Does the membership of an organization’s CEO on its board of directors affect the organization’s performance?  CEO board membership can have two opposing effects.  On one hand, it can reduce the independence of the board from management and thereby impede the ability of the board to monitor the CEO’s failure to pursue shareholder interests (Fama and Jensen 1983, http://dx.doi.org/10.1086/467037).  On the other hand, it can establish a “unity of command” in the firm (Finkelstein and D’Aveni 1994, https://doi.org/10.2307/256667) and facilitate the use of management’s private information to enhance the firm’s value.

The vast majority of research on this question has focused on the consequences of CEO board membership for for-profit firms.  The consequences of non-profit CEO board membership have received less attention.  Yet, the consequences of CEO board membership for non-profits may be even more important than the consequences for for-profits.  Fama and Jensen (1983) suggest that the agency problems between management and the intended beneficiaries of non-profits are so great that non-profits’ management should never serve on their boards.  Yet because of the breadth of non-profits’ goals, “unity of command” may be even more important in non-profits as well.

In this paper, we examine the consequences of non-profit CEO board membership (and board size) in a large sector primarily composed of non-profits:  the hospital industry.  We match data from five sources – the AHA Hospital Survey, the AHA National Health Care Governance Survey, IRS Form 990s, Medicare, and the RAND Transparency Initiative – to create a unique snapshot of the characteristics, governance, and performance of 1,744 nonprofit hospitals in 2018-22.  We describe two aspects of the governance of non-profit hospitals using the AHA Governance Survey and IRS Form 990s – the number of members of the board of directors and whether the CEO is a member of the board.  We examine the association of these aspects of hospital governance (holding constant other hospital characteristics and county fixed effects) with each hospital’s price to commercially-insured patients; operating margin; quality; service to low-income patients; and CEO pay.

We find a strong positive association between CEO board membership and non-profit hospital prices, operating margins, and CEO pay, with a weaker positive (negative) association between CEO board membership and quality (service to low-income patients).  Our results are consistent with Brickley, Van Horn, and Wedig (2010, https://doi.org/10.1016/j.jebo.2010.06.008) who show a positive association between non-profit hospital CEO pay premia and boards with “insider” members.  Our results are also consistent with several recent papers that examine the incentives of non-profit hospital CEOs, including the following:

1. Jenkins, Short and Ho (2024, https://doi.org/10.1371/journal.pone.0306571) find a positive association between non-profit hospital CEO pay and revenues in excess of costs, but not the volume of charity care;
2. Lewellen, Phillips, and Sertsios (2023, https://haslam.utk.edu/wp-content/uploads/2023/04/Lewellen-Nonprofit-Governance-LPS-March-24_2023.pdf) find that “CEO pay and turnover are sensitive to financial performance but are unresponsive (or less responsive) to nonfinancial goals, including the quality of medical treatment, patient satisfaction, and charity provision”; and
3. Mulligan, Choksy, et al. (2022,https://doi.org/10.1371/journal.pone.0264712) find non-profit hospital CEO pay has “a stronger relationship with financial performance than with non-profit performance measures.”

Our findings provide evidence for both of the competing effects of CEO board membership hypothesized in corporate governance theory.  On one hand, we find evidence that CEO board membership contributes to agency problems.  CEO board membership is positively associated with CEO pay.  Hospitals with CEOs on their boards have higher prices and operating margins – goals for which other studies show CEOs are rewarded – despite the fact that higher prices and margins may not be in society’s overall interest (e.g., Brot-Goldberg, Cooper, et al. 2024, https://www.nber.org/papers/w32613).  Conversely, hospitals with CEOs on their boards have less service to low-income patients – a goal for which other studies show CEOs are not rewarded – despite the fact that service to low-income patients may be in society’s overall interest.  On the other hand, we also find evidence that CEO board membership contributes to at least one socially beneficial measure of hospital performance – quality.  Our finding that CEO board membership is positively associated with quality is inconsistent with a pure agency explanation.

For these reasons, CEO board membership involves a tradeoff for non-profit hospitals.  On one hand, it is associated with higher CEO pay, higher prices and margins, and lower service to low-income patients, all of which signal an agency problem.  On the other hand, it is associated with higher quality, which may be socially valuable.  Based on the magnitudes of our estimates, we conclude that the social cost from agency problems associated with CEO board membership likely exceed the social benefits from higher quality, although our calculations should be viewed with caution given that the associations we measure do not necessarily represent causal effects.