Board Capture in the Nonprofit Sector? An Exploration of a Possible Cause and Solution

Abstract

Responsibilities of nonprofit boards of directors include financial oversight and supervision of the CEO. Lack of expertise and the nature of supervision may contribute to board capture, characterized by longer executive tenure despite poor financial performance. We explore whether nonprofit boards are as responsive as for-profit boards to poor CEO financial performance. Our analyses indicate that the boards of the largest nonprofits do not definitively respond to changes in revenues, expenses, savings, or profit in CEO tenure decisions. This contrasts sharply with our for-profit results. However, changes in net assets as a fraction of total assets provoke nonprofit board responses comparable to those of for-profit boards. These findings are consistent with a pattern wherein nonprofit boards provide relatively lax financial oversight most of the time, but act when financial results threaten long-term solvency. Nonprofit boards might consider compensating board members to improve supervision of CEOs.

Keywords

nonprofit governance nonprofit board performance nonprofit board capture

Liu (2017) maintains that the most important responsibilities of nonprofit boards of directors include providing financial oversight and selecting, supporting, evaluating, and replacing the chief executive officer (CEO). Given the increasing scale, scope, and importance of work done by the nonprofit sector, CEOs should be expected to meet or exceed financial and other performance standards and accept that their tenure is linked to their nonprofit organization’s overall health and financial sustainability (Stewart & Diebold, 2017). Nonetheless, nonprofit boards may be ill-equipped to carry out these responsibilities (Stewart et al., 2021). Individual board members are asked to play multiple roles simultaneously, often requiring skills they do not possess (Wright & Millesen, 2008). Previously effective boards may become less effective when the organization’s external environment changes dramatically (McMullin & Raggo, 2020), or volunteer board members may wish to limit their investment in time-intensive board responsibilities (Stewart, 2017). Such challenges to effective governance are exacerbated by intense and growing competition for executive talent in the nonprofit sector (McKee & Froelich, 2016). Under these circumstances, nonprofit boards may fail to monitor financial performance rigorously and otherwise supervise CEOs.

Across sectors, causal factors have been offered to explain board abdication of CEO oversight. Often characterized as rubber stamping of executive initiatives by passive boards, board capture, or circumstances that make boards ineffective in their monitoring and resource provisioning functions (Hillman & Dalziel, 2003), has been variously attributed to information asymmetries crafted by or favoring management (Nili & Kastiel, 2017), conflicts of interest (i.e., appointment, retention, and compensation decisions by management) on the part of board members (Blodgett et al., 2013), government funding (i.e., securing public funding becomes the CEO’s top priority; Guo, 2007), and directors’ behavior biased by a desire to enhance social or other relationships with management or board members aligned with management (Page, 2009).

We raise the possibility that the boards of the largest nonprofits unwittingly neglect their duties to oversee the financial health of their organizations. A lack of expertise and the time-intensive and stressful nature of executive supervision may combine to cause nonprofit board capture, manifesting as longer executive tenure in the face of poor financial performance. Using an extensive panel of nonprofit data, we explicitly evaluate whether nonprofit boards are as responsive as for-profit boards to poor CEO performance on financial measures. Our analyses indicate that nonprofit boards are much less responsive across a range of plausible operationalizations of financial performance, and that the largest, most sophisticated nonprofit organizations show little evidence that their boards incorporate changes in revenues, expenses, savings, or profit into CEO turnover decisions. This contrasts sharply with the results of our for-profit analyses, in which changes in all those variables produce statistically and materially significant estimates. However, when we measure performance in terms of impacts to net assets as a fraction of total assets, nonprofit boards’ responsiveness is comparable to that of for-profit boards. These findings are consistent with a pattern wherein boards of the largest nonprofits provide relatively lax financial oversight most of the time, but act when financial results start to affect long-term solvency.

Literature Review

The Generic Nature of Organizations and Its Implications for Board Practice

In 2017, Bromley and Meyer observed,

As traditionally separate sectors shift toward formal, and more standardized, forms of organization, the historical distinctions between [business, government, and nonprofit organizations] come to increasingly rest in legal and scientific definitions rather than in functional purposes. Today, we know a nonprofit is such because it has the appropriate legal status. It becomes harder and harder to determine an organization’s form . . . based on functional activity alone . . . [S]ector blurring is not simply a transfer of new practices into the world of nonprofits and government. All sectors are changing in similar ways. (p. 957, 939)

Across sectors, Pope et al. (2018, p. 1312) document “an expanded focus on transparency, accountability, and trustworthiness—and more positively, on values, citizenship, and leadership . . . [and] the result has been a considerable degree of cross-sectoral organizational isomorphism.” Nonetheless, numerous authors and some participants in the sector itself argue that nonprofit organizations are materially different from private and public entities (e.g., McFarlan, 2017; Trautman & Ford, 2019), which could lead to different performance standards across sectors (e.g., Boland et al., 2020; Kugel & Mercado, 2024). We take issue with this “sectoral exceptionalism” and the negative impact it may have on board supervision and nonprofit performance. Given the increasingly important work of nonprofit organizations, nonprofit boards should be expected to execute their fiduciary responsibilities in a manner that is similar, if not superior, to that of for-profit boards.

Scholars who disagree with this assertion most commonly cite three domains wherein material cross-sectoral differences are thought to reside: sources of funds, the profit incentive, and the distribution of earnings/ownership. We consider each one in turn.

Sources of Funds

While donations are not a focus of for-profit organizations, nonprofit organizations’ reliance on individual and institutional donors is remarkably similar to for-profit organizations’ reliance on individual and institutional investors. In both settings, these individuals and institutions range in sophistication from individual donors or investors to large foundations or private equity firms. While nonprofit organizations do not sell ownership stakes (i.e., equity), they often rely on operating cash flows and debt for capital (Bahceli, 2020). As with for-profit organizations, investment income can be an important component of their financial position. Moreover, “social mutual funds,” “social venture capital funds,” and “foundation growth capital funds” now serve as important financial intermediaries in the nonprofit sector (Kaplan & Grossman, 2010), emphasizing accountability (i.e., impact return on investment). Indeed, in many nonprofit organizations, operating and capital “fundraising” is substantially augmented with a combination of debt, investment income, operating surplus, and earnings from for-profit subsidiaries (Lyons et al., 2007) and an overall alignment with the for-profit sector (Meier & Schnurbein, 2024). Thus, like many in the private sector, donor-investors in the nonprofit sector have two bottom lines, both a financial and a social return (MacQuillin & Sargeant, 2019).

The Profit Incentive

It is often asserted that nonprofit organizations have “no profit incentive” (Stewart & Diebold, 2017; Young, 2013). When profit is understood simply as an excess of revenue over expenses, however, solvency demands that all organizations earn some profit over the long term (Bowman, 2011). Furthermore, retaining sufficient earnings as net assets allows nonprofit organizations to be able to act on opportunities and reduce financial vulnerability (Calabrese, 2012). Nonprofit service reductions and closures due to financial exigencies are common (O’Leary, 2021). For these reasons, we argue that nonprofit organizations have a clear, if less emphasized, “profit” incentive.

Is profit-making the central purpose of nonprofit organizations? Certainly, the answer is no, but neither is it the central purpose of many for-profit firms (George et al., 2023). The purpose of many top-performing, private, for-profit entities is to maximize their value to society as a whole (Sisodia et al., 2014). These entities are as mission-driven as nonprofit organizations and, according to Drucker et al. (2015), arguably more so, given that nonprofit organizations are prone to mission revision, dilution, or abandonment in the face of financial pressure or donor influence.

The Distribution of Earnings/Ownership

Arguably, the most compelling distinguishing characteristic of nonprofit organizations and their for-profit counterparts is the “prohibition on the distribution of profit to [the nonprofit organization’s] members or supporters” (Lyons et al., 2007, p. 99). Without equity or profit incentives, this logic contends that nonprofit board members acting as agents may fail to act on behalf of a broad set of principals who are widely dispersed and who have little means of independently acquiring information about the performance of the organization (van Puyvelde et al., 2012). Adhering to the prescriptions of stewardship and stakeholder theories may alleviate some of the agency costs that arise from information asymmetries and goal misalignment (Coule, 2015), yet the absence of profit incentives and the associated lack of investor oversight underline the need for board supervision of nonprofit CEOs (Kaplan & Grossman, 2010).

Board oversight is especially important given that the National Center for Charitable Statistics (2024) reports that over 1.8 million nonprofit organizations are registered in the United States with assets exceeding six trillion dollars. The value created by those organizations is less clear. At a societal level, Kaplan and Grossman (2010) argue that we are earning a low rate of return on the trillions of dollars invested in the nonprofit sector. Independent of whether nonprofits are efficient, it is difficult to measure the true effectiveness of nonprofit organizations when outcomes of interest are distal, and the term “social impact” takes on different meanings in different contexts (Ebrahim & Rangan, 2014; Hall et al., 2015; Rawhouser et al., 2019; Stephan et al., 2016). At a more granular level, without sound financial management and a focus on creating measurable impact, individual nonprofit organizations are likely to underfulfill their missions and underserve those in great need of their services. Bradlach et al. (2008) succinctly capture this sentiment:

U.S. nonprofits are being asked to take on an increasing share of society’s most important and difficult work. At the same time, the expectations being placed on these organizations to show results—by their staff members, their boards, and public and private donors—are rising. (p. 1)

The implications for board practice are clear. Like many for-profit boards, nonprofit boards have the challenging task of supervising CEOs through a lens of organizational performance that considers financial sustainability a necessary but insufficient achievement (Sisodia et al., 2014). Given the difficulty of measuring social impact and the centrality of financial health for long-term mission fulfillment, financial performance remains a salient indicator of CEO effectiveness in nonprofit organizations. In this regard, we argue that nonprofit boards should hold CEOs accountable for financial performance in a manner that is at least comparable with that of for-profit boards.

There is an extensive literature confirming the positive relationship between for-profit CEO tenure and financial performance (Cao et al., 2021), but this relationship is less well examined in the nonprofit sector. Stewart and Diebold (2017, p. 757) find that poor financial performance, operationalized as significant expense reductions, predicts nonprofit executive turnover. Type-specific evidence similarly suggests a relationship, with Stilwell (2021) finding that university boards respond to financial performance, and Brickley and van Horn (2002) finding much the same in nonprofit hospitals. That said, the extant literature is largely parochial in the type of nonprofit studied, limited in sample size, and disparate in how it measures financial performance.

Volunteer Board Capacity to Effectively Supervise Executive Leadership

While nonprofit board members are presumably recruited “for their skills, expertise, fundraising ability, and social networks,” joining a volunteer board is often motivated by desired personal outcomes (Ward & Miller-Stevens, 2021, p. 316). For others, the desirability of nonprofit board service is related to prestige, power, or “social cachet” (Woodroof et al., 2021). Incongruity between organizational needs and candidate motivations can lead to the all-too-frequent nonprofit error of recruiting board members in response to their articulated passion for the mission or their donor potential rather than their demonstrated expertise and commitment (McFarlan, 2017).

Nonprofit board members can reasonably expect to invest between 8 and 15 hr per month in routine board tasks (Ha et al., 2017; Klein, 2020), with leadership roles requiring a greater time commitment. In particular, searches for a new CEO often require extraordinary investments of time and effort. In the for-profit sector, CEO searches typically last 90 days (Murray, 2022), but searches for university presidents often take 6 months or more of fairly constant participation by the search committee and its chair (Korn, 2024).

Once a new nonprofit CEO is hired, the effort required to manage the executive transition can be “daunting” (Stewart et al., 2021). Even if board members possess the expertise and skills needed by the organization, internal and external inertial forces make substantive organizational change difficult and hazardous (Hager & Yoon, 2024). In addition, the amount of time required to complete important tasks, such as replacing and managing the transition of the CEO, may be prohibitive. “Unlike the for-profit board, where time demands are somewhat predictable, the longer a board member remains on a nonprofit board, the more time is demanded . . . sometimes . . . more time than a member’s regular paid business position” (Epstein & McFarlan, 2011, p. 18).

Clearly, some board members lack the requisite level of commitment. Nonprofit executives and CEOs of publicly traded for-profit organizations routinely comment on the lack of meaningful preparation and participation among their board members (Woodroof et al., 2021). Unprepared or unengaged members impede substantive discussions and critical decision-making in board meetings (van Puyvelde et al., 2018). More problematic is the task of recruiting board members to take on critical, highly visible leadership roles such as managing executive evaluations or transitions. Reluctance to take on these roles could result in accepting subpar CEO performance and “kicking the can down the road.” Consistent with this view, BoardSource (2021) reports that 47% of nonprofit executives had no formal, written performance evaluation in the previous year and one in five had never had a formal, written evaluation of their performance.

Unsurprisingly, board members’ investment of time and effort in preparing for and participating in board activities is positively related to the effective supervision of CEOs and the financial performance of organizations (Minichilli et al., 2009). Gazley and Nicholson-Crotty (2018) similarly find positive relationships between board member training, engagement in strategic decision-making, and selective recruitment on board performance. Yet considerable empirical evidence suggests that nonprofit boards underperform in their oversight of the CEO (Stewart, 2017). “[B]oards generally have too little time, too little experience with nonprofit management principles, too little expertise in the business of the organization (field of service), and too little skill at governing as a group to be able to handle the governing role well” (Allison, 2002, p. 348).

Method

We now turn to a discussion of our methods and the data we use to examine how both nonprofit and for-profit boards respond to financial performance when deciding to replace CEOs.

Analysis

To determine whether nonprofit boards are as responsive as for-profit boards to financial performance, we use Cox proportional hazard regressions with multiple failures, which allow for the possibility that a single organization experiences turnover more than once during our study. Survival, or time-to-event, models such as the ones we use are still relatively rare in research on nonprofit executive turnover; probit regressions (e.g., Firth et al., 2014) and logistic regressions (e.g., Brickley & van Horn, 2002) are more common. Turnover is inherently binary, so probit and logit models are capable of estimating the likelihood of its occurrence at a particular point in time (Lee, 2019), while survival models combine the ability to handle binary outcomes with an explicit treatment of how the likelihood of turnover changes with CEO tenure. Consequently, survival models are better suited to research in executive turnover, because we are interested not only in the correlation between predictors and CEO replacement, but also whether those predictors affect the chance of turnover conditional on how long the CEO has already served.

A key concept in survival models is the baseline hazard rate, which represents the underlying risk of an event (such as CEO turnover) over time, assuming no influence from covariates. One way to think about this is to imagine the data being reorganized so that every CEO’s first year on the job is aligned at the same starting point. With the data structured this way, the baseline hazard rate essentially reflects the proportion of CEOs who turn over after their first year, second year, and so on. This gives us a time-dependent “expected” probability of turnover based solely on tenure. Given a baseline hazard rate, the model can then estimate how each covariate influences the risk of turnover relative to the baseline.

This time-to-event information in Cox models affords two additional advantages. The first advantage is that these models account for the right-censoring of data, which occurs when a research subject has not experienced the outcome at the end of the relevant period but could possibly experience it later (Cox, 1972). Accounting for right-censoring avoids the bias that occurs when simpler models treat the last sampled year as equivalently informative regardless of how long each CEO had served prior to the final observation (Annesi et al., 1989).

The second advantage is that total hazards are the product of individual coefficient estimates and the baseline hazard rate; allowing Cox models the flexibility to account for poor financial performance having a greater absolute impact on turnover earlier or later in a CEO’s tenure, even as they assume that its proportional impact is the same. Logistic models treat failure after 1 year exactly the same as failure after a decade, and while they can be modified to incorporate tenure information, those changes quickly run into problems of overfitting, making time-to-event models generally preferred (van der Net et al., 2008).

Data and Variables

Our data on nonprofit organizations and their leadership come from the Internal Revenue Service 990 Electronic Filing (eFile) Database and the Exempt Organization Business Master files from 2009 to 2019 (Lecy, 2023). We combine information from the header, summary, and Part VIII portions of the eFile data by employer identification number and tax year, and then merge in nonprofit industry classifiers from the business master file using the same identifiers. As tax years are defined differently from years in our for-profit dataset, we align the two by defining the year of each nonprofit observation as the calendar year of the beginning of the tax period in cases where the fiscal year ends before June and the calendar year of the end of the tax period in all other cases.

Our for-profit data comes from Compustat and Execucomp. These datasets provide information on financial performance and leadership for all S&P 1500 firms. We multiply all for-profit financial variables by one million and all for-profit compensation variables by one thousand so that they are measured in dollars, which are the same units we use for our nonprofit variables. We only include data from 2009 to 2019 to match the availability of data for nonprofit organizations.

To ensure similar statistical power in our two regressions and to focus our nonprofit results on the most sophisticated boards, we limit our nonprofit data to only the largest organizations. We accomplish this by calculating an annual ranking of nonprofit organizations by revenue and retaining every observation from each organization whose revenue was in the top 0.55% of any year for which data is available. Retaining any firm that crosses a particular size threshold is similar to the approach taken by Execucomp, which retains data for any firm that enters the S&P 1500 in any year. The statistical and economic significance of our estimates are naturally sensitive to this assumption, but tests that vary this size cutoff confirm the overall nature of our conclusions. As we have no theoretical reason for focusing on a different size range of nonprofits, we choose to emphasize statistical comparability by including approximately the same number of for-profit and nonprofit observations.

For the nonprofit organizations, this approach results in a study population of 10,429 organizations with average total revenue of $452 million and a median of $257 million. On average, these organizations have 19.6 voting board members and a median of 15. Nonprofit CEO compensation averages $984,000 with a median of $701,000. About 75% of our observed nonprofit organization-years have CEOs with names typical of males.

In the for-profit study population, our approach results in 10,702 firms with average total revenue of $8.14 billion and a median of $194 billion. These for-profit boards consist, on average, of 8.5 voting board members and a median of 9. On average, for-profit CEO compensation is $6.7 million with a median of $4.9 million. About 96% of the organization-years in our for-profit population have CEOs whose names are typical of males.

Dependent Variable

The outcome we are interested in for both sectors is executive leadership turnover. We define replacing a permanent CEO with a temporary or interim CEO as one transition, and the removal of the interim CEO for a newly appointed CEO as a separate transition. A CEO who was appointed as temporary but was later named the permanent CEO is one transition. In our data, a single organization can have multiple transition events but can only have one transition event per organization-year.

We identified nonprofit CEOs using the key employee data from Part VIII of the eFile data. Our first cut identified the titles of “president,” “executive director,” “CEO,” or “chief executive officer,” including arbitrary capitalizations of these words. We then dropped any organization-year in which more than two individuals were identified, and any other key employees listed in organization-years when a single person was identified. If exactly two people were identified, we retained the person whose title included CEO, and if no such CEO title was present, we dropped the organization-year. To retain as much information as possible, we then added observations using the following sequential identification strategies: titles starting with the first set of keywords, identifiable misspellings of those keywords, variations on the word “manager” that included those keywords, and variations on the word “manager” when no other previous identification had worked.

We identified for-profit CEOs using the “CEOANN” flag in Execucomp, which identifies whether a CEO served for all or most of a year and is determined from company reports filed with the U.S. Securities and Exchange Commission. We then hand-corrected errors, such as CEOs of divisions occasionally being co-listed with the enterprise CEO. We also removed six for-profit organizations that listed permanent co-CEOs.

Independent Variables

While revenue and profitability are the two most common financial variables used in for-profit CEO turnover research, financial indicators that might affect nonprofit performance and CEO turnover are more numerous (Hung & Hager, 2019; Prentice, 2016). Consequently, we explore changes in five separate financial indicators. We operationalize all our financial variables as annual changes, and we winsorize them at the 1% and 99% levels to limit the influence of outliers. Each of our independent variables is lagged at least 1 year, because CEO changes are recorded as happening during the first year of the new CEO’s tenure, denoted as Year t in our notation.

Revenue. Revenue in both populations is measured as total revenue. Because revenue is strictly non-negative, we measure its annual change as the following difference of logs, allowing for changes over multiple calendar years to be annualized if an observation is missing:

Revenue Change = \frac{1}{n} \cdot [\ln (R e v e n u e_{t}) - \ln (R e v e n u e_{t - n})]

Expenses

Total expenses are reported directly within the nonprofit data. In the for-profit data, we calculate total expenses as revenue minus net income. Expenses are considered a good measure of the level of activity in nonprofit organizations and are often used to normalize nonprofit financial ratios (Prentice, 2016). However, as activity levels and expenses are clearly driven by revenue in for-profit organizations, exploration of how expenses affect turnover in that context is rare. Expenses are non-negative, so we measure expense changes as follows:

Expense Change = \frac{1}{n} \cdot [\ln (E x p e n s e_{t}) - \ln (E x p e n s e_{t - n})]

Profitability

Profit is defined as net income in the for-profit data and revenue less expenses in the nonprofit data. Because both profit variables commonly take negative values, we normalize them to reduce skew and retain interpretability without resorting to transformations, such as the inverse hyperbolic sine, which are known to introduce arbitrary unit-dependence (Chen & Roth, 2024). In this case, profit is normalized by revenue to create a profit margin, and its change is measured as follows:

P r o f i t M a r g i n C h a n g e = \frac{1}{n} \cdot [(\frac{P r o f i t_{t}}{R e v e n u e_{t}}) - (\frac{P r o f i t_{t - n}}{R e v e n u e_{t - n}})]

Savings

A more common approach to considering profit in nonprofit financial statement analysis is recasting revenue less expenses as savings, and normalizing it by expenses rather than revenue (Prentice, 2016). This ratio is often called the savings indicator, though we prefer calling it savings margin here because of its similarity with profit margin. We operationalize the change in savings margin as follows:

Savings Margin Change = \frac{1}{n} \cdot [(\frac{P r o f i t_{t}}{E x p e n s e s_{t}}) - (\frac{P r o f i t_{t - n}}{E x p e n s e s_{t - n}})]

Net Assets

Finally, we estimate the effect of changes in net assets, or owners’ equity in the for-profit context. Net assets and equity are both measured as total assets minus total liabilities, and because they are sometimes negative or zero, we normalize them by total assets before measuring their annualized change as follows:

Net Asset Change = \frac{1}{n} \cdot [(\frac{N e t A s s e t s_{t}}{T o t a l A s s e t s_{t}}) - (\frac{N e t A s s e t s_{t - n}}{T o t a l A s s e t s_{t - n}})]

Controls

To provide an equivalent set of for-profit and nonprofit tests, we limit our controls to only those variables that we believe are comparable across contexts.

Size of the Board

In the nonprofit setting, Stewart and Diebold (2017) find that the larger the board, the longer the tenure of executives; Stilwell (2021) replicates these findings in the private, nonprofit university setting. These findings are consistent with an idea examined by Aggarwal et al. (2012), namely that powerful nonprofit CEOs build larger boards to entrench themselves and insulate themselves from scrutiny. To control for this, our tests include the natural log of the number of voting members of the board.

Tenure of Executive

Lengthy tenure appears to enhance boards’ trust of and reliance on the CEO. Board independence has been found to decline over an executive’s tenure; conversely, the intensity of board supervisory engagement “jumps” when a new executive is appointed (Graham et al., 2020). Tenure has also been found to be important in determining board responses to malfeasant CEO behavior, with longer tenures associated with more lenient responses (Tillotson & Tropman, 2014). As discussed above, CEO tenure is accounted for in survival models through the baseline hazard function and therefore should not be added as a separate control variable.

Total CEO Compensation

Core et al. (1999) and others (e.g., Bouteska & Mefti-Wali, 2021) report that CEOs earn greater compensation when governance structures are less effective. On the other hand, compensation has been characterized as an indicator of CEO performance expectations and has, in the nonprofit sector, been a predictor of turnover (Stewart & Diebold, 2017). Others argue that CEO compensation is an indicator of CEO power relative to the board (Shen et al., 2010) and thus has a negative impact on turnover. Regardless of which view is correct, we include the natural log of total CEO compensation as a control in our models.

CEO Gender

Several studies find that, across sectors, male CEOs are more likely to be replaced than female CEOs in circumstances of declining performance (Cooper, 2017; Liu, 2014). In contrast, some studies report that male candidates are more likely to be appointed as CEOs (Norris-Tirrell et al., 2018) and female CEOs are more likely to be dismissed (e.g., Stewart & Diebold, 2017), while Gupta et al. (2020) find that male and female CEOs are equally likely to be fired in underperforming firms.

Despite these inconsistent findings, we include a control for CEO gender in our tests. In the for-profit data, we rely on the CEO gender flag in the Execucomp dataset. For nonprofit organizations, we use the people-parser R package (Lecy, 2020) to obtain the likely gender of each CEO by comparing names with census records. We operationalize the results as an indicator that takes a value of 1 if the CEO has a typically male name and a value of 0 otherwise.

We control for the size of each organization by including the natural log of revenue in our models. We also use indicator variables to control for the month of the fiscal year end and industry membership. Our for-profit industries are defined using the 12 industry Fama-French mapping of SIC codes (Fama & French, 1997), while our nonprofit industries are defined using the 12 major National Taxonomy of Exempt Entities industry classifications. We drop all mutual benefit and religious nonprofit organizations, as well as any nonprofit organizations for which the industry is unknown, because these organizations are likely to have idiosyncratic approaches to choosing and monitoring their CEO.

Tables 1 and 2 present summary statistics of transformed variables for both populations.

Table 1.

Nonprofit Summary Statistics of Transformed Variables.

Variable	Obs.	Mean	Min	Median	Max
Revenue Change t-1	10,429	0.0639	-0.838	0.0497	1.189
Revenue Change t-2	10,429	0.0659	-0.838	0.0505	1.189
Revenue Change t-3	10,429	0.0736	-0.838	0.0535	1.189
Expense Change t-1	10,429	0.0637	-0.569	0.0484	1.015
Expense Change t-2	10,429	0.0645	-0.569	0.0487	1.015
Expense Change t-3	10,429	0.0670	-0.569	0.0480	1.015
Profit Margin Change t-1	10,429	0.000471	-0.563	0	0.581
Profit Margin Change t-2	10,429	0.00179	-0.563	0	0.581
Profit Margin Change t-3	10,429	0.00710	-0.563	0.00136	0.581
Savings Margin Change t-1	10,429	0.00169	-0.892	0	0.940
Savings Margin Change t-2	10,429	0.00314	-0.892	0	0.940
Savings Margin Change t-3	10,429	0.00816	-0.892	0.00142	0.940
Net Assets Change t-1	10,429	0.00455	-0.465	0.00596	0.372
Net Assets Change t-2	10,429	0.00423	-0.465	0.00640	0.372
Net Assets Change t-3	10,429	0.00394	-0.465	0.00639	0.372
Board Size t-1	10,429	2.776	0	2.708	10.81
Size t-1	10,429	19.43	7.574	19.32	23.64
Compensation t-1	10,429	12.78	0	13.41	16.55
Male Name t-1	10,429	0.754	0	1	1

Table 2.

For-Profit Summary Statistics of Transformed Variables.

Variable	Obs.	Mean	Min	Median	Max
Revenue Change t-1	10,702	0.0561	-0.591	0.0505	0.810
Revenue Change t-2	10,702	0.0642	-0.591	0.0543	0.810
Revenue Change t-3	10,702	0.0721	-0.591	0.0579	0.810
Expense Change t-1	10,702	0.0549	-0.700	0.0494	0.818
Expense Change t-2	10,702	0.0635	-0.700	0.0548	0.818
Expense Change t-3	10,702	0.0681	-0.700	0.0562	0.818
Profit Margin Change t-1	10,702	0.00253	-0.793	0.00142	0.925
Profit Margin Change t-2	10,702	0.00274	-0.793	0.00128	0.925
Profit Margin Change t-3	10,702	0.00683	-0.793	0.00274	0.925
Savings Margin Change t-1	10,702	0.00204	-0.747	0.00161	0.760
Savings Margin Change t-2	10,702	0.00151	-0.747	0.00148	0.760
Savings Margin Change t-3	10,702	0.00458	-0.747	0.00317	0.760
Equity Change t-1	10,767	-0.00735	-2.071	-0.000650	2.404
Equity Change t-2	10,767	-0.00748	-2.071	-0.00148	2.612
Equity Change t-3	10,767	-0.00536	-2.071	-0.00109	2.612
Board Size t-1	10,702	2.080	0	2.079	3.401
Compensation t-1	10,702	15.21	0	15.35	18.79
Size t-1	10,702	21.38	11.51	21.33	26.96
Male Name t-1	10,702	0.956	0	1	1

Statistical Models

Our regression models define their hazard as CEO turnover in Year t, with organizations retained after a turnover event using the ordered failure event approach of Andersen and Gill (1982). The models use two general formulations: One tests whether the previous year’s financial performance affects CEO turnover, and the other includes three previous years of financial performance data rather than just one. Thus, we estimate the following two models for each of our five financial variables:

\begin{array}{l} h (t | x) = h_{0} (t) \cdot \exp (β_{1} \cdot F i n C h g_{t - 1} + γ_{1} \cdot S i z e_{t - 1} + γ_{2} \cdot B d S i z e_{t - 1} \\ + γ_{3} \cdot C o m p_{t - 1} + γ_{4} \cdot M a l e N a m e_{t - 1} + φ_{1} \cdot I n d_{i} + φ_{2} \cdot F Y E_{i}) \end{array}

\begin{array}{l} h (t | x) = h_{0} (t) \cdot \exp (β_{1} \cdot F i n C h g_{t - 1} + β_{2} \cdot F i n C h g_{t - 2} + β_{3} \cdot F i n C h g_{t - 3} + γ_{1} \cdot S i z e_{t - 1} \\ + γ_{2} \cdot B d S i z e_{t - 1} + γ_{3} \cdot C o m p_{t - 1} + γ_{4} \cdot M a l e N a m e_{t - 1} + φ_{1} \cdot I n d_{i} + φ_{2} \cdot F Y E_{i}) \end{array}

where $h_{0}$ denotes the baseline hazard rate; β’s are slopes on financial indicator variables; FinChg is the change in either revenue, expenses, profitability, savings, or net assets depending on the variable examined in each test; γ’s are slopes on control variables; Size represents organization scale through revenue; BdSize represents the number of voting board members; Comp represents CEO compensation; MaleName represents CEO gender; φ’s are vectors of slopes on indicator variables; Ind_i is a vector of industry membership flags; and FYE_i is a vector of fiscal year end month flags. We report all coefficients from these tests as hazard ratios, which represent multiplicative shifts in the baseline chance of turnover from a one-unit change in each independent variable. As such, their statistical significance is evaluated relative to 1, which would indicate no change in the chance of turnover.

Results

We first investigate board responses to changes in revenue (Table 3). The two left-hand models show that revenue changes in nonprofit organizations have little effect on CEO replacement decisions. While the point estimates in both models suggest that revenue changes in nonprofit organizations might influence how those boards view the need to replace their CEOs, none of the hazard ratios are statistically significantly different from 1. As such, these tests do not support the argument that boards of the largest nonprofits respond to changes in revenue when choosing to replace their CEOs.

Table 3

Comparison of Reaction to Revenue Changes (Nonprofit − Total Revenue; For-profit − Total Revenue).

Variable	Nonprofit	Nonprofit3 years	For-profit	For-profit3 years
Revenue Change t-1	0.875 (0.116)	0.864 (0.119)	0.366*** (0.060)	0.389*** (0.066)
Revenue Change t-2		0.876(0.122)		0.814(0.131)
Revenue Change t-3		0.967(0.121)		0.892(0.141)
Size t-1	1.053(0.034)	1.056(0.034)	1.025(0.020)	1.027
Size t-1	1.053(0.034)	1.056(0.034)	1.025(0.020)	(0.019)
Board Size t-1	0.989(0.047)	0.988(0.047)	1.396** (0.154)	1.375** (0.152)
Compensation t-1	0.941*** (0.007)	0.941*** (0.007)	0.937** (0.019)	0.937** (0.019)
Male Name t-1	1.011(0.065)	1.011(0.065)	1.269(0.169)	1.277(0.169)
Observations	10,429	10,429	10,702	10,702
FYE Month FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes

Note. Clustered, robust standard errors are in parentheses.

p < .05, **p < .01, ***p < .001.

This is particularly interesting when compared with for-profit boards’ responsiveness to changing revenues. In both of our for-profit models, we find that prior year revenue changes have a strong, statistically significant effect on the likelihood of CEO removal (β = 0.366 and 0.389, p < .001). This predictor is log-natural transformed, necessitating care when interpreting the coefficient. Here, a for-profit CEO whose firm recorded 5% revenue growth would be between 4.5% and 4.9% less likely to be removed the following year, all else equal (1 − β ^ ln(1.05)). Revenue changes from 2 and 3 years prior have nonsignificant estimated effects, which implies that for-profit boards respond most strongly to recent revenue changes. Combined, these results indicate that the decision to replace a CEO in our sample of nonprofit boards is significantly less influenced by revenue changes than a similar decision in a for-profit board.

Our controls have relatively stable estimates in each of the five models, with compensation being the most consistently important control. This suggests that both nonprofit and for-profit boards display a remarkably similar tendency to be tentative when considering replacing more highly compensated CEOs. This effect is relatively small, with a 5% increase in pay being linked to roughly a 0.3% reduction in turnover hazard (1 − β ^ ln(1.05)). Interestingly, the number of voting board members plays very different roles in these two populations. We find no evidence that the size of the board has any impact on CEO replacement decisions in the nonprofit sector, but larger boards are much more likely to replace a CEO in the for-profit sector. No other control variables were significant at the 0.05 level.

Based on the findings of Never (2013) and Stewart and Diebold (2017), one might expect expense changes to play a role in nonprofit CEO replacement decisions, because expense changes are often employed as measures of organizational distress. In contrast, our tests of the role of expense changes on the decision to retain a CEO by nonprofit boards show that, in fact, most of these estimates are less impactful (i.e., closer to 1) than our corresponding estimates for revenue changes. Broadly, these tests once again find no support for the argument that the boards of the largest nonprofits respond to changes in expenses when choosing to replace their CEOs (Table 4).

Table 4.

Comparison of Reaction to Expense Changes (Nonprofit − Total Expense; For-profit − Revenue Less Net Income).

Variable	Nonprofit	Nonprofit3 years	For-profit	For-profit3 years
Expense Change t-1	0.891(0.163)	0.908(0.164)	0.721* (0.109)	0.732* (0.110)
Expense Change t-2		1.008(0.183)		0.895(0.124)
Expense Change t-3		0.866(0.155)		0.845(0.121)
Size t-1	1.050(0.034)	1.050(0.033)	1.015(0.020)	1.016(0.020)
Board Size t-1	0.989(0.047)	0.988(0.047)	1.460** (0.167)	1.443** (0.165)
Compensation t-1	0.941*** (0.007)	0.941*** (0.007)	0.938** (0.019)	0.937** (0.019)
Male Name t-1	1.011(0.065)	1.010(0.065)	1.252(0.170)	1.260(0.170)
Observations	10,429	10,429	10,702	10,702
FYE Month FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes

Note. Clustered, robust standard errors are in parentheses.

p < .05, **p < .01, ***p < .001.

One might question whether our tests of how expense changes relate to turnover make conceptual sense in a for-profit context. On one hand, large positive changes in expenses could signal organizational inefficiency or mismanagement and could be detrimental to CEO tenure. On the other hand, large positive changes in expenses could signal increased activities that will generate revenue in future periods and could protect CEOs from dismissal. Given this ambiguity, we refrain from discussing these results in detail. Instead, we reiterate that the results show material differences in how nonprofit and for-profit boards interpret changes in expenses when deciding whether to retain or replace a CEO.

Next, we investigate board responsiveness to changes in profit margin (Table 5). Once again, both the 1- and 3-year nonprofit models show no statistically discernible effect of profit margin changes on the likelihood of CEO turnover. While few researchers might have expected nonprofit boards to focus on profit margins, all the estimated hazard ratios are less than 1, and a change to profit margin 2 years earlier has a particularly impactful point estimate. Overall, though, these tests show no evidence that the boards of the largest nonprofits respond to changes in profitability when making CEO replacement decisions.

Table 5.

Comparison of Reaction to Profit Margin Changes (Nonprofit − Rev Less Exp/Revenue; For-profit − Net Income/Revenue).

Variable	Nonprofit	Nonprofit3 years	For-profit	For-profit3 years
Profit Margin Change t-1	0.934(0.227)	0.803(0.222)	0.491*** (0.097)	0.421*** (0.087)
Profit Margin Change t-2		0.642(0.185)		0.526** (0.113)
Profit Margin Change t-3		0.913(0.234)		0.922(0.196)
Size t-1	1.049(0.034)	1.055(0.034)	1.020(0.020)	1.022(0.021)
Board Size t-1	0.989(0.047)	0.991(0.047)	1.460** (0.164)	1.447** (0.164)
Compensation t-1	0.941*** (0.007)	0.941*** (0.007)	0.937*** (0.019)	0.937** (0.020)
Male Name t-1	1.011(0.065)	1.012(0.066)	1.234(0.168)	1.232(0.168)
Observations	10,429	10,429	10,702	10,702
FYE Month FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes

Note. Clustered, robust standard errors are in parentheses.

p < .05, **p < .01, ***p < .001.

For-profit boards are highly responsive to changes in profit margins in our tests. Both the 1- and 3-year models show statistically and economically significant responses to a single year’s change in profit margin (β = 0.491 and 0.421, p < .001), with the 3-year model indicating that changes 2 years earlier are additionally influential. Because this predictor can take negative values, it was normalized rather than log transformed, making its interpretation more straightforward. A 5% increase in profit margins over 1 year lowers the hazard of removal by 3.5% or 4.2% (1 − β ^ 0.05). This effect attenuates only slightly in the second year, with a 5% increase in profitability 2 years earlier being estimated to cause a reduction of 3.2% in the likelihood of turnover. These results further strengthen the case that nonprofit boards are significantly less responsive to financial performance than for-profit ones with respect to CEO replacement decisions.

If the lack of responsiveness of nonprofit boards in our results is driven by our use of measures that fail to capture the kind of financial performance important in nonprofit organizations, then a test using the savings indicator (i.e., savings margin) should perform better. Our results using savings margin changes do improve marginally for the nonprofit data, but are again underwhelming, especially when focusing on the previous year’s change, where neither model estimates a statistically significant effect of savings margin changes on CEO turnover (Table 6). However, the estimated effect of changes in the savings margin 2 years earlier does show statistical significance at the 10% level, which aligns with the timing of the large point estimate we saw in the profit margin tests. This suggests that savings margin is a better indicator of CEO turnover in nonprofit organizations than profit margin, which is not surprising. That said, our estimate for board responsiveness to savings margin is delayed by an additional year compared to what we would expect from a truly responsive board.

Table 6.

Comparison of Reaction to Savings Margin Changes (Nonprofit − Rev Less Exp/Expenses; For-profit − Net Income/Expenses).

Variable	Nonprofit	Nonprofit3 years	For-profit	For-profit3 years
Savings Margin Change t-1	0.905(0.138)	0.816(0.149)	0.470*** (0.089)	0.330*** (0.081)
Savings Margin Change t-2		0.739(0.133)		0.448*** (0.099)
Savings Margin Change t-3		0.943(0.146)		0.690(0.147)
Size t-1	1.050(0.034)	1.055(0.034)	1.019(0.020)	1.021(0.020)
Board Size t-1	0.989(0.047)	0.991(0.047)	1.467** (0.164)	1.452** (0.163)
Compensation t-1	0.941*** (0.007)	0.941*** (0.007)	0.937** (0.019)	0.938** (0.020)
Male Name t-1	1.011(0.065)	1.013(0.066)	1.234(0.168)	1.233(0.167)
Observations	10,429	10,429	10,702	10,702
FYE Month FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes

Note. Clustered, robust standard errors are in parentheses.

p < .05, **p < .01, ***p < .001.

We again find strong evidence that for-profit boards base their CEO replacement decisions on financial performance. Positive changes in the savings margin 1 year earlier correspond with a reduced hazard of removal in both for-profit models, with a 5% increase in savings margin reducing the likelihood of dismissal by 3.7% or 5.4% (1 − β ^ 0.05). We also find evidence that the savings margin in prior years is a particularly sticky signal affecting turnover decisions, with savings margin changes in Year t-2 also significantly affecting the hazard of dismissal. Even though savings margin is a nonprofit financial ratio, our for-profit results indicate that it has a larger, more enduring, and more statistically stable influence on CEO turnover than profit margin, suggesting a potential role for savings margin in the for-profit CEO turnover literature. Together, these results show that boards of the largest nonprofits may care about changes in savings margin but are less reactive to them than for-profit boards, both because we estimate a less responsive hazard ratio and because our sample of nonprofit boards take an extra year to respond.

Our final tests investigate how boards respond to changes in the ratio of net assets (or equity) to total assets. They show our only unambiguous evidence that nonprofit boards use financial performance to make CEO replacement decisions (Table 7). In fact, we estimate the magnitude and significance of net asset changes for nonprofit boards to be comparable with those of for-profit boards. Both our 1- and 3-year nonprofit models estimate that a 5% increase in the proportion of net assets to total assets decreases the likelihood of dismissal of a nonprofit CEO by about 3.8% (1 − β ^ 0.05). While additional lags of net asset changes are not estimated to be significant drivers of nonprofit CEO turnover, our results suggest that, once a nonprofit CEO’s financial performance substantially affects the long-term financial health of the organization, boards of the largest nonprofits do incorporate that information into their CEO replacement decisions.

Table 7.

Comparison of Reaction to Net Asset Fraction Changes (Nonprofit − Net Assets/Total Assets; For-profit − Equity/Total Assets).

Variable	Nonprofit	Nonprofit3 years	For-profit	For-profit3 years
Net Asset Frac Change t-1	0.461** (0.133)	0.459** (0.130)	0.321** (0.113)	0.325** (0.113)
Net Asset Frac Change t-2		0.649(0.205)		0.517* (0.169)
Net Asset Frac Change t-3		1.183(0.365)		1.061(0.355)
Size t-1	1.051(0.034)	1.052(0.034)	1.015(0.020)	1.015(0.020)
Board Size t-1	0.991(0.047)	0.991(0.047)	1.475** (0.167)	1.472** (0.167)
Compensation t-1	0.941*** (0.007)	0.941*** (0.007)	0.938** (0.019)	0.938** (0.019)
Male Name t-1	1.011(0.065)	1.011(0.065)	1.239(0.169)	1.240(0.168)
Observations	10,429	10,429	10,702	10,702
FYE Month FE	Yes	Yes	Yes	Yes
Industry FE	Yes	Yes	Yes	Yes

Note. Clustered, robust standard errors are in parentheses.

p < .05, **p < .01, ***p < .001.

In our for-profit analysis, changes in equity as a fraction of assets are also estimated to be very important. In both the 1- and 3-year models, a 5% increase in the equity fraction corresponds with approximately a 5.5% decrease in hazard of dismissal (1 − β ^ 0.05). Our 3-year model additionally suggests that for-profit boards further incorporate changes to equity 2 years earlier with only limited attenuation. While this for-profit response differs in magnitude and persistence from the nonprofit response, these tests provide support for the idea that, with respect to CEO replacement decisions, the boards of the largest nonprofits treat impacts to net assets in a way that is comparable to how for-profit boards treat impacts to owners’ equity.

Discussion

We control for several factors—size of the board and the tenure, compensation, and gender of the CEO—typically offered as causes of board capture, and we find that these factors cannot explain the observed differences in oversight. CEO compensation has a relatively similar impact between sectors, and both nonprofit and for-profit boards exhibit a similar hesitancy to replace highly compensated CEOs. Shen et al. (2010) suggest that CEO compensation is a valid measure of CEO power, and our results are consistent with their interpretation. In the for-profit context, the number of voting board members is also highly significant. However, we find no evidence that the size of the board has any impact on CEO turnover in the nonprofit sector. We also find no statistically significant evidence that differences in the size (i.e., total revenue) of these organizations or the CEO’s gender influence turnover.

Our broader results are consistent with the observation that volunteer nonprofit board members, relative to their highly compensated for-profit counterparts, may not possess the necessary time, expertise, or incentives to appropriately supervise the CEOs of nonprofit organizations. To us, the most glaring difference between these boards—volunteer versus paid members—might account for our finding that nonprofit boards are much less responsive to financial performance than for-profit boards. This is particularly salient given that we restricted our data to include only the largest and presumably most sophisticated nonprofit organizations in the United States. Lax financial oversight and other leadership shortcomings contribute to the demise of about 30% of nonprofits within their first 10 years of existence (Ebarb, 2019). We must strive to do better.

In response to these and related concerns, some have suggested compensating nonprofit board members as a possible solution (Blodgett et al., 2013; Unda & Ranasinghe, 2021). Generally, the anticipated positive consequences of nonprofit board compensation are (1) clearly defined roles and responsibilities of board members, (2) individual board member accountability, (3) professional behavior on the part of board members (e.g., preparation, attendance, participation), (4) a diverse pool of potential board members, and (5) the removal of volunteer immunity. The first four of these are basic good governance practices appropriate to all boards (Letts et al., 1999). Those in favor of compensating nonprofit board members argue that compensation (i.e., transforming board members from volunteers to paid supervisors) will cause board members to take their responsibilities more seriously and make board service a financially viable option for those who would otherwise be unable to volunteer. That said, we suggest that the fifth consequence, the removal of volunteer immunity, introduces a more compelling incentive to better performance than compensation per se.

The current attitude of most nonprofit scholars and practitioners is that nonprofit board members should be volunteers. Some of the arguments against compensating board members are practical, such as adhering to donor expectations about their contributions going to the mission, or the potential for compensation to encourage board capture (Philanthropy Roundtable, 2024). On this last point, Blodgett et al. (2013) assert that compensating nonprofit board members creates conflicts of interest and attenuates fiduciary responsibilities. In short, these authors imagine that board compensation will motivate board members to prioritize personal financial gain over their responsibilities as board members.

Other arguments against the wider adoption of compensation for nonprofit board members are legal. According to Philanthropy Roundtable (2024), “[t]he federal Volunteer Protection Act of 1997 provides broad—though not total—immunity from tort claims that might be filed against unpaid volunteers of nonprofit organizations.” If board members are paid, they would lose the protections granted by this law and potentially be discouraged from working with the organization. While some view this as a negative consequence of board member compensation, Dale (1997) has made a compelling case to the contrary. He observes that the Act generally prevents the recovery of damages unless the victim can prove that the harm was caused by willful or criminal misconduct, gross negligence, reckless misconduct, or conscious, flagrant indifference to the rights or safety of the individual harmed by the volunteer. He goes on to argue that the Act results in inadequate risk management procedures and financial irresponsibility. Compensating nonprofit board members removes these impediments to genuine legal accountability.

If, as our results suggest, fiduciary performance by nonprofit boards is relatively ineffective, we should begin to reconsider the prevailing norm of volunteer boards. This is especially salient given the growing economic importance of nonprofit organizations and the blurring of lines between sectors. The loss of immunity associated with compensating nonprofit board members could improve incentives to recruit and develop competent, committed board members who are better able to supervise CEOs. Improved board performance, and its effects on nonprofit impact, should be championed by increasingly sophisticated donors who understand the importance of competent financial oversight.

Future Research

There are several important limitations of our analysis that future research might seek to address. Given that we only observe CEO turnover, and not any details about why the CEO was replaced, the results might be importantly different if future work was able to observe and control for whether retirement, resignation, a better career opportunity, or some other reason caused the turnover event. We are also importantly limited by what CEO characteristics we can observe. The IRS data does not include information on CEO race or age, and both could be important control variables in future analysis. Deeper insight into the nature of the board members, including their networks, their net worth, or their skillsets could also illuminate differences in the effectiveness of their financial oversight.

This effort has also raised a number of important questions. The first concerns the impact of board performance on the longevity of nonprofit organizations. Given that large nonprofit boards appear to be unresponsive except to indicators of significant financial stress, we wonder what actions boards take to reduce their risk of disbandment. Do these organizations return to less dire or even satisfactory financial performance, and how long does that transition take?

It may also be fruitful to examine how nonprofit board oversight and the factors influencing it change as our focus shifts from the largest nonprofits to smaller organizations. Unreported sensitivity tests show that expanding the nonprofit sample to include the largest 1% of organizations generally increases the statistical and economic significance of the response of nonprofit boards, but that further expanding the sample decreases the magnitude of the estimated responses. These tests suggest that there may be a group of nonprofit organizations clustered around the top 1% of nonprofits by revenue that have stronger responses to financial signals, but we see no extant theoretical reason why this might be. Inconsistent results in the literature could well be a function of the size and sophistication of the organizations under study, and this broad sample could serve as an excellent foundation for examining whether that is the case. In addition, some research suggests that boards in different nonprofit industries (e.g., hospital systems, colleges and universities, arts organizations) respond differently to financial performance. Examining and comparing nonprofit board oversight by industry may give us a better understanding of how nonprofit boards operate and how governance can be improved.

Given our recommendation that nonprofit organizations consider compensating their board members, another important area of research is to carefully evaluate the longitudinal performance of CEOs, boards, and organizations wherein board members are compensated, especially if natural or explicitly controlled experiments focused on board compensation can be identified. As Cole (2023, p. 1) observes, “Survey results and anecdotal experience indicate that compensating directors is more common among private foundations, professional and trade associations and educational and cultural institutions,” but nonprofit organizations that compensate their boards are rare. Nonetheless, nonprofit board compensation is becoming more common, especially among credit unions and hospital systems. Tracking and evaluating the performance of these organizations with respect to longevity, financial performance, and the impact of financial information on CEO turnover will yield important insights into the impact of compensating nonprofit board members.

Conclusion

Our results indicate that boards of the largest nonprofits are much less responsive to financial performance than for-profit boards. We find little evidence that these boards incorporate changes in revenues, expenses, savings, or profit margins into CEO turnover decisions. This contrasts sharply with the results of our for-profit CEO turnover analyses, in which all of our financial performance variables produce statistically and materially significant estimates. In addition, the nonprofit estimates are universally weaker than the corresponding for-profit estimates.

The only case in which we observe nonprofit boards reacting to financial information both quickly and decisively is when we estimate changes in the ratio of net assets to total assets, suggesting that the largest nonprofit organizations may only use financial signals once they are large enough to substantially affect the organization’s long-term financial health. Year-over-year changes in revenue, expenses, and savings appear to be largely uninformative in the prediction of CEO turnover at the largest nonprofits.

Given the importance of nonprofit organizations in our society, and the substantial investment we collectively make in their success, we believe that the boards of nonprofit organizations should hold their CEOs to a higher standard of fiscal performance. We acknowledge that there is a long history of scholarship advising against compensating board members; however, we see compensation as an effective remedy for the observed lack of oversight. While it is heartening that nonprofit boards do seem to respond when results affect solvency, earlier reactions to a broader set of signals by highly motivated and fiduciary board members would prevent at least some nonprofit financial exigencies.

Footnotes

Acknowledgements

Monique Bourque and Rebecca Sullivan were instrumental in assisting us in the early phases of this effort. We are most grateful to Professor Jesse Lecy, who provided the Internal Revenue Service eFile data that we analyze herein. In particular, his gracious response to our request for an electronic version of key employee information was central to our ability to conduct this research.

Data Availability Statement

The data that produce the findings reported in this article are available at the National Center for Charitable Statistics (NCCS) IRS 990 Efile Data (https://www.urban.org/tags/national-center-charitable-statistics-data), Compustat Financials (https://wrds-www.wharton.upenn.edu/), and Execucomp ().

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.

ORCID iDs

Kawika Pierson

Debra Jones Ringold

Author Biographies

Kawika Pierson applies advanced methods to interesting data across management disciplines, working to uncover compelling explanations for observed behavior.

Colin Birkhead researches how social conditions within and around organizations shape macro-organizational behavior.

Debra Jones Ringold has explored many areas of marketing and public policy.

References

Aggarwal

R. K.

Evans

M. E.

Nanda

(2012). Nonprofit boards: Size, performance and managerial incentives. Journal of Accounting and Economics, 53(1–2), 466–487.

Allison

(2002). Into the fire: Board and executive transitions. Nonprofit Management & Leadership, 12(4), 341–351. https://doi.org/10.1002/nml.12402

Andersen

P. K.

Gill

R. D.

(1982). Cox’s regression model for counting processes: A large sample study. Annals of Statistics, 10, 1100–1120.

Annesi

Moreau

Lellouch

(1989). Efficiency of the logistic regression and Cox proportional hazards models in longitudinal studies. Statistics in Medicine, 8(12), 1515–1521.

Bahceli

(2020, August 11). Facing pandemic squeeze, universities hit bond markets for cheap cash. Reuters. https://www.reuters.com/article/idUSKCN257219

Blodgett

M. S.

Melconian

L. J.

Peterson

J. H.

(2013). Evolving corporate governance standards for healthcare nonprofits: Is board of director compensation breach of fiduciary duty? Brooklyn Journal of Corporate, Financial & Commercial Law, 7(2), 443–474.

BoardSource. (2021). Leading with Intent: BoardSource Index of Nonprofit Board Practices. https://leadingwithintent.org

Boland

C. M.

Harris

E. E.

Petrovits

C. M.

Yetman

M. H.

(2020). Controlling for Corporate Governance in Nonprofit Research. Journal of Governmental & Nonprofit Accounting, 9(1), 1–44. https://doi.org/10.2308/JOGNA-17-017

Bouteska

Mefteh-Wali

(2021). The determinants of CEO compensation: New insights from United States. Journal of Applied Accounting Research, 22(4), 663–686. https://doi.org/10.1108/JAAR-08-2020-0176

10.

Bowman

(2011) Financial capacity and sustainability of ordinary nonprofits. Nonprofit Management and Leadership, 22(1), 37–51. https://doi.org/10.1002/nml.20039

11.

Bradlach

J. L.

Tierney

T. J.

Stone

(2008). Delivering on the promise of nonprofits. Harvard Business Review, 86, 1–21.

12.

Brickley

J. A.

van Horn

R. L.

(2002). Managerial incentives in nonprofit organizations: Evidence from hospitals. The Journal of Law and Economics, 45(1), 227–249.

13.

Bromley

Meyer

J. W.

(2017). “They are all organizations”: The cultural roots of blurring between the nonprofit, business, and government sectors. Administration & Society, 49(7), 939–966. https://doi.org/10.1177/0095399714548268

14.

Calabrese

T. D.

(2012). The accumulation of nonprofit profits: A dynamic analysis. Nonprofit and Voluntary Sector Quarterly, 41(2), 300–324.

15.

Cao

Syed

(2021). A meta-analysis of the relationship between chief executive officer tenure and firm financial performance: The moderating effects of chief executive officer pay and board monitoring. Group & Organization Management, 46(3), 530–563. https://doi.org/10.1177/1059601121989575

16.

Chen

Roth

(2024). Logs with zeros? Some problems and solutions. The Quarterly Journal of Economics, 139(2), 891–936.

17.

Cole

D. F.

(2023, June 18). To compensate or not to compensate not-for-profit board members. Dana F. Cole Company LLP. https://www.danacole.com/news/not-profit/compensate-or-not-compensate-not-profit-board-members

18.

Cooper

(2017). Corporate social responsibility, gender, and CEO turnover. Managerial Finance, 43(5), 528–544. https://doi.org/10.1108/MF-02-2016-0049

19.

Core

J. E.

Holthausen

R. W.

Larcker

D. F.

(1999). Corporate governance, chief executive officer compensation, and firm performance. Journal of Financial Economics, 51, 371–406.

20.

Coule

T. M.

(2015). Nonprofit governance and accountability: Broadening the theoretical perspective. Nonprofit and Voluntary Sector Quarterly, 44(1), 75–97.

21.

Cox

D. R.

(1972). Regression models and life-tables. Journal of the Royal Statistical Society: Series B (Methodological), 34(2), 187–202.

22.

Dale

H. P.

(1997). On the occasion of the tenth anniversary of Peter Swords as president of the Nonprofit Coordinating Committee. Century Association (transcript obtained through personal communication).

23.

Drucker

P. F.

Hesselbein

Kuhl

J. S.

(2015). Peter Drucker’s five most important questions: Enduring wisdom for today’s leaders. John Wiley & Sons.

24.

Ebarb

T. S.

(2019, September 7). Nonprofits fail—here’s seven reasons why. National Association of Nonprofit Organizations and Executives. https://nanoe.org/nonprofits-fail

25.

Ebrahim

Rangan

V. K.

(2014). What Impact? A Framework for Measuring the Scale and Scope of Social Performance. California Management Review, 56(3), 118–141. https://doi.org/10.1525/cmr.2014.56.3.118

26.

Epstein

M. J.

McFarlan

F. W.

(2011). Joining a nonprofit board. Jossey-Bass.

27.

Fama

E. F.

French

K. R.

(1997). Industry costs of equity. Journal of Financial Economics, 43(2), 153–193.

28.

Firth

Wong

S. M. L.

Yang

(2014). The double-edged sword of CEO/chairperson duality in corporatized state-owned firms: Evidence from top management turnover in China. Journal of Management & Governance, 18(1), 207–244. https://doi.org/10.1007/s10997-012-9225-6

29.

Gazley

Nicholson-Crotty

(2018). What Drives Good Governance? A Structural Equation Model of Nonprofit BoardPerformance. Nonprofit and Voluntary Sector Quarterly, 47(2), 262–285. https://doi.org/10.1177/0899764017746019

30.

George

Haas

M. R.

McGahan

A. M.

Schillebeeckx

S. J.

Tracey

(2023). Purpose in the for-profit firm: A review and framework for management research. Journal of Management, 49(6), 1841–1869.

31.

Graham

J. R.

Kim

Leary

(2020). CEO-board dynamics. Journal of Financial Economics, 137, 612–636.

32.

Guo

(2007). When government becomes the principal philanthropist: The effects of public funding on patterns of nonprofit governance. Public Administration Review, 67(3), 458–473.

33.

Gupta

V. K.

Mortal

S. C.

Silveri

Sun

Turban

D. B.

(2020). You’re fired! Gender disparities in CEO dismissal. Journal of Management, 46(4), 560–582. https://doi.org/10.1177/0149206318810415

34.

Hudson

Rehr

(2017). Association and nonprofit boards: Maximizing effective service. Heidrick & Struggles.

35.

Hager

M. A.

Yoon

(2024). Manager control over outcomes? Nature and nurture over time. Nonprofit and Voluntary Sector Quarterly, 53(5), 1261–1281.

36.

Hall

Millo

Barman

(2015). Who and what really counts? Stakeholder prioritization and accounting for social value. Journal of Management Studies, 52(7), 907–934. https://doi.org/10.1111/joms.12146

37.

Hillman

A. J.

Dalziel

(2003). Boards of directors and firm performance: Integrating agency and resource dependence perspectives. Academy of Management Review, 28(3), 383–396.

38.

Hung

Hager

M. A.

(2019). The impact of revenue diversification on nonprofit financial health: A meta-analysis. Nonprofit and Voluntary Sector Quarterly, 48(1), 5–27.

39.

Kaplan

R. S.

Grossman

A. S.

(2010, October). The emerging capital market for nonprofits. Harvard Business Review, 88(10). https://hbr.org/2010/10/the-emerging-capital-market-for-nonprofits

40.

Klein

(2020, February 19). Board members: How hard should they work? Nonprofit Quarterly. https://nonprofitquarterly.org/board-members-how-hard-should-they-work

41.

Korn

(2024, January 6). Wanted: New college presidents. Mission: Impossible. The Wall Street Journal. https://www.wsj.com/us-news/education/wanted-new-college-presidents-mission-impossible-0fdc54bd

42.

Kugel

Mercado

J. M.

(2024). Good governance in not-for-profit organizations: A review of the literature on boards of directors. Journal of Governmental & Nonprofit Accounting, 13(1), 56–82. https://doi.org/10.2308/JOGNA-2022-011

43.

Lecy

(2020). People Parser [Computer software]. https://github.com/Nonprofit-Open-Data-Collective/peopleparser

44.

Lecy

(2023). National Center for Charitable Statistics (NCCS), IRS 990 Efile Data. https://www.urban.org/tags/national-center-charitable-statistics-data

45.

Lee

Y. J.

(2019). Scarce as hen’s teeth: Women CEOs in large nonprofit organizations. Nonprofit Management and Leadership, 29(4), 601–610.

46.

Letts

C. W.

Ryan

W. P.

Grossman

(1999). High Performance Nonprofit Organizations. Wiley and Sons.

47.

Liu

(2017). Note on nonprofit board performance. Darden Business Publishing, University of Virginia.

48.

Liu

(2014). Outside options and CEO turnover: The network effect. Journal of Corporate Finance, 28, 201–217.

49.

Lyons

North-Samardzic

Young

(2007). Capital access of nonprofit organisations. Agenda: A Journal of Policy Analysis and Reform, 14(2), 99–110.

50.

MacQuillin

Sargeant

(2019). Fundraising ethics: A rights-balancing approach. Journal of Business Ethics, 160(1), 239–250. https://doi.org/10.1007/s10551-018-3872-8

51.

McFarlan

F. W.

(2017). Issues in non-profit governance. Harvard Business Publishing.

52.

McKee

Froelich

(2016). Executive succession planning: Barriers and substitutes in nonprofit organizations. Annals of Public and Cooperative Economics, 87(4), 587–601.

53.

McMullin

Raggo

(2020) Leadership and governance in times of crisis: A balancing act for nonprofit boards. Nonprofit and Voluntary Sector Quarterly, 49(6), 1182–1190. https://doi.org/10.1177/0899764020964582

54.

Meier

D. S.

Schnurbein

G. von

. (2024). From mission to market: Assessing sector overlap between nonprofits and for-profits. Nonprofit and Voluntary Sector Quarterly, 08997640241300509. https://doi.org/10.1177/08997640241300509

55.

Minichilli

Zattoni

Zona

(2009). Making boards effective: An empirical examination of board task performance. British Journal of Management, 20, 55–74.

56.

Murray

(2022, August 8). Many companies aren’t prepared to replace underperforming CEOs. Insights by Stanford Business. https://www.gsb.stanford.edu/insights/many-companies-arent-prepared-replace-underperforming-ceos

57.

National Center for Charitable Statistics. (2024, September 3). Nonprofit trends and impacts. https://urbaninstitute.github.io/nccs/resources/nonprofit-trends-and-impacts

58.

Never

(2013). Divergent patterns of nonprofit financial distress. Nonprofit Policy Forum, De Gruyter, 5(1), 67–84.

59.

Nili

Kastiel

(2017). Information capture: “Captured boards”: The rise of “super directors” and the case for a board suite. University of Wisconsin Law Review, 1392, 19–68.

60.

Norris-Tirrell

Rinella

Pham

(2018). Examining the career trajectories of nonprofit executive leaders. Nonprofit and Voluntary Sector Quarterly, 47(1), 146–164. https://doi.org/10.1177/0899764017722023

61.

O’Leary

A. S.

(2021, November 23). The data show what we know: The nonprofit helpers need help. National Council of Nonprofits. https://www.councilofnonprofits.org/articles/data-show-what-we-know-nonprofit-helpers-need-help

62.

Page

(2009). Unconscious bias and the limits of director independence. University of Illinois Law Review, 2009(1), 237–294.

63.

Philanthropy Roundtable. (2024). Should You Compensate Your Board Members? https://www.philanthropyroundtable.org/resource/should-you-compensate-your-board-members

64.

Pope

Bromley

Lim

Meyer

J. W.

(2018). The pyramid of nonprofit responsibility: The institutionalization of organizational responsibility across sectors. Voluntas, 29, 1300–1314. https://doi.org/10.1007/s11266-018-0038-3

65.

Prentice

C. R.

(2016). Why so many measures of nonprofit financial performance? Analyzing and improving the use of financial measures in nonprofit research. Nonprofit and Voluntary Sector Quarterly, 45(4), 715–740.

66.

Rawhouser

Cummings

Newbert

S. L.

(2019) Social impact measurement: Current approaches and future directions for social entrepreneurship research. Entrepreneurship Theory and Practice, 43(1), 82–115. https://doi.org/10.1177/1042258717727718

67.

Shen

Gentry

R. J.

Tosi

H. L.

(2010). The impact of pay on CEO turnover: A test of two perspectives. Journal of Business Research, 63(7), 729–734.

68.

Sisodia

Sheth

Wolfe

(2014). Firms of endearment: How world-class companies profit from passion and purpose. Pearson Education.

69.

Stephan

Patterson

Kelly

Mair

(2016) Organizations driving positive social change: A review and an integrative framework of change processes. Journal of Management, 42(5), 1250–1281. https://doi.org/10.1177/0149206316633268.

70.

Stewart

(2017). Exploring board perspectives on non-profit executive turnover. Voluntary Sector Review, 8(2), 169–185. https://doi.org/10.1332/204080517X14942368265346

71.

Stewart

A. J.

Adams

T. H.

McMillian

Burns

(2021). No room for failure: Investigating board leadership in nonprofit executive transitions. Nonprofit Management & Leadership, 31(3), 525–545. https://doi.org/10.1002/nml.21436

72.

Stewart

A. J.

Diebold

(2017). Turnover at the top: Investigating performance-turnover sensitivity among nonprofit organizations. Public Performance & Management Review, 40(4), 741–764. https://doi.org/10.1080/15309576.2017.1340900

73.

Stilwell

M. L.

(2021) Do private universities’ trustees hold their presidents accountable? A study of university performance and presidential turnover [Unpublished doctoral dissertation]. Creighton University.

74.

Tillotson

A. R.

Tropman

(2014). Early responders, late responders, and non-responders: The principal-agent problem in board oversight of nonprofit CEOs. Human Services Organizations: Management, Leadership & Governance, 38(4), 374–393. https://doi.org/10.1080/23303131.2014.916244

75.

Trautman

Ford

(2019). Nonprofit Governance: The Basics. Akron Law Review, 52(4). https://ideaexchange.uakron.edu/akronlawreview/vol52/iss4/2

76.

Unda

L. A.

Ranasinghe

(2021). To pay or not pay: Board remuneration and insolvency risk in credit unions. Pacific-Basin Finance Journal, 66, 101128. https://doi.org/10.1016/j.pacfin.2019.03.005

77.

van der Net

J. B.

Janssens

A. C. J.

Eijkemans

M. J.

Kastelein

J. J.

Sijbrands

E. J.

Steyerberg

E. W

. (2008). Cox proportional hazards models have more statistical power than logistic regression models in cross-sectional genetic association studies. European Journal of Human Genetics, 16(9), 1111–1116.

78.

van Puyvelde

Brown

W. A.

Walker

Tenuta

. (2018). Board effectiveness in nonprofit organizations: Do interactions in the boardroom matter? Nonprofit and Voluntary Sector Quarterly, 47(6), 1296–1310.

79.

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(3), 431–451.

80.

Ward

K. D.

Miller-Stevens

(2021). Public service motivation among nonprofit board members and the influence of primary sector of employment. Nonprofit and Voluntary Sector Quarterly, 50(2), 312–334. https://doi.org/10.1177/0899764020952161

81.

Woodroof

P. J.

Howie

K. M.

Peasley

M. C.

(2021). Nonprofit quality: What is it and why should nonprofits care? Journal of Philanthropy and Marketing, 26, e1682. https://doi.org/10.1002/nvsm.1682

82.

Wright

J. E.

Millesen

J. L.

(2008). Nonprofit board member role ambiguity: Investigating its prevalence, antecedents and consequences. American Review of Public Administration, 38(3), 322–338.

83.

Young

D. R.

(2013). If Not for Profit, for What? A Behavioral Theory of the Nonprofit Sector Based on Entrepreneurship (2nd ed.). Lexington Books.