Nonprofit Hospital CEO Compensation

Abstract

Background:

Past research has documented that increases in profits and health system size, as well as increases in the reward generosity for improving these metrics play an important role in explaining increases in nonprofit hospital CEO pay between 2012 and 2019.

Objectives:

To test whether hospital quality measures play a supplemental role in determining CEO pay.

Research Design:

We estimated linear regressions for 2012 and 2019 of the log of CEO wages on system or independent hospital characteristics, including quality. The regressions were used to construct a Oaxaca decomposition of factors associated with CEO compensation.

Subjects:

One thousand forty-seven nonprofit health systems and independent hospitals in 2012 and 812 in 2019.

Measures:

CEO compensation, hospital profits, charity care, hospital size, teaching status, system status, 30-day mortality rate for pneumonia patients, hospital-wide 30-day readmission rate.

Results:

We find that better quality was more closely associated with higher pay among hospital CEOs in 2012 versus 2019. The inclusion of these quality measures in the analysis somewhat reduced the observed relative return for leading larger hospitals or health systems in 2012, but not in 2019. The link between quality and CEO pay is weaker in 2019 than in 2012.

Conclusions:

The results suggest that nonprofit hospital CEOs are being rewarded more for leading large hospitals or systems, but not for providing higher quality care.

Hospital CEO salaries have rapidly increased in the past 2 decades. On the basis of data from 22 major nonprofit medical centers, mean CEO compensation rose 92% from $1.6 million (M) to $3.1M between 2005 and 2015.¹ A recent study estimated that mean salaries for a larger, more diverse sample of CEOs of nonprofit hospitals and health care systems grew 30% (inflation-adjusted), from $1M in 2012 to $1.3M in 2019.² CEO compensation packages for hospitals significantly exceed that of nonprofit leaders in other sectors, where mean CEO pay hovered between $120,000 and $180,000 in 2019.³ Furthermore, nonprofit hospital CEO compensation is growing faster than wages for other health care workers. Mean wages for all health care workers grew by only 5.4 percentage points between 2012 and 2019.^4,5

Nonprofit hospitals are generally controlled by a governing board, which incentivize CEOs by choosing the relevant factors along with their corresponding weights to determine compensation packages. Surveys of these CEOs indicate that boards commonly use care quality, financial performance, and (to a lesser extent) community benefit to evaluate the organization’s performance.⁶ While CEOs reported that boards use quality to measure organizational performance, it is unclear if care quality is associated with CEO pay. A cross-sectional study of 2009 nonprofit hospital CEO compensation found that process of care quality measures (eg, appropriate antibiotic selection) were not associated with higher CEO pay, although CEOs were rewarded for higher patient satisfaction scores.⁷ Other literature provides mixed results on the association of CEO compensation with financial performance, but a strong association with organization size.^8,9 Intentional or unintentional emphasis on financial performance and organization size in determining CEO pay may be fueling the observed increase in health system consolidation, which is contributing to rising health care prices. It is therefore important to empirically measure the relative contributions of financial performance, size, and quality of care in determining leadership pay in nonprofit hospitals.

A recent analysis found a strong association of nonprofit CEO salaries with both profits and organization size in cross-sectional analyses of data from 2012 to 2019.² More than one quarter of the increase in CEO pay between these years was attributable to an inflation-adjusted rise in profits and organizational size. Another substantial portion of the pay increase was attributable to a rise in compensation generosity for leading a hospital system with 500 or more beds. However, the study did not include hospital quality measures.

This study updates the literature by adding quality of care measures to the recent analysis of nonprofit hospital CEO compensation in 2012 and 2019, while quantifying the hospital characteristics most closely associated with the leadership pay increase between those years.

METHODS

Data

We obtained hospital financial data and facility characteristics (ie, size) from the National Academy for State Health Policy’s Hospital Cost Tool (HCT). Hospitals’ 2012 and 2019 IRS base 990 and Schedule H tax filings compiled by Candid allowed us to identify CEOs and their total compensation.¹⁰ We used the American Hospital Association (AHA) annual survey to identify teaching hospitals and hospitals affiliated with health systems.¹¹

We identified 2 measures of hospital quality in the Center for Medicare and Medicaid Services (CMS) hospital database archives that were reported for a large number of hospitals in both 2012 and 2019: 30-day mortality rate for pneumonia patients and hospital-wide 30-day readmission rate. We used the 2013 sample containing mortality and readmissions data from July 2009 to June 2012, and the 2020 dataset containing mortality data from July 2016 to June 2019 and readmission data from July 2018 to June 2019. Exhibit 1, Supplemental Digital Content 1, http://links.lww.com/MLR/D33 provides detailed data merge processes.

HCT and CMS data are publicly available. Candid and AHA data may be purchased from their respective organizations.^10–13

CEO Identification

We identify a CEO for each hospital or system using position titles and compensation reported in the 990s. This task is challenging, because large systems consisting of many hospitals may have multiple executives with titles such as, “CEO,” “President,” or “Executive Director,” while some hospitals do not explicitly list a CEO. Hospitals must report compensation information for all current officers, other key personnel, and highly compensated employees in their 990s.¹⁴

To identify the CEO, we begin with Song et al¹⁵ method to identify executives using their titles. For cases where Song criteria did not select any executive, the variety of titles used for selection was expanded. We designate the “CEO” as the highest paid executive whose title was flagged using this methodology.

We excluded clinical staff (eg, medical staff presidents) from CEO by removing observations with titles reflecting involvement in direct patient care, as these physician specialists could be highly compensated due to their performance of complex procedures, rather than overall performance of the hospital or system (see Exhibit 2, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, clinician exclusion criteria). We manually exclude academic personnel (eg, university presidents and coaches) reported on university hospitals’ 990s.

The IRS requires compensation from “related organizations” to be reported on 990 forms, which sometimes leads to system-level executives being reported on multiple hospitals’ 990s within the system. In these cases, we keep the observation with the highest reported total compensation, because our analysis is based on total compensation from all related organizations (see Exhibit 3, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, CEO selection including treatment of personnel with multiple observations). Total compensation consists of W-2 and/or 1099-MISC compensation, including base, bonus, and incentive pay; other reportable compensation; retirement and other deferred compensation; and nontaxable benefits.

Explanatory Variables

We hypothesized that profit, charity care, and the size of the hospital or system would be associated with CEO pay. To measure system size, we constructed categorical variables for the number of beds in each independent hospital or hospital system (<100, 100–299, 300–499, or 500 or more). We calculated the percentage of an independent hospital’s or system’s beds located in a teaching hospital and categorized them as: nonteaching, <50% of beds within teaching hospitals, and 50% or more beds within teaching hospitals.

Hospital quality measures (30 d pneumonia mortality and hospital-wide readmission rates) were aggregated to the system level using the number of inpatient discharges as weights.

Empirical Strategy

We estimated linear regressions for 2012 and 2019 of the log of CEO wages on system or independent hospital profits, charity care, size, teaching status, system status, and quality of care. The coefficients are interpreted as the percentage change in CEO wages associated with a one-unit change in each independent variable.

For our main specification, we measured size using the number of beds in each independent hospital or hospital system, and as robustness checks, we used the number of adjusted discharges and hospitals in the system.

Hospital systems and independent hospitals may be determining compensation based on different criteria given the differences in organizational complexity. However, there are not enough systems in our sample to precisely estimate the determinants of CEO pay for only hospital systems. Instead, we included an indicator for observations reflecting a hospital system, and we performed a robustness check using only independent hospitals.

We reviewed the proxy statements of the 5 largest for-profit hospital systems to study how quality measures were used when setting CEO compensation. When quality of care was used in setting CEO pay, it generally involved the quality of the CEO’s system versus the industry. Only one proxy statement provided details explaining how quality was factored into compensation. We used their calculations as a basis to categorize the pneumonia mortality and readmission rates into 3 groups: hospitals and systems with average quality (ratings within the 45–55 percentile range of the sample median), and those with rates lower or higher than average in each year.¹² Furthermore, the distributions of pneumonia mortality rates and hospital-wide readmission rates vary across years. Categorizing these variables enables meaningful comparisons of CEO compensation incentives in 2019 relative to those in 2012 and aligns with the methods outlined in the proxy statements of for-profit hospitals suggesting that quality-based compensation changes may be determined by differences in quality relative to other hospitals.

We seek to understand the difference in CEO pay between 2012 and 2019 that is explained by growth over time in characteristics such as hospital or system size, versus increases in the generosity of compensation in 2019 versus 2012 associated with a unit increase any of these attributes. The results from our regressions analysis were used to construct a Oaxaca decomposition.¹⁶

The Oaxaca decomposition has long been used by labor economists to break down differences in mean wages between 2 groups (eg, males vs. females) into the portion that is attributable to differences in pay for people with the same qualifications in each group (eg, college graduate), versus differences in pay resulting from differences in the qualifications of people in these 2 groups (eg college versus high school graduate). We apply the decomposition to decompose differences in CEO compensation that occurred due to changes in the means of covariates in our analysis and changes in the coefficients. We begin by estimating regressions of CEO pay on hospital characteristics for the years 2012 and 2019. The overall change in the mean log CEO compensation between 2012 and 2019 can be expressed as the following where X is a vector of the means of all covariates and α ₁₂ and β ₁₉ are vectors of the coefficients in each year.

$${\overline{\mathit{CEO}}}_{\mathrm{19}}-{\overline{\mathit{CEO}}}_{\mathrm{12}}=({{\beta }_0-a}_0)+{\beta }_{19}{\stackrel{\circledR }{X}}_{19}+{{\beta }_{19}\stackrel{\circledR }{X}}_{12}-{\beta }_{19}{\stackrel{\circledR }{X}}_{12}-a_{12}{\stackrel{\circledR }{X}}_{12}$$

These coefficients and mean explanatory variables can be rearranged to yield the following.

These 3 components are mutually exclusive and exhaustive, comprising the entire difference of mean log CEO compensation.

RESULTS

Our final sample consists of 1047 independent hospitals and health systems in 2012 and 812 in 2019. The largest system contained 101 hospitals in 2019. Between 2012 and 2019, the mean compensation of CEOs grew by 34%, from roughly $1M (in 2019 dollars) to $1.3M. Over the same period, registered nurses’ mean wages grew by only 2.3%, from $75,652 (in 2019 dollars) to $77,460.^17,18 Figure 1 graphs the levels of CEO compensation by year. As CEOs of independent and smaller hospital systems in 2012 consolidated into larger health systems by 2019, the distribution of CEO compensation shifted to the right.

FIGURE 1.

Distribution of CEO Compensation Levels in 2012 and 2019.

Oaxaca Decomposition

Table 1 presents the Oaxaca decomposition, which includes regression results from analyses estimating the association between CEO compensation and hospital or system characteristics in 2012 and 2019, along with the sample means in each year. (see Exhibits 4–6, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, summary statistics and visual representations of Table 1). The sample size, including independent hospitals and health care systems, dropped between 2012 and 2019 due to consolidation. The total difference in log CEO compensation is 0.193 from 2012 to 2019, and 0.147 of this difference was attributable to the difference in the intercept terms. The difference in the intercept term represents the difference in the base case across years, which is the difference in log pay for a hospital with fewer than 100 beds that recorded 0 profits, 0 charity care, and had worse than average quality ratings in 2019 versus 2012.

TABLE 1.

Oaxaca Decomposition

	Mean		Coefficient
	2012	2019	2012	2019	Change in log wage attributable to change in covariates	Change in log wage attributable to change in coefficients
Profit ($ Mill)	100.201	159.706	0.000433**	0.000425***	0.0253**	−0.001
Charity care ($ Mill)	15.896	18.336	0.001	0.0001	0.0001	−0.018
100–299 beds	0.290	0.244	0.850***	0.936***	−0.0435*	0.025
300–499 beds	0.118	0.113	1.106***	1.156***	−0.006	0.006
500+ beds	0.194	0.235	1.343***	1.571***	0.0649*	0.044
Teaching >0 and <50%	0.034	0.067	0.249*	0.310**	0.00997*	0.002
Teaching >50%	0.093	0.081	0.347***	0.216*	−0.002	−0.012
System	0.284	0.361	0.152*	0.112	0.009	−0.011
Mortality average	0.114	0.105	0.106	−0.019	0.0002	−0.014
Mortality lower than average	0.447	0.446	0.119**	0.079	−0.0001	−0.018
Readmission average	0.123	0.167	0.016	0.011	0.001	−0.001
Readmission lower than average	0.442	0.411	0.116**	0.0930*	−0.003	−0.010
Intercept term			12.44	12.59		0.147**
Total					0.0547***	0.1386***
No. observations	1047	812	1047	812

^* P<0.05.

^** P<0.01.

^*** P<0.001.

Table 1 presents the results of the analysis in which we regress hospital characteristics on log CEO compensation. The first 2 columns present the means of each variable in 2012 and 2019, while the third and fourth columns present the regression coefficients. Change in log wage attributable to change in covariates can be expressed as (mean 2019—mean 2012)*(Coefficient in 2019). Change in log wage attributable to change in coefficients can be expressed as (coefficient 2019–coefficient 2012)*Mean 2012.

Wage Increase Attributable to Covariate Changes

Mean profits rose 59.4% between 2012 and 2019 (from $100.2M to $159.7M), contributing to a 0.0253 increase in log compensation (P=0.002). Over the same period, mean charity care increased only 15.4% (from $15.9M to $18.3M), and this increase had a negligible contribution to the change in log wages. Mean bed size rose 30.7%, from 388.2 to 507.6. The median hospital was an independent hospital in both years, although the mean number of hospitals in an observation rose from 2.2 to 3.0. In 2012, 19.4% of hospitals were in the 500 or more beds category and in 2019, this share rose to 23.5%. The larger share of CEOs managing systems with 500 or more beds in 2019 versus 2012 was associated with 0.0649 greater log compensation (P=0.033). In 2012, 29% of hospitals had 100–299 beds. By 2019, the share of hospitals with 100–299 beds fell to 24.4%. Because hospitals consolidated between 2012 and 2019, much of this differential was offset by a 0.0435 reduction in log wages associated with fewer CEOs heading systems with 100 to 299 beds (P=0.025).

Hospital-wide readmission rates fell slightly between 2012 and 2019, while the pneumonia mortality rate increased. However, these changes did not contribute meaningfully to the change in log compensation. In total, differences in the means of the covariates contributed to 28.3% (0.0547) of the total difference in log wages (P<0.001).

Wage Increase Attributable to Coefficient Changes

A $1M increase in hospital profits was associated with a 0.04% (P=0.002 in 2012; P=0.000 in 2019) increase in CEO pay in both years. Charity care was not significantly associated with CEO compensation in 2012 or 2019. Increases in the reward for charity care spending or profits did not significantly contribute to the rise in CEO compensation from 2012 to 2019.

Hospital size had a large, statistically significant association in both years with CEO compensation. In 2012, hospitals with 100–299 beds paid CEOs 85% more (P=0.000) than the reference group of hospitals with fewer than 100 beds. Hospitals and systems with 300–499 beds paid CEOs 111% (P=0.000) more than the reference group, and hospitals with 500 or more beds paid their CEOs 134% (P=0.000) more. By 2019, CEOs were paid 157% (P=0.000) more than the reference group for leading a hospital with 500 or more beds, 116% (P=0.000) more for hospitals with 300–499 beds, and 94% (P=0.000) more for hospitals with 100–299 beds. The increase in the reward for leading a hospital with 500 or more beds and 100–299 beds accounted for increases in log wages of 0.044 and 0.025, respectively, from 2012 to 2019. In total, 0.075 of the 0.193 difference in CEO compensation was attributable to the rise in generosity of pay for heading a system with more beds.

CEOs of nonprofit hospitals or systems with more than 50% of beds at teaching hospitals were paid 34.7% more than CEOs of nonteaching hospitals or systems in 2012 and 21.6% more in 2019. Hospital systems with >0% but <50% of beds at teaching hospitals paid their CEOs 24.9% more than nonteaching hospitals in 2012 and 31% more in 2019. The coefficients for each of the teaching categorical variables were precisely estimated in both years, but they did not contribute significantly to the change in log CEO compensation from 2012 to 2019.

In 2012, CEOs of organizations with lower pneumonia mortality rates received higher salaries. For organizations with mortality rates that were lower than average in 2012, CEO pay was 11.9% (P=0.003) greater relative to independent hospitals or systems with mortality rates that were higher than the average. In 2019, CEOs of hospitals with lower-than-average mortality rates were paid 7.9% more relative to hospitals with higher-than-average rates. Lower-than-average readmission rates were also associated with 11.6% higher pay (P=0.003) in 2012 and 9.3% higher pay (P=0.042) in 2019 relative to CEOs in the higher-than-average readmission group.

CEOs of systems were paid 15.2% (P=0.011) more than CEOs of independent hospitals in 2012. By 2019, they were paid 11.2% more than independent hospitals. While individually, the coefficients in 2012 are not significantly different from 2019, a Wald test indicates that the 2019 coefficients are jointly significantly different from the 2012 coefficients (P=0.0012) (see Exhibit 7, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, Wald test).

Sensitivity Analyses

A robustness check analyzing only independent hospitals produced similar results to our primary regression (see Exhibit 8, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, analysis of independent hospitals). In another robustness check, we estimate our regressions using a generalized linear model with a log link function resulting in similar findings (see Exhibit 9, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, analysis using GLM).

Additional robustness checks using discharges and the number of hospitals instead of bed size to measure hospital size had similar results (see Exhibit 10, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, analysis using alternative measurements). Readmissions were omitted from the regression using the number of hospitals, because the correlation between readmission rates and the number of hospitals is relatively stronger than their correlation with CEO compensation, resulting in multicollinearity issues (see Exhibit 11, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, correlation with CEO compensation). A cross-tabulation of beds and hospitals is available in the supplemental digital content (see Exhibit 12, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, cross-tabulation). Sensitivity analyses using the cube root transformation of profit to account for rightward skew produced similar results. These alternative specifications support our finding that hospital size is the predominant factor in determining CEO compensation level within our analysis. In addition to confirming the relatively small association of profits and quality with CEO compensation, these specifications suggest that there may have been a minimal effect of charity care on CEO compensation in 2012.

DISCUSSION

Summary of Key Findings

This analysis of 1047 nonprofit health systems and independent hospitals in 2012 and 812 in 2019 found that most of the increase in log CEO compensation between 2019 and 2012 was attributable to an increase in pay for the base case hospital. CEOs who exceeded this base case by increasing bed size raised their pay even more, as the pay generosity for leading larger systems rose over time. Furthermore, the reward for leading an organization with lower than average readmission rates and mortality rates for pneumonia fell over time.

The substantial increase in base case salaries for nonprofit hospital CEOs likely contributes to growing disparities between health care executive salaries and nurses’ pay.¹ The additional rewards associated with leading larger, more profitable health care systems likely incentivize CEOs to pursue these objectives, which may explain the considerable uptick in health care system consolidation in recent years.

Limitations

CEOs of hospital systems may be managing many other types of facilities (eg, stand-alone clinics, ambulatory surgery centers, and imaging centers) that we lack information on but would likely boost their pay. The presence of these other facilities may be captured by our measure of hospital size. However, if all CEOs are responsible for managing more nonhospital facilities in 2019 versus 2012, this effect may be captured in the intercept term of our regressions.

Identifying the CEO was often challenging, because not all hospitals explicitly identify a CEO in their 990s. Many 990s had missing titles or used other titles for the CEO (eg, President/COO, Executive Officer) (see Exhibit 3, Supplemental Digital Content 1, http://links.lww.com/MLR/D33, CEO identification process). In some cases, these challenges could have led us to selecting a higher paid executive who was not the current CEO.

We may have underestimated some health care systems’ profits. When we aggregated the HCT hospital-level data to system-level, some hospitals were lost from the analysis, because we could not find a matching facility in the Candid dataset. However, 91.9% of the nonprofit hospitals in the HCT data were matched to the final dataset in 2012, and in 2019, 93.1% were matched, so this shortfall in profits is likely small.

Many mergers and acquisitions took place between 2012 and 2019, and when a hospital was acquired between 2012 and 2019, they are no longer in our sample as an individual hospital. They are instead aggregated within the acquiring system. However, because mergers and acquisitions increase number of beds, the resulting rise in compensation may be captured by the categorical bed size variables.

Another limitation was that we were not able to include other quality measures. All other available mortality measures in our analysis were highly correlated with pneumonia mortality and could not be included in the same regression. Previous research found that higher patient satisfaction was associated with greater CEO compensation.⁷ But inclusion of this measure would have reduced our cohort of hospitals and systems analyzed by 6% due to missing satisfaction data. This sample size reduction would have caused inconsistencies in the measurement of other variables. In addition, the hospital composite star rating was not available for 2012, so we were unable to include it in our analysis. However, the addition of 2 measures of organizational quality to previously conducted analyses plays an important role in explaining differences over time in CEO compensation.² Notably, CEOs in 2019 were not rewarded as generously for lower-than-average readmission or mortality rates for pneumonia patients relative to 2012.

Policy Implications

Hospital boards select the criteria that determine CEO pay. A survey of hospital boards and managers found that hospitals with boards that paid greater attention to clinical quality had management that better monitored quality performance.¹⁹ Other research has found that quality improvement interventions can be effective at reducing readmissions, but the net costs of these programs vary and are sometimes quite high.²⁰ Thus, while board members appear generally concerned with the provision of high-quality care, we are unaware of studies documenting the relative monetary returns to CEOs of devoting time and energy to improving profits or increasing organization size versus improving hospital or system quality. In addition, our own limited review of the proxy statements of 5 for-profit systems revealed that while quality was mentioned in the statements, it was often vague about what role, if any, quality played in directly determining compensation values. CEOs may find the rewards for increasing profits and building hospital or system size are easier to attain and focus their energies there.