumstances known to be associated with organizational performance, a systems model can offer a more fully specified model with greater precision in hypothesis testing. Lecy et al. (2012) observed that a central challenge to any multidimensional model is whether an accessible, user friendly, and understandable system of evaluation is available (Ebrahim, 2005); whether the latent behaviors are measurable; and whether aggregation assumptions can be made (e.g., to simplify the proposed behavioral model; see also Kaplan & Elliott, 1997).

Study Design

Structural equation models (SEMs) offer one type of contingency framework for testing multidimensional models of organizational performance. A SEM brings cross-sectional analysis closer to the point of causal inference, and is also well-suited to identifying the reciprocal and latent relationships that characterize nonprofit board performance. Reciprocal relationships can be found, for example, in the connection between board self-assessment and performance. A high performing board is expected to self-assess, but the self-assessment in turn influences board actions respecting performance measurement. With respect to latent factors, evidence relating board size to performance is weak and inconclusive, as noted above (Doherty & Hoye, 2011; Hu, Tam, & Guo-Sze Tan, 2010; Iecovich, 2005). But Brown (2005) has suggested that an indirect association between board size and strategic performance may exist, as board size may offer opportunities for a board to have a greater strategic orientation.

Researchers have also observed that board performance is related to CEO job satisfaction, which may in turn influence how CEOs rate their boards (Nobbie & Brudney, 2003). Structural equation modeling, therefore, is well-suited to examining simultaneous dependent relationships and testing simultaneously the multiple theories that might explain them (Hair et al., 2010). SEMs can also test the influence of external characteristics on internal behavior, a key ingredient in socio-ecological models (see, for example, Ostrom, 2009). For example, Mausolff and Spence (2008) observed associations between organizational effort at performance measurement and stakeholder assessments of program effectiveness.

Data and Method

Although boards comprise a group of individuals, governance itself is based on group actions. To understand how boards perform as a unit, a common practice in board research is to ask the CEO or Executive Director to assess the board as a whole (see, for example, BoardSource, 2012; Ostrower & Stone, 2010). This approach has its limitations as it relies on one perspective and cannot discern variations in how individuals on the same board perform. It is, however, theoretically sound in cases where a CEO is expected to be more objective about performance than board members themselves, and when CEO characteristics will be included in the analysis.¹ In such cases, possible sources of CEO bias should be included as control variables.

This analysis uses a dataset of 1,585 U.S.-based member-serving nonprofit organizations whose CEO participated in an extensive 2013 survey of governance practices (Gazley & Bowers, 2013). Associations are the common name for private sector organizations serving some collective interest (e.g., recreational clubs, fraternities and sororities, professional and learned societies, producer cooperatives, credit unions). Most are tax exempt under different sections of the U.S. Tax Code depending on the primary nature of their mission. They can serve individuals (many are professional and occupational societies), businesses or other nonprofit organizations (often as trade associations), or both. Despite this diversity, most associations hold in common three characteristics: a dues-paying membership base, nonprofit status, and a mission focused on some form of collective action serving the interests of its members.

The data came from a national survey of 13,391 tax-exempt member-serving organizations matched with 990 Form financial data, the annual tax form for charitable organizations. This sample was drawn from all 21,326 U.S. member-serving associations using a random, stratified design. A total of 1,585 responses were analyzed, for a response rate of 12% (margin of error < 2.4% at 95% confidence; a small amount of missing data may reduce analyzed sample sizes further). The criteria for eligibility in the larger dataset of U.S. 990 tax form filers are available from the American Society of Association Executives Foundation.² In the survey, the CEO or executive director of each nonprofit organization was asked to describe both the organization’s external environment and a broad range of governance characteristics, including board structure, selection procedures and challenges, deliberative processes, and meeting characteristics. They also assessed their boards on board relations with staff or with members, the board’s performance of fiduciary duties, its strategic orientation, and board development and self-assessment practices.

In the systems model hypothesized in Figure 1, organizational characteristics such as board dynamics, organizational capacity, and labor dynamics (center of figure) are hypothesized to be associated with key legal and industry dynamics (left side of figure). These all, in turn, are hypothesized to be associated with board performance (right side of figure). The relationships are expected to be complex and multilateral. For example, nonprofit tax status could be associated with the extent to which boards self-assess, with charities (501(c)(3)s in the United States) possibly more likely to self-assess due to the greater transparency expected of the charitable part of the nonprofit sector. In another example, trade associations might impose different representational expectations than, say, learned societies, resulting in different board selection criteria. Organizational characteristics and behaviors such as tax status might shape a board’s choice of structure and operating norms (Coombes, Morris, Allen, & Webb, 2011). These choices, in turn, lead to intermediate outcomes such as board development and self-assessment, reflections of the amount of internal emphasis on good governance practices. These good governance practices, in turn, could shape the board’s performance outcomes. And all CEOs might rate the board of a thriving organization more highly than a struggling one regardless of these other internal considerations.

Figure 1.

Hypothesized model of board performance indicators.

As we note above, the structural dynamics of board performance are likely to be quite complex, but as a first cut at these processes, this article implements a relatively simple form of structural equation modeling commonly known as path analysis. Path analysis can reveal direct and indirect effects of exogenous factors on outcomes like board performance, through mediators such as board dynamics, capacity, and so on. Path analysis imposes the assumptions that the errors across these equations are uncorrelated and that causality is unidirectional, meaning that it cannot reveal any reciprocal relationships, say, between performance and board dynamics. Despite this circumstance, it is a good place analytically to begin investigating the complex processes that we are interested in and provides a solid foundation for future research.

This path analysis uses a two-stage process to first examine, in separate models, the degree to which Legal Structure and Industry Dynamics (e.g., tax status, structure, and membership characteristics) relate to the latent constructs Board Dynamics, Organizational Capacity, and Labor Dynamics. A second stage then investigates the direct association between the first set of (industry and legal) characteristics and board performance when the impact of the latent factors is accounted for. The remainder of this section describes the measurement of these variables. All of these models are estimated as ordinary least squares regressions that include sample weights and calculate robust standard errors.

Dependent Variable

Board performance is operationalized as a multi-faceted concept, measured using multiple indicators. The survey asked CEOs to rate their boards on 19 separate performance measures, based on a comprehensive scan of the governance literature with refinements offered by survey pretesters. Each measure could be rated on “needs improvement (score of 1),” “satisfactory (2),” or “excellent (3).” These measures were analyzed in two ways: as an aggregate single performance measure, and in thematic subgroupings of performance characteristics, created via a principal components analysis (displayed in Table 1, along with descriptive statistics). Table 1 suggests that boards are rated by their staff leadership in distinct but related areas of performance. While the results show that ratings vary depending on the performance measure, they also reveal four clusters: (a) a set of performance indicators related to the board’s Culture Orientation (e.g., its attention to interpersonal relationships and to fostering a culture of responsibility), as well as (b) a group of indicators related to Strategic Orientation, (c) to Performance Orientation, and (d) to Membership Orientation. Naturally, these groupings will vary study by study as they depend on researcher choice in the variable design. But these findings reinforce the point that board performance does not occur on a single dimension as boards serve a variety of fiduciary, political, and strategic roles (Axelrod, 2015; Chait et al., 2011).

Table 1.

Descriptive Statistics and Principal Components Analysis for Dependent Variable “Board Performance” (Rotated, Varimax With Kaiser Normalization, 59% Variance Explained, selected factors boldfaced for emphasis).

Dependent variable (scale = 1-3)	M (SD)	Culture-oriented	Strategy-oriented	Performance-oriented	Member-oriented
Overall quality of board relations with staff	2.49 (0.659)	.671	.256	.046	.216
Overall quality of relations among board members	2.46 (0.664)	.761	.143	.063	.201
Overall quality of board relations with the membership	2.08 (0.800)	.482	.094	.036	.621
Accountability to members	1.89 (0.860)	.228	.182	.160	.740
Direct outreach and engagement of membership	1.60 (0.801)	.131	.227	.143	.737
Stewardship over the organization’s resources	2.48 (0.713)	.583	.294	.064	.172
Willingness to take responsibility for difficult decisions	2.34 (0.773)	.645	.335	.083	.135
Ability to make decisions based on organizational interests	2.37 (0.758)	.702	.300	.080	.075
Collegiality of the board atmosphere	2.49 (0.741)	.775	.134	.069	.012
Board’s ability to set performance standards for itself	1.11 (0.777)	.114	.123	.863	−.016
Board’s record of enforcing self-imposed performance standards	1.00 (0.896)	.127	.103	.861	.067
Securing feedback on its own performance from key constituencies	0.97 (0.811)	−.019	.078	.710	.300
Effectiveness at strategic thinking	1.78 (0.809)	.257	.742	.158	.036
Board participation in advocacy/public policy	1.53 (1.014)	−.021	.442	.031	.345
Effectiveness at aligning resources with strategic needs	1.97 (0.787)	.262	.705	.075	.174
Ability to serve as a catalyst for change	1.85 (0.797)	.301	.669	.107	.147
Understanding of organization’s external environment and trends	2.02 (0.789)	.280	.646	.067	.188
Leading in a way that maintains the public trust in nonprofits	2.34 (0.797)	.398	.473	−.025	.242
Ability to achieve strategic plan	1.87 (0.918)	.204	.710	.089	.045
Sum of all performance factors (range 2-57 on a possible scale of 0-57, normally distributed)	36.65 (8.842)

Independent Variables

Board Characteristics and Operation

Descriptive statistics of all independent variables are displayed in Table 2. As noted, board performance is in part a function of dynamics among board members, the capacity of the board to make good decisions, and the industry and labor dynamics with which it must contend. These are latent constructs in that they cannot be measured directly but must instead be proxied with observable variables that represent the underlying construct. A factor analysis creates these three internal characteristics of boards. As in this case, when indicators included in the creation of a factor variable are dichotomous, a polychoric principal components analysis is used (Kolenikov & Angeles, 2004). When measures include dichotomous and continuous variables, the modeling uses a standard regression-based principal components analysis.

Table 2.

Descriptive Statistics for Independent Variables.

Variable	n	M	SD	Minimum	Maximum
Board dynamics	1,574	−0.01039	0.898246	−3.24315	1.709302
Organizational capacity	1,560	−5.94E-10	1	−3.31728	5.260643
Labor dynamics	1,584	−1.37E-08	1	−2.21114	0.781555
Trade association	1,585	0.329338	0.470121	0	1
501(c)(3)	1,585	0.490852	0.500074	0	1
501(c)(6)	1,585	0.425868	0.49463	0	1
National organization	1,585	0.319874	0.466575	0	1
International organization	1,585	0.090221	0.286588	0	1
Chapter organization	1,585	0.378549	0.485179	0	1
Public sector organization	1,585	0.028391	0.16614	0	1
Education organization	1,585	0.123028	0.328574	0	1
Membership growth	1,585	0.388644	0.487596	0	1
Competition for members	1,585	0.786751	0.409732	0	1

Board Dynamics

The latent construct for “Board Dynamics” was built on 11 survey questions describing board actions that can influence performance (producing an eigenvalue of 2.20). The survey captured the following structural characteristics: board size, whether the CEO serves as the president, and whether the board has term limits (weakly but positively associated with performance in Gazley & Bowers, 2013). Selection criteria include whether the board entertains external nominations or direct external appointments (found to be associated with weaker performance in Gazley & Bowers, 2013). Positively related selection criteria include whether board nominees are screened before election, and whether bylaws permit competitive elections.

Variables also reflect the strength of board dynamics once elected, including whether board members assess their own performance, the amount of attention paid to board training and development, the board’s willingness to report its performance to members, and the amount of board time spent on strategic-level decision making. Experts consider strategic leadership and team learning to be among the key dimensions of organizational learning. They also suggest that organizational learning is a complex and dynamic process and the links to outcomes are multidimensional, thus best captured through SEM and similar modeling (Yang, Watkins, & Marsick, 2004). Research has also attempted, although with limited success due to small samples, to identify relationships between task environment and resulting board structure and activities (Iecovich, 2005).

Organizational Capacity

A latent measure of “Organizational Capacity” was developed, with heavy emphasis on the staff’s ability to support healthy board dynamics. Human capital capacity measures include number of full time equivalent staff (logged to reduce outlier effects), whether staffing is mainly paid (vs. volunteer), and whether staff have association management training. This variable also includes whether the organization is operating under a strategic plan that would guide staff. These last three factors were positively related to performance in Gazley and Bowers (2013). Finally, respondents were asked to estimate the total time (in hours per week) staff spend supporting the board. These variables produced an eigenvalue of 1.43.

Labor Dynamics

The latent variable of “Labor Dynamics” is built on two survey questions, the stability of the staff (i.e., low turnover) and whether the CEO is contemplating leaving the organization. Numerous board studies find a strong relationship between staff turnover and staff satisfaction with the board (BoardSource, 2012; Ostrower, 2007). This construct returned an eigenvalue of 1.42.

Legal and Industry Dynamics

Working next on relationships between variables in the center and the left side of the hypothesized model, an additional research question examines ways in which “Legal” and “Industry” dynamics influence board performance through their influence on the latent board characteristics described above. These include eight variables reflecting tax classification, structure, and subsector characteristics. The first point of comparison is whether an organization has 501(c)(3) status, has 501(c)(6) status, and/or is classified as a trade association. Second, under “structure” is whether an organization focuses activities exclusively in the United States or if some activities and resources are located in other countries. International associations were found to have lower performing boards in Gazley and Bowers (2013), perhaps due to greater leadership and management challenges associated with geographical diversity.

Also included is whether the organization has a single or centralized membership structure versus local chapters, a dynamic that might influence governance choices. For subsector characteristics, associations representing primarily public employees and those serving educational professionals are compared with all other associations, based on Gazley’s (2014) finding that these organizations appear to have stronger board performance, perhaps due to a stronger public service ethos. Finally, organizations that report high recent membership growth, and those that report significant competition for members, are compared, in both cases because these situations might demand greater effort from their boards.

Findings

The results of the analyses are presented in Tables 3 and 4 and in Figure 2. The discussion of the findings begins with the signs and significance levels for key predictors, then turns to the path diagram to discuss substantive effects. The first table presents the findings from models of the influence of legal and industry factors on Board Dynamics (column 1), Organizational Capacity (column 2), and Labor Dynamics (column 3).

Table 3.

Relationships of Sector and Legal Factors to Board Dynamics, Organizational Capacity, and Labor Dynamics.

Variables	Board dynamics	Organizational capacity	Labor dynamics
Trade Association	−0.0916	0.00542	0.0798
	(−1.61)	(0.10)	(1.28)
501(c)(3)	0.00558	−0.142	−0.186*
	(0.06)	(−1.17)	(−1.72)
501(c)(6)	0.155	0.246*	−0.0602
	(1.56)	(2.07)	(−0.57)
National	0.302***	0.387***	0.0783
	(6.01)	(7.29)	(1.39)
International	0.241**	0.596***	0.0530
	(2.96)	(6.59)	(0.57)
Chapter organization	0.208***	0.298***	0.0247
	(4.57)	(5.86)	(0.46)
Public sector association	0.0971	0.241*	0.0376
	(0.79)	(2.08)	(0.24)
Education association	0.123*	−0.123	0.0535
	(1.73)	(−1.58)	(0.65)
Membership growth	0.181***	0.142**	0.0680
	(4.00)	(2.86)	(1.29)
Competition for members	−0.0268	−0.0674	0.0324
	(−0.48)	(−1.14)	(0.53)
Sample weights	−0.0845	0.0151	0.0360
	(−1.88)	(0.25)	(0.71)
Intercept	−0.229	−0.336**	−0.0441
	(−1.75)	(−2.05)	(−0.30)
n	1,574	1,560	1,584
R ²	.08	.12	.01

Note. Values in parentheses represent t statistics.

p < .1. **p < .05. ***p < .01.

Table 4.

Relationships of External Factors and Board Dynamics, Organizational Capacity, and Labor Dynamics With Performance.

Variables	Coefficient (T-score)
Board dynamics	1.888***
	(8.51)
Capacity	0.915***
	(4.56)
Labor dynamics	0.949***
	(5.23)
Trade association	1.894***
	(4.19)
501(c)(3)	0.210
	(0.25)
501(c)(6)	0.0292
	(0.04)
National	−0.822*
	(−1.95)
International	−1.797**
	(−2.70)
Chapter organization	0.207
	(0.53)
Public sector association	2.201**
	(2.03)
Education association	0.586
	(1.00)
Membership growth	2.660***
	(6.94)
Competition for members	0.756*
	(1.70)
Sample weights	−0.423
	(−1.09)
Intercept	27.95***
	(25.07)
N	1,479
R ²	.14

p < .1. **p < .05. ***p < .01.

Figure 2.

Path analysis of relationship between external factor and board dynamics, organizational capacity, and labor dynamics, and aggregate board performance.

As noted earlier, some factors are associated with more than one of these latent constructs. For example, an organization that operates on a national or international scale (compared with a regional or local organization) is both positively and significantly associated with “Board Dynamics” and “Organizational Capacity.” Similarly, a chapter structure is associated positively with both constructs. Together, these results suggest that larger more complex organizations may have boards that receive more training, greater levels of transparency, and better leadership along with the other factors hypothesized to contribute to positive board dynamics. Similarly, they may have greater internal organizational capacity, including a larger and more professionalized staff and a stronger strategic orientation, among others. High membership growth was also positively associated with both “Board Dynamics” and “Organizational Capacity,” and there is likely to be a reciprocal relationship between these variables in the sense that these two dynamics support one another. Having a 501(c)(6) tax status was also positively associated with “Organizational Capacity” but did not reach the traditional threshold of statistical significance in the “Board Dynamics” model.

Just one of the left-hand variables predicted “Labor Dynamics.” Although all boards of directors must deal with labor volatility, structure and subsector status do not in themselves have a direct relationship to staff and CEO turnover. However, tax status was a significant predictor of this latent construct in that 501(c)(3) public charities appear to experience greater staff volatility, possibly arising from higher turnover in their leadership.

Table 4 also displays associations between board performance and legal and industry dynamics, in addition to latent constructs. The results suggest that being a trade association is positively and significantly associated with board performance, as is serving mainly public sector members. The results suggest that both of these membership subsectors expect higher performance from their boards.

The findings also indicate that both membership growth and strong competition for members are positively associated with board performance. The first of these characteristics may, admittedly, be somewhat endogenous to the dependent variable in the sense that CEOs may reward boards for membership growth in their ratings. The latter finding suggests that boards may improve their performance to make their organizations more attractive to members in competitive environments.

The findings presented in Table 4 also suggest that Board Dynamics, Organizational Capacity, and Labor Dynamics are all highly significant and positively associated with Board Performance. This result, coupled with those presented in Table 3, provides support for expectations that (a) board dynamics, capacity, and labor issues relate to performance and that (b) some legal and industry factors also relate to board performance both directly and indirectly through their influence on the operation and capacity of those boards. Therefore, much of the earlier literature that takes a multiplex view of board performance is supported. The path diagram presented in Figure 2 makes it easier to visualize and precisely estimate both the direct and indirect associations. The paths are labeled with standardized coefficients, signifying the SD change in the dependent variable given a 1 SD change in the independent variable, or a 0 to 1 change in the case of dichotomous variables. Direct effects are represented by the beta coefficient on a single path; indirect effects are the product of all the paths linking two variables; and total impacts are calculated as the sum of direct and indirect effects. Only paths found to have statistically significant effects are displayed.

This section explains how to interpret the direct and indirect effects displayed in Figure 2. One sees, for example, that being an organization that is National in scope has a negative direct relationship to Board Performance (.04*, p < .10). However, the variable can still have an indirect relationship with Board Performance that is positive. This effect is seen through the positive relationship between “National” and both “Board Dynamics” (.15***, p < .01) and “Organizational Capacity” (.11***, p < .01). The analysis suggests that being a National organization indirectly predicts an increase in Board Performance by .03 SDs (.15 × .21) through its positive impact on Board Dynamics, plus an increase of .01 SDs (.11 × .12) through Organizational Capacity. In other words, despite its negative direct effect, being a National organization has a small positive relationship to Board Performance overall (.03 + .01 – .03 = .01), because it is associated with better Board Dynamics and increased Organizational Capacity.

In the case of internationally focused organizations, their negative direct relationship to Board Performance (.05 SD) overwhelms positive indirect effects through Board Dynamics and Organizational Capacity. Specifically, the analysis suggests that being an international organization predicts higher Board Performance indirectly by .016 SDs due to a positive association with Board Dynamics (.14 × .12) and by .014 SDs due to the relationship with Organizational Capacity (.07 × .21). However, that still leaves a negative total relationship of .01 SDs (.016 + .014 – .05 = –.01).

Organizational structure, as measured by having Chapters versus a more centralized structure, appears to have only an indirect association with Board Performance. Having chapters is related to better Board Dynamics, which is in turn positively associated with Board Performance, creating an indirect positive relationship between Chapters and Board Performance of .02 SDs. Similarly, the Chapter variable has an indirect positive relationship of .013 SDs through its association with higher Organizational Capacity.

As the lack of a path straight from Chapters to Board Performance indicates, there is no direct relationship between these variables once the mediating influence of Board Dynamics and Organizational Capacity are controlled for. Thus, the total impact for having a Chapter structure is .033 SDs. Membership Growth, reflecting not only a healthy organization with more resources but also possibly higher member expectations of boards, has both a direct and an indirect positive association with Board Performance. The variable first has a positive and significant association with Board Dynamics, which in turn has a positive association with Board Performance. Taken together, these suggest that a 1 SD change in Membership Growth indirectly predicts higher Board Performance by .01 SDs through Board Dynamics. Even after controlling for that impact, however, Membership Growth still has a positive and significant direct relationship to performance of .18 SDs. That result creates a rather large total predictive effect for this variable, of .19 SDs. Quite interestingly, this Membership Growth variable does not have a direct association with Organizational Capacity or Labor Dynamics. Future research should test why that absence of an effect might occur, but it may simply be that this model’s inclusion of other variables associated with Membership Growth has already captured the statistical impact.

Interestingly, variables measuring the subsector or industry in which an organization operates tended to have fewer paths of connection to Board Performance than those measuring structure, scope, and resources. Being an education organization had only a small indirect effect of .02 SD through Board Dynamics. Alternatively, being a public-serving association was not associated with Board Dynamics, Organizational Capacity, or Labor Dynamics and, thus, had only a small direct relationship to Board Performance (.05 SDs).

Turning to variables measuring tax status of organizations, the path running directly between Trade Association and Board Performance indicates a relatively large direct relationship (.12 SDs). However, being a Trade Association was negatively associated with Organizational Capacity, which means it had a small, but negative (–.005) indirect relationship to Board Performance through that variable. Nonetheless, taking both direct and indirect influence together, it suggests a relatively large positive total effect of .115 SDs for being a Trade Association.

Having 501(c)(3) charity status has only a small indirect positive relationship (.01 SDs) with Board Performance through Labor Dynamics, though it is interesting to note that this is the only variable in the model which had a direct association with Labor Dynamics. The results suggest that charities are more sensitive to labor dynamics than are noncharitable tax-exempt organizations, although reasons must still be determined (controlling for budget size, which, we do not, might help). Finally, statistically significant effects for having 501(c)(6) status (vs. all other tax statuses) and Competition for Members was not found and those variables are not displayed in this model.

Moving from right to left, Figure 2 also displays the direct relationship between Board Dynamics (e.g., board functions and characteristics) and Board Performance. Compared with Table 2, the relative substantive connection of these variables to Board Performance is easier to assess in the path diagram because it presents standardized coefficients. Those coefficients suggest that Organizational Capacity and positive Labor Dynamics have positive relationships with Board Performance that are relatively large but significantly smaller than the total impact observed for Board Dynamics (as well as for Membership Growth). A 1-SD improvement in Board Dynamics is associated with a .21-SD increase in Board Performance, giving this measure the largest impact in the analysis.

Exploring the Components of Performance

As noted above, our aggregate measure of board performance comprises four distinct factors: the degree to which the board cultivates an effective culture, maintains a strategic orientation, focuses on member relations, and/or effectively sets performance goals and responds to performance information. Because these components reflect different foci and competencies, it is likely that they are related to different external factors such as tax status and sector, and internal characteristics such as board dynamics, organizational capacity, and labor dynamics. Four additional path diagrams (Figures 3 to 6) explore the impact of these factors on the various components of board performance.

Figure 3.

Relationship between external factor and board dynamics, organizational capacity, and labor dynamics, and culture orientation.

Figure 4.

Relationship between external factor and board dynamics, organizational capacity, and labor dynamics, and strategic orientation.

Figure 5.

Relationship between external factor and board dynamics, organizational capacity, and labor dynamics, and performance orientation.

Figure 6.

Relationship between external factor and board dynamics, organizational capacity, and labor dynamics, and membership orientation.

To facilitate the analysis, only direct effects are displayed (indirect effects will be identical to Figure 2). The main overall finding is that while all direct relationships remain positive, as found in Figure 2, the strength of the effect varies according to changes in the performance dimension. For example, with respect to an effective board culture (Figure 3), we also see fewer exogenous factors predicting this component of performance. Only two such factors, being a trade association and reporting significant membership growth, continue to have a direct positive relationship to board culture. However, these variables continue to have an indirect relationship to a healthy board culture through their relationship to board dynamics and organizational capacity, although we also note that the effect of board dynamics weakens. In other words, capturing board dynamics helps to capture an elusive latent construct of “board culture,” a useful finding for governance scholars.

When examining how a CEO rates a board on its strategic performance, Figure 4 presents a path diagram with a set of relationships quite similar to the paths presented in Figure 2. This is not surprising given that strategic orientation is the largest single contributor to the total measure of performance. The major change is that two additional exogenous factors, including being an internationally focused organization and reporting significant competition for members, have a direct relationship to the strategic component of performance. This result suggests that market dynamics play a stronger direct effect on a board’s strategic orientation. The substantive effect of these variables is quite small, however.

In Figure 5, the strength of a board’s performance measurement orientation depends much more heavily on indirect effects than on direct effects. Being a nationally focused or public-sector-serving organization cease to have a direct influence on the disaggregated component of total performance, although the former continues to have an indirect effect via its influence on board dynamics. The effects of the other endogenous factors, including organizational capacity and labor dynamics, shrink by about 50% relative to the aggregate model of performance in Figure 2. Finally, Figure 6 presents the path diagram for the component of total performance that captures the Board’s attention to member relations. Here, again the pattern of relationships is essentially unchanged from those in Figure 2.

The biggest takeaway from these additional analyses is that the direct and indirect effects of both exogenous and endogenous predictors of Board performance are surprisingly consistent across the different subcomponents of that construct. We also identify four unique components of performance, including the creation of a productive culture, a strategic orientation, effective collection and use of performance information, and attention to member relations. Also of note is that effects are generally positive across the board, with few exceptions. If there are organizational circumstances that cause a negative drain on board performance, they are not captured in these models. The consistency of predictions across the different components of performance suggests that board dynamics, organizational capacity, and labor dynamics are important for all rather than only certain elements of performance.

Discussion and Conclusion

This study has taken a careful approach to sampling and weighting survey data to produce fairly generalizable conclusions about board leadership in a variety of member-serving nonprofit associations. The usual limitations researchers encounter when using only filing nonprofits still apply, for example, that smaller and non-filing organizations are left out. The chief analytical limitation is the cross-sectional nature of the data, which precludes causal arguments. However, the model does capture temporally exogenous organizational characteristics such as tax status and membership characteristics that certainly precede board dynamics in their causal order. Future researchers might consider including these relationships as causal hypotheses in panel studies. On the contrary, some relationships found here may be reciprocal, in the sense that boards are being rated as successful because they have achieved certain healthy dynamics but are also pursuing healthy dynamics so that they will be successful (to put it another way, a board not only conducts a self-assessment to improve itself but also might improve itself due to the self-assessment). So a second limitation must be acknowledged, the potential for common method bias.

Another important limitation is the variation among respondents regarding their expectations of their boards. We have provided what the literature suggests is a good effort at an objective list of board expectations. However, while all nonprofit CEOs/executive directors are in a position to evaluate their boards, they will not necessarily agree on what represents success. This study does not compare expected performance with realized performance, but this effort has rarely been attempted in any governance research. In addition, we cannot determine cases where boards are being rated as high performing by their CEOs simply because the organization is healthy, although this systems model goes further than much of the prior research by attempting to control for such response bias by including the chief influences on such a response in the model (including CEO intentions to quit and an objective variable, membership growth).

Three conclusions can be made from this analysis. First is the empirical reinforcement of a widely held conceptual and practical view of board performance as a complex phenomenon. While this fact is practically self-evident in this era of increasingly sophisticated governance research, empirical evidence on specifically how and where board performance is influenced has been harder to produce, as noted in this article’s literature review. These results suggest that systems models of board performance are worth the effort, particularly when compared with simpler analytic approaches where indirect effects cannot be identified.

Second is the finding that, while some of the more visible structural and legal characteristics of boards do matter, there are also latent influences on board performance that are more challenging to measure and observe, but worth the effort. We note, for example, the lack of a direct connection between charitable tax status and board performance, yet the mediating effect of tax status on both an organization’s capacity to support a healthy board and on healthy labor dynamics. This finding also reinforces the value in studying boards of directors across different parts of the tax-exempt legal spectrum. We suggest that the choice of tax status, not to mention other legal and structural choices, may change stakeholder expectations of their boards of directors.

In contrast to organizations serving the education sector, associations that primarily serve government employees, a subsector with distinct expectations of public service and transparency, realize their positive connection to board performance directly, rather than through the mediating effect of board dynamics (also see Gazley, 2014). Also, an indirect effect is found between chapter structure and board performance via the mediating effect of board dynamics. This finding will be of interest to association leaders because chapters are considered anecdotally to be more challenging to lead and manage (Gazley & Kissman, 2015). The lesson that can be learned here is that chapter-based associations may need to invest more heavily in board dynamics to achieve their representational and strategic goals related to supporting a healthy chapter structure—but the greater effort pays off in a stronger board overall.

Third is the finding that, after comparing direct and indirect effects on performance of a variety of legal, structural, and functional dynamics of membership associations, the strongest independent effect comes from the ways in which boards design their own internal dynamics. In a word, investing in board development pays off, as does related efforts to be careful with board member selection and self-assessment. Organizational capacity, particularly investing in professional staff development, does support strong boards. But the ways in which boards support themselves perhaps have the greatest impact on their ability to meet their fiduciary obligations.

Footnotes

Acknowledgements

The authors gratefully acknowledge support from the ASAE Foundation to produce the original data on which this and other publications are based, and the helpful comments of anonymous reviewers.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The ASAE Foundation funded the survey on which this analysis is based.

Notes

Author Biographies

Beth Gazley is professor, Indiana University School of Public and Environmental Affairs. Her research addresses nonprofit management and governance, volunteerism, and intersectoral relations.

Jill Nicholson-Crotty is associate professor, Indiana University School of Public and Environmental Affairs. Her research focuses on both public and nonprofit management and the role of those sectors in the policy process.

References

Association Forum of Chicago. (2012, March). Fiduciary and management duties for the association executive and governing body. Chicago, IL: Association Forum.

Axelrod

N. R.

(2015). The governing board: Key responsibilities for boards and board members. Washington, DC: ASAE Association Management Press.

BoardSource. (2012). Nonprofit governance index 2012: CEO survey of BoardSource members. Washington, DC: Author.

Boesso

Cerbioni

Menini

Parbonetti

(2015). Philanthropy by decree. Nonprofit Management and Leadership, 25, 197-213.

Bradshaw

Murray

Wolpin

(1992). Do nonprofit boards make a difference? An exploration of the relationships among board structure, process, and effectiveness. Nonprofit and Voluntary Sector Quarterly, 21, 227-249.

Bradshaw

Toubiana

(2014). The dynamics of nested governance: A systems perspective. In Cornforth

Brown

W.A.

(Eds.), Nonprofit governance: Innovative perspectives and approaches (pp. 231-248). New York, NY: Routledge.

Brown

W. A.

(2002). Inclusive governance practices in nonprofit organizations and implications for practice. Nonprofit Management and Leadership, 12, 369-385.

Brown

W. A.

(2005). Exploring the association between board and organizational performance. Nonprofit Management and Leadership, 15, 317-339.

Brown

W. A.

(2007). Board development practices and competent board members: Implications for performance. Nonprofit Management and Leadership, 17, 301-317.

10.

Brown

W. A.

Guo

(2010). Exploring the key roles for nonprofit boards. Nonprofit and Voluntary Sector Quarterly, 39, 536-546.

11.

Chait

R. P.

Ryan

W. P.

Taylor

B. E.

(2011). Governance as leadership: Reframing the work of nonprofit boards. Washington, DC: BoardSource.

12.

Coombes

S. M. T.

Morris

M. H.

Allen

J. A.

Webb

J. W.

(2011). Behavioural orientations of non-profit boards as a factor in entrepreneurial performance: Does governance matter? Journal of Management Studies, 48, 829-856.

13.

Cornforth

(2012). Nonprofit governance research: Limitations of the focus on boards and suggestions for new directions. Nonprofit and Voluntary Sector Quarterly, 41, 1116-1135.

14.

Cornforth

Simpson

(2002). Change and continuity in the governance of nonprofit organizations in the United Kingdom: The impact of organizational size. Nonprofit Management and Leadership, 12, 451-470.

15.

Doherty

Hoye

(2011). Role ambiguity and volunteer board member performance in nonprofit sport organizations. Nonprofit Management and Leadership, 22, 107-128.

16.

Ebrahim

(2005). Accountability myopia: Losing sight of organizational learning. Nonprofit and Voluntary Sector Quarterly, 34, 56-87.

17.

Eisinger

(2002). Organizational capacity and organizational effectiveness among street-level food assistance programs. Nonprofit and Voluntary Sector Quarterly, 31, 115-130.

18.

Engle

(2013). The strategic decision-making process of the board and its impact on decision outcomes (Unpublished doctoral thesis). Case Western Reserve University, Cleveland, OH.

19.

Erhardt

N. L.

Werbel

J. D.

Shrader

C. B.

(2003). Board of director diversity and firm financial performance. Corporate Governance, 11, 102-111.

20.

Gazley

(2014). Good governance practices in professional associations for public employees: Evidence of a public service ethos? Public Administration Review, 74, 736-747.

21.

Gazley

Bowers

(2013). What do high performing boards look like? Results from ASAE’s national study. Washington, DC: ASAE and the Center for Association Leadership.

22.

Gazley

Chang

W. K.

Bingham

L. B.

(2010). Board diversity, stakeholder representation and collaborative performance in community mediation centers. Public Administration Review, 70, 610-620.

23.

Gazley

Kissman

(2015). Transformational governance: How boards achieve extraordinary change. Hoboken, NJ: John Wiley.

24.

Green

J. C.

Griesinger

D. W.

(1996). Board performance and organizational effectiveness in nonprofit social services organizations. Nonprofit Management & Leadership, 6, 381-402.

25.

Hair

J. F.

Black

Babin

B. J.

Aderson

(2010). Multivariate data analysis (7th ed.). Upper Saddle River, NJ: Prentice-Hall.

26.

Herman

R. D.

Renz

D. O.

(1997). Multiple constituencies and the social construction of nonprofit organization effectiveness. Nonprofit and Voluntary Sector Quarterly, 26, 185-206.

27.

Herman

R. D.

Renz

D. O.

(2000). Board practices of especially effective and less effective local nonprofit organizations. American Review of Public Administration, 30, 146-160.

28.

Heugens

P. P. M. A. R.

Lander

M. W.

(2009). Structure! Agency! (And other quarrels): A meta-analysis of institutional theories of organization. The Academy of Management Journal, 52, 61-85.

29.

H. W.

Tam

O. K.

Guo-Sze Tan

(2010). Internal governance mechanisms and firm performance in China. Asia Pacific Journal of Management, 27, 727-749.

30.

Iecovich

(2005). Environmental and organizational features and their impact on structural and functional characteristics of boards in nonprofit organizations. Administration in Social Work, 29, 43-59.

31.

Ingley

Van Der Walt

(2005). Do board processes influence director and board performance? Statutory and performance implications. Corporate Governance, 13, 632-653.

32.

Kaplan

Elliott

P. R.

(1997). A didactic example of multilevel structural equation modeling applicable to the study of organizations. Structural Equation Modeling: A Multidisciplinary Journal, 4, 1-24.

33.

Kolenikov

Angeles

(2004). The use of discrete data in PCA: Theory, simulations, and applications to socioeconomic indice. Chapel Hill: Carolina Population Center, University of North Carolina.

34.

Lakey

B. M.

Hughes

S. R.

Flynn

(2004). Governance committee. Washington, DC: BoardSource.

35.

Lecy

J. D.

Schmitz

H. P.

Swedlund

(2012). Non-governmental and not-for-profit organizational effectiveness: A modern synthesis. Voluntas: International Journal of Voluntary and Nonprofit Organizations, 23, 434-457.

36.

Lichtsteiner

Lutz

(2012). Use of self-assessment by nonprofit organization boards: The Swiss case. Nonprofit Management and Leadership, 22, 483-506.

37.

Mausolff

Spence

(2008). Performance measurement and program effectiveness: A structural equation modeling approach. International Journal of Public Administration, 31, 595-615.

38.

Miller-Millesen

J. L.

(2003). Understanding the behavior of nonprofit boards of directors: A theory-based approach. Nonprofit and Voluntary Sector Quarterly, 32, 521-547.

39.

Nobbie

P. D.

Brudney

J. L.

(2003). Testing the implementation, board performance, and organizational effectiveness of the policy governance model in nonprofit boards of directors. Nonprofit and Voluntary Sector Quarterly, 32, 571-595.

40.

Ostrom

(2009). A general framework for analyzing the sustainability of social-ecological systems. Science, 325, 419-422.

41.

Ostrower

(2007). Nonprofit governance in the United States: Findings on performance and accountability from the first national representative study. Washington, DC: Urban Institute.

42.

Ostrower

Stone

M. M.

(2010). Moving governance research forward: A contingency-based framework and data application. Nonprofit and Voluntary Sector Quarterly, 39, 901-924.

43.

Panel on the Nonprofit Sector. (2007). Principles for good governance and ethical practice. Washington, DC: Independent Sector.

44.

Renz

D. O.

(2013, January 1). Reframing governance II. Nonprofit Quarterly. Retrieved from http://www.nonprofitquarterly.org/governancevoice/21572-reframing-governance-2.html

45.

Salipante

(2014). Lessons for governance research and practice: Examining behavioral processes and conceptual resources. In Cornforth

Brown

W.A.

(Eds.), Nonprofit governance: Innovative perspectives and approaches (pp. 272-288). New York, NY: Routledge.

46.

Sowa

J. E.

Selden

S. C.

Sandfort

J. R.

(2004). No longer unmeasurable? A multidimensional integrated model of nonprofit organizational effectiveness. Nonprofit and Voluntary Sector Quarterly, 33, 711-728.

47.

Stone

M. M.

Ostrower

(2007). Acting in the public interest? Another look at research on nonprofit governance. Nonprofit and Voluntary Sector Quarterly, 36, 416-438.

48.

Van Puyvelde

Caers

Du Bois

Jegers

. (2012). The governance of nonprofit organizations: Integrating agency theory with stakeholder and stewardship theories. Nonprofit and Voluntary Sector Quarterly, 41, 431-451.

49.

Yang

Watkins

K. E.

Marsick

V. J.

(2004). The construct of the learning organization: Dimensions, measurement, and validation. Human Resource Development Quarterly, 15, 31-55.