Hospital Compensation: Nationally Representative Data on Chief
Executives' Compensation (Letter Report, 08/16/94, GAO/HEHS-94-189).
Hospital-reported data showed that chief executives received an average
of $129,000 in compensation for overseeing hospital operations in 1991.
Actual compensation ranged from $31,000 to $849,000. Differences in
compensation are influenced by the hospital's patient load, the number
and relative size of nearby hospitals, and the hospital's geographic
locations and ownership type. In general, executive compensation is
higher at hospitals with greater financial success, greater numbers of
patients discharged, higher numbers of similarly sized hospitals nearby,
location in large cities, or for-profit operations. Data on executive
compensation from related businesses at not-for-profit hospitals showed
that relatively few executives received such payments. These payments,
however, can be large. Compensation rose from six to 138 percent over
the amounts received for hospital administration for the executives who
received payments from such related businesses as hospitals and other
health care facilities, as well as foundations and property management
firms. Dollar amount increases ranged from $13,000 to $530,000.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: HEHS-94-189
TITLE: Hospital Compensation: Nationally Representative Data on
Chief Executives' Compensation
DATE: 08/16/94
SUBJECT: Hospitals
Executive compensation
Health care cost control
Income statistics
Hospital administration
Hospital bed count
Salary increases
Reporting requirements
Financial management
IDENTIFIER: Medicaid Program
**************************************************************************
* This file contains an ASCII representation of the text of a GAO *
* report. Delineations within the text indicating chapter titles, *
* headings, and bullets are preserved. Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are *
* identified by double and single lines. The numbers on the right end *
* of these lines indicate the position of each of the subsections in the *
* document outline. These numbers do NOT correspond with the page *
* numbers of the printed product. *
* *
* No attempt has been made to display graphic images, although figure *
* captions are reproduced. Tables are included, but may not resemble *
* those in the printed version. *
* *
* A printed copy of this report may be obtained from the GAO Document *
* Distribution Facility by calling (202) 512-6000, by faxing your *
* request to (301) 258-4066, or by writing to P.O. Box 6015, *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time. *
**************************************************************************
Cover
================================================================ COVER
Report to the Chairman, Subcommittee on Oversight and Investigations,
Committee on Energy and Commerce, House of Representatives
August 1994
HOSPITAL COMPENSATION - NATIONALLY
REPRESENTATIVE DATA ON CHIEF
EXECUTIVES' COMPENSATION
GAO/HEHS-94-189
Hospital Compensation
Abbreviations
=============================================================== ABBREV
AHA - American Hospital Association
ARF - area resource file
CPI - Consumer Price Index
HCFA - Health Care Financing Administration
HCRIS - Medicare Hospital Cost Report Information System
IRS - Internal Revenue Service
MSA - metropolitan statistical area
SEC - Securities and Exchange Commission
Letter
=============================================================== LETTER
B-256200
August 16, 1994
The Honorable John D. Dingell
Chairman, Subcommittee on Oversight
and Investigations
Committee on Energy and Commerce
House of Representatives
Dear Mr. Chairman:
This report responds to your request for information on the
compensation paid to the chief executives of our country's hospitals.
In December 1993,\1 we testified on executive compensation before
your Subcommittee, which has been looking into various aspects of the
financial operations of the health care industry. This report
presents additional information concerning compensation hospital
chief executives received from 1989 through 1991. More specifically,
it addresses (1) the compensation hospital chief executives received
for overseeing hospital operations, (2) the factors that influenced
these compensation levels, and (3) compensation paid to executives of
not-for-profit hospitals by businesses related\2 to those hospitals.
--------------------
\1 Hospitals: Chief Executives' Compensation, 1989-1991
(GAO/T-HRD-94-70, Dec. 7, 1993).
\2 Related businesses are those that share a common governing board
or set of officers. In other words, a related business is one the
hospital directly or indirectly owns or controls or, conversely, one
that directly or indirectly owns or controls the hospital.
BACKGROUND
------------------------------------------------------------ Letter :1
In recent years, the media have focused public attention on seemingly
high salaries paid to some health care and other chief executives.
In particular, examples of salaries approaching $1 million paid to
hospital chief executives have often been highlighted and sometimes
cited as contributors to rising health care costs. However, little
was known about how representative these salaries were of the
industry as a whole or about the various factors that influenced
chief executives' compensation.
Compensation for hospital executives is most often set by a
hospital's governing board, which is usually composed of community
volunteers. Compensation levels reflect the board's desire to
attract, retain, and motivate executives who will implement board
decisions regarding the hospital's organizational strategy and
policy, mission, and financial soundness. Appendix III contains more
information on the board's role in setting executive compensation.
SCOPE AND METHODOLOGY
------------------------------------------------------------ Letter :2
To obtain nationally representative data, we surveyed 429 hospitals
that participated in Medicare, a federally funded health care program
that accounted for 35 percent of hospital patient care revenue in
fiscal year 1991. Our survey, which included for-profit,
not-for-profit, and state and local government hospitals, yielded
results that could be projected to the country's 5,300
Medicare-participating hospitals. We received an 86-percent response
rate from our survey, which covered tax years 1989, 1990, and 1991.
We developed an econometric model to quantify the impact of various
factors on the level of compensation hospitals pay their chief
executives. Among the factors we included in our analysis were the
annual number of inpatient days and patients discharged, the
hospital's financial performance and ownership type, the number and
relative size of nearby hospitals, and geographic location.
Beginning with tax year 1992, the Internal Revenue Service (IRS)
required not-for-profit hospitals to report the source and amount of
payments from related businesses to executives and other key
personnel whose compensation exceeded $100,000 of which $10,000 or
more came from related businesses. To determine the compensation
paid to chief executives by related businesses, we asked 194
not-for-profit hospitals that responded to our survey for their
reports to IRS for 1992. Of these hospitals, 137 (71 percent)
responded; 112 supplied the reports and 25 stated theirs were not yet
due and so, not available. Appendixes I and II contain more detail
on our scope and methodology.
RESULTS IN BRIEF
------------------------------------------------------------ Letter :3
Hospital-reported data showed that chief executives received an
average of $129,000 in compensation, including cash (salary, fees,
and bonuses), benefits, and allowances, for overseeing hospital
operations during 1991. Executives in 1991 administered a hospital
that averaged about 180 beds, with net patient revenue of $42.3
million and net income of $1.8 million. Overall, one-fourth of chief
executives received less than $63,000, while an equal number received
over $176,000. Actual compensation ranged from $31,000 to $849,000.
Differences in compensation amounts are influenced by the hospital's
patient load, the number and relative size of nearby hospitals, and
the hospital's geographic location and ownership type. Except in the
smallest hospitals, compensation amounts are also partly determined
by the hospital's financial performance. In general, executive
compensation is higher at hospitals with greater financial success,
greater numbers of patients discharged, higher numbers of similarly
sized hospitals nearby, location in large cities, or for-profit
operations.
Data on executive compensation from related businesses at
not-for-profit hospitals showed that among the 112 hospitals we
examined, relatively few executives received such payments. These
payments, however, can be large. Compensation increased from 6 to
138 percent over the amounts received for hospital administration for
the executives who received payments from such related businesses as
hospitals and other health care facilities, as well as foundations
and property management firms. Dollar amount increases ranged from
$13,000 to $530,000.
COMPENSATION FOR OVERSEEING
HOSPITAL OPERATIONS
------------------------------------------------------------ Letter :4
In 1991, hospital executives received an average of $129,000 (plus or
minus $9,000) in compensation\3 for overseeing hospital operations.
Over 93 percent of reported compensation was in the form of cash
payments. The median compensation of $112,291 was somewhat lower
than the $129,000 average, indicating that a relatively small number
of executives received relatively high amounts.\4
As shown in figure 1, from 1989 to 1991 the average change in
executive compensation, not adjusted for inflation, increased at less
than the rate for general hospital operating costs but at a greater
rate than the Consumer Price Index (CPI) and the Medical Care
Index.\5
Figure 1: Percentage Change in
Costs and Compensation, 1989-91
(See figure in printed
edition.)
Note: Change in hospital operating costs is the average change for
the period 1989-91.
Sources: Change in chief executive compensation was calculated from
responses to our survey. Change in hospital operating costs was
calculated by the Prospective Payment Assessment Commission. Changes
in the CPI and Medical Care Index were calculated from Bureau of
Labor Statistics data.
When we adjusted for inflation, we found that compensation for
executives grew almost 9 percent from 1989 to 1991.\6 Compensation
for executives of urban hospitals grew slightly faster, or about 10
percent, adjusted for inflation, during the same period.
The range of actual annual compensation for hospital chief executives
responding to our survey was from about $31,000 at a 48-bed hospital
in Texas to about $849,000 at an 880-bed hospital in New York. Table
1 shows the compensation, adjusted for inflation, chief executives
received for the 3 years covered by our survey.
Table 1
Chief Executive Compensation, 1989-1991
(In 1991 dollars)
25 25
percen percen
t t
Sampli receiv receiv
Averag ng ed ed
e error\ more less
Year (mean) a Median than than
-------------------- ------ ------ ------ ------ ------
1989 $120,1 $8,860 $95,90 $153,4 $64,16
94 4 28 8
1990 126,11 12,136 101,11 156,30 62,520
2 5 0
1991 128,75 8,317 112,29 175,95 62,971
4 1 7
------------------------------------------------------------
\a Sampling errors are computed at the 95-percent confidence level.
Compensation received by chief executives in 1991 was about 1.4 times
greater than amounts received by the hospitals' other top management
and highly compensated employees. While the average compensation for
chief executives was $129,000, the average for top management was
$97,000, and the average for other highly compensated employees was
$86,000. Top management includes vice presidents, chief financial
officers, and chief operating officers. Other highly compensated
employees include medical directors, facilities and services
managers, nurses, and physicians working on the hospital staff.
Most of the compensation reported--about 93.6 percent--was in the
form of cash payments. Benefits constituted 4.8 percent of the
total, and allowances constituted 1.6 percent. The amounts reported
as benefits are understated because many hospitals did not report
them.
IRS officials have expressed concerns about instances of abuse in
reporting taxable income, which can be significant for involved
individuals. At congressional hearings in summer 1993,\7 IRS
officials said that audits of large not-for-profit educational and
health care systems and media evangelists revealed compensation,
benefits, and allowances provided to executives that were not
accurately reported as taxable compensation to the individuals.
Examples from hospitals included $20,000 in country club dues and
catered meals as well as substantial payments for personal expenses
such as liquor, china, perfume, crystal, and theater and airline
tickets for a chief executive; and deferred compensation
arrangements. However, IRS officials stated that, overall,
not-for-profits comply fully with reporting requirements.
We asked Medicare officials who audit annual cost reports filed by
participating hospitals whether concerns related to excessive or
unreasonable compensation had been raised in any review of the 429
hospitals in our sample, regardless of ownership. No such concerns
were reported in audits or other reviews of these hospitals conducted
since 1989.
In some instances, chief executives were not employed by the hospital
at all but instead were employees of a company that provided
management services by contract with the hospital. Between 6 and 15
percent of hospitals had such arrangements. In these cases the
hospitals did not report specific information on compensation paid to
the chief executive and so were not included in our analysis.
Appendix IV provides information on management services contracts.
--------------------
\3 Compensation includes three different components: cash, benefits,
and allowances. Cash refers to all cash payments received by the
individual for such items as salary, fees, and bonuses; severance
payments; payments for accumulated, but unused, leave; payments of
amounts previously deferred; and forgiven loan balances. Benefits
include payments made by the employer on behalf of the individual to
pension or health insurance plans. Allowances include, under certain
conditions, such items as the value of housing, automobiles, or other
assets provided by the hospital and the value of payments for life
insurance, travel, and tuition. Allowances are income if they must
be reported on one's personal tax return.
\4 Median compensation is the value that falls midway between the
highest and lowest amounts, meaning that half of executives received
more than $112,291 and half received less.
\5 CPI measures the average change in prices in a market basket of
goods and services, while the Medical Care Index, a component of CPI,
measures the average change in prices for medical care commodities
and services.
\6 Compared with the overall average compensation increase, the
median increase during the same period was slightly less for
medium-size and large hospitals (8.5 percent and 8.3 percent,
respectively) and much less for small hospitals (3.2 percent).
\7 Hearings on federal tax laws applicable to the activities of
tax-exempt charitable organizations were held by the Subcommittee on
Oversight of the Committee on Ways and Means, House of
Representatives, on June 15 and August 2, 1993.
COMPENSATION VARIED BY HOSPITAL
CHARACTERISTICS, PATIENT LOAD,
FINANCIAL PERFORMANCE, AND
LOCATION
------------------------------------------------------------ Letter :5
Our analysis confirms some of what could be considered conventional
wisdom. For example, chief executives employed by large or
for-profit hospitals are likely to receive the highest compensation.
On the other hand, chief executives employed by small or
government-owned hospitals are likely to receive the lowest
compensation. In addition, significant regional variation exists in
the compensation hospitals pay their chief executives.
However, we also found several relationships that are less well
known. For example, while executive compensation increases with the
number of patients discharged, chief executives at large hospitals do
not necessarily receive higher compensation than executives at
medium-size hospitals. Additionally, executive compensation is
affected by the hospital's financial performance in all but the
smallest hospitals, is higher in areas with greater competition
between hospitals, and is not significantly affected by the
hospital's involvement in medical education.
SIZE AND PATIENT LOAD
---------------------------------------------------------- Letter :5.1
For all sizes of hospitals the compensation paid to chief executives
rises with the number of patients discharged annually. For example,
among urban hospitals, a 10-percent increase in the number of
patients discharged annually is associated with a 2.2-percent
increase in compensation. If all else is equal, an executive of an
urban hospital with 12,200 patients discharged annually would be
expected to receive approximately $6,300 more than an executive of a
hospital with 10,200 patients discharged.
The higher compensation associated with greater numbers of patients
discharged in part reflects the concomitant increase in the
responsibility and complexity of the executives' jobs. These results
also suggest that hospitals reward executives for reducing the
average number of days patients stay in the hospital.\8 Because
Medicare and many private insurers pay hospitals a flat fee per
diagnosis, regardless of the hospitals' true cost for treating the
patient, hospitals can improve their financial performance by
reducing the length of patient stays.
The relationship between compensation and number of beds is less
consistent. Although executives of small hospitals earn less than
their counterparts at medium-size hospitals, we found no evidence
that compensation at large hospitals is higher compared with
compensation at medium-size hospitals, after controlling for other
influencing factors.
The number of beds is an important determinant of executive
compensation only for hospitals with fewer than 100 beds. Even after
considering the number of patients served, executives of small
hospitals earn 25 percent less than executives of hospitals with
between 100 and 500 beds. Appendix V provides information on
executive compensation at hospitals of varying size.
--------------------
\8 An increase in annual number of patients discharged accompanied by
no change in the total number of inpatient days implies a shorter
average length of stay.
FINANCIAL PERFORMANCE
---------------------------------------------------------- Letter :5.2
For all but the very smallest hospitals (fewer than 50 beds), a
hospital's financial performance--as measured by the proportion of
patient revenues realized as profits--is another significant
determinant of executive compensation.\9 On average, a chief
executive at a hospital rated strongly (75th percentile) on this
financial performance measure would receive approximately $4,500 (3.6
percent) more than an executive at a hospital rated less strongly
(the 25th percentile). The compensation differential is even
greater, or almost $6,400 (4.5 percent), between two urban hospitals
with those same relative financial performance ratings (75th versus
25th percentile).
--------------------
\9 This proportion, called the "operating margin," is the ratio of
net patient revenue less associated costs to net patient revenue.
Our methodology allowed for the possibility that executive
compensation could affect a hospital's operating margin, for example,
if the hospital paid a premium to attract a particularly skilled
executive who improved financial performance. See appendix II for
more details.
COMPETITION
---------------------------------------------------------- Letter :5.3
Executive compensation tends to be higher at hospitals in more
competitive hospital markets, as measured by the number and relative
size of area hospitals.\10 Higher compensation may result from a
greater demand for hospital executives in those areas or from a
demand for more talented and, thus, more highly paid executives to
meet the market challenge.
For example, our results suggest that average compensation would be
1.6 percent higher in a county with 4 hospitals, each with 180 beds,
than in a county with the same number of hospital beds divided among
3 hospitals. However, competitiveness, and thus executive
compensation, depends not only on the number of hospitals in the
county, but also on their relative size. Compensation is higher in
counties where the hospitals are nearly equal in size than in
counties with the same number of hospitals of disparate size. For
example, average compensation would be 4 percent higher for the
executives in a county where 3 hospitals each had 250 beds than for
executives in a county with 1 hospital of 550 beds and 2 others of
100 beds each.
--------------------
\10 The positive and significant coefficient on sole community
hospitals noted in table II.3 does not contradict this finding.
Local competitiveness is measured at the county level while "sole
community" status is defined by the Health Care Financing
Administration (HCFA) using its own criteria. Thus, hospitals may be
defined as monopolists by the Herfindahl index without being
designated as sole community hospitals by HCFA and vice versa.
OWNERSHIP AND INVOLVEMENT IN
MEDICAL EDUCATION
---------------------------------------------------------- Letter :5.4
When size and other characteristics are equal, compensation is
highest at for-profit hospitals, next highest at not-for-profit
hospitals, and lowest at government hospitals. This compensation
pattern may reflect factors such as scope of responsibility, job
security, and nonmonetary benefits that vary by ownership type.
Overall, for-profit hospitals pay their executives approximately 12
percent more (almost $12,500) than do not-for-profit hospitals. The
difference in compensation between for-profit and not-for-profit
hospitals is even greater (over $30,000) in urban areas. Government
hospitals typically pay 9 percent less, in both rural and urban
areas, than otherwise similar not-for-profit hospitals. Appendix VI
provides information on executive compensation at hospitals with
differing ownership. No significant differences exist between the
compensation paid by hospitals involved in medical education and the
compensation paid by hospitals that are not.
LOCATION
---------------------------------------------------------- Letter :5.5
We measured the regional variation in compensation between the nine
regions defined by the U.S. Bureau of the Census, which are shown in
figure 2 and listed in table II.3. Regional variation in executive
salaries is most evident for urban hospitals. Among otherwise
similar hospitals, executives at urban hospitals in the West North
Central States earn about 15 percent more ($22,700) than executives
at urban hospitals in the South Atlantic States. Executives of urban
hospitals in the Pacific, Mountain, East North Central, and West
South Central States earn 8 to 20 percent less (approximately $12,800
to $30,000) than do executives of urban hospitals in the South
Atlantic States.
Figure 2: Census Divisions and
Regions for the United States
(See figure in printed
edition.)
Source: U.S. Bureau of the Census.
When nonurban hospitals are included, regional differences are not as
evident. Only hospitals in the Pacific States pay significantly less
(24 percent, or approximately $26,000) to executives than do
hospitals in the South Atlantic States. Executive compensation in
all other states is approximately the same as in the South Atlantic
States.
Among urban hospitals, executive compensation increases with the size
of the city. Our analysis indicates that an executive of a hospital
in a city the size of Buffalo, New York (with an urban population of
approximately 1.19 million), would, on average, earn 5.7 percent more
than an executive at a similar hospital in a city the size of
Binghamton, New York (with an urban population of approximately
264,000). The disparity reflects the combined impact of differences
in the cost and quality of living, hospital market competitiveness,
and other factors between cities of unequal size.
RELATED BUSINESSES ADD TO SOME
HOSPITAL EXECUTIVES'
COMPENSATION
------------------------------------------------------------ Letter :6
Executives who oversee hospital operations may also oversee the
operations of businesses related to the hospital and be compensated
in part by them. Related businesses include other hospitals, parent
corporations, foundations, research institutes, medical equipment
companies, home health agencies, pharmacies, management and
consulting firms, diagnostic centers, and property management firms
specializing in building rentals and parking lots. At the hospitals
we examined, relatively few executives received such payments, though
in some cases these payments equaled 50 percent or more of the
compensation received for overseeing hospital operations.
We estimate that almost 60 percent (plus or minus 6 percent), or
about 3,200, of the country's Medicare- participating hospitals,
regardless of ownership type, had one or more related businesses in
1991. Among not-for-profit hospitals, half had at least one related
business. Hospitals with related businesses were, on average,
related to 2 other businesses, but the number ranged up to 24. Most
related businesses (66 percent) were not-for-profit; the remainder
were for-profit.
Of the 112 not-for-profit hospitals that supplied data, 4 of 74
medium-size and 7 of 34 large hospitals reported payments to their
executives from businesses related to the hospitals. The remaining
four small hospitals reported no such payments to their executives.
Although payments from related businesses may involve a small number
of executives, such payments can be large. For example, in the cases
we examined, payments from related businesses increased executive
compensation from 6 to 138 percent. Dollar increases for executives
ranged from $13,000 to $96,000 at medium-size hospitals and from
$14,000 to $531,000 at large hospitals.
The executives in the cases we examined received additional payments,
on average, from 2 related businesses, but the actual number ranged
up to 6. The businesses making these payments were generally
not-for-profit, rather than for-profit, enterprises. Payments to
executives came primarily from health care organizations but also
from foundations and property management firms.
The examples below demonstrate the range, in combined compensation,
of how payments from related businesses can affect chief executive
compensation.
A chief executive at a large not-for-profit hospital with 880 beds
had managerial and other duties for the hospital's fund-raising
body and two property management firms. Total compensation in
1992 for this executive was 79 percent above that received for
managing the hospital. The chief executive received $675,829
for administering the hospital and $530,553 for responsibilities
to three of the hospital's related businesses, bringing his
total to $1,206,382.
At a hospital with 513 beds, the chief executive received $127,244
in compensation for hospital-related administrative duties in
tax year 1992. This executive also had administrative
responsibilities for a related health care corporation from
which he received $14,162, bringing the total payment package to
$141,406, a 12- percent increase over his hospital compensation.
---------------------------------------------------------- Letter :6.1
This work was done under the direction of Sarah F. Jaggar, Director
of Health Financing and Policy Issues. Please call Cheryl A.
Williams, Evaluator-in-Charge at (503) 235-8451 or James C.
Cosgrove, Senior Economist at (202) 512-7029 if you have any
questions. Other major contributors are listed in appendix VIII.
As arranged with your office, unless you publicly announce its
contents earlier, we plan no distribution of this report until 30
days after its issue date. At that time we will send copies to the
Secretary of Health and Human Services; the American Hospital
Association, which took an active role in encouraging hospitals to
respond; executives of participating hospitals; and other interested
congressional committees. Copies will also be made available to
others upon request.
Sincerely yours,
Leslie G. Aronovitz
Associate Director, Health
Financing Issues
SCOPE AND METHODOLOGY
=========================================================== Appendix I
We conducted a nationwide survey of 429 Medicare-participating
for-profit, not-for-profit, and state and local government hospitals
of varying sizes.\11 Many hospitals file annual reports with the
Internal Revenue Service (IRS) or, in some cases, with the Securities
and Exchange Commission (SEC). These reports contain compensation
and other data related to the organizations' operations. We
requested copies of these publicly available reports or, for
hospitals not subject to IRS or SEC reporting, the same information
on a questionnaire we developed patterned after these agencies'
reporting requirements. (See app. VII for a copy of our survey
instrument.) We asked hospitals to send data for the tax years 1989,
1990, and 1991.
To ensure adequate representation from hospitals of varying sizes, we
selected a stratified random sample. We based our strata on bed
size: small hospitals had from 1 to 100 beds; medium-size hospitals,
101 to 500 beds; and large hospitals, over 500 beds. We received
complete responses from 368 hospitals, or 86 percent. Tables I.1 and
I.2 show the distribution of responding hospitals by size and
ownership type.
Table I.1
Survey Response Rate by Hospital Size
Percen
No t
Univer Sample respon Respon respon
Hospital size se size se ses se
-------------------- ------ ------ ------ ------ ------
Small 2,491 84 18 66 78.6
Medium 2,596 260 39 221 85.0
Large 241 85 4 81 95.3
Total 5,328 429 61 368 85.8
------------------------------------------------------------
Table I.2
Survey Response Rate by Hospital
Ownership Type
(Numbers are aggregated across size
strata)
Percen
Survey No t
s Respon respon respon
Hospital ownership type mailed ses se se
---------------------------- ------ ------ ------ ------
Not-for-profit 285 257 28 90.1
For-profit 54 37 17 68.5
Government 90 74 16 82.2
Total 429 368 61 85.8
------------------------------------------------------------
To obtain data on payments to executives from businesses related to
the hospital, we requested 1992 IRS reports from the 194
not-for-profit hospitals that sent us 1991 IRS reports and that
reported they had related businesses. The remaining 174 hospitals
either did not send us IRS reports, did not have related businesses,
or were for-profit.
Beginning with tax year 1992, the IRS required not-for-profit
hospitals to report the source and amount of payments from related
businesses to executives and other key personnel when compensation
for these individuals exceeded $100,000 of which $10,000 or more came
from related businesses. Of the 194 hospitals, 137 (71 percent)
responded to our request, though 25 of these stated their reports
were not yet due, and so, not available. As a result, we received
112 usable responses.\12
Table I.3 shows the distribution of responding hospitals by size.
Table I.3
Responses to Request for 1992 IRS
Reports by Hospital Size
Execut
ives
receiv
Reques No Report Usable ing
ts respon s not respon paymen
Hospital size mailed se due ses ts
-------------------- ------ ------ ------ ------ ------
Small 9 4 1 4 0
Medium 129 36 19 74 4
Large 56 17 5 34 7
------------------------------------------------------------
While IRS reporting requirements increase public access to
compensation information, data are not available in all cases. For
example, three hospitals did not report executive compensation of
$100,000 or more, part or all of which was paid by related
businesses. One hospital official said IRS criteria require
disclosure only when the executive received payments both from the
hospital and a related business, not when all payment came from the
related parent corporation. In another instance, a hospital official
stated that disclosure was not required when the payments were made
by two members of the same holding company. Lastly, one hospital
official stated that disclosure was not required of amounts paid
under contract with a related management services firm. We discussed
these situations with an IRS official who stated in these instances
hospitals should have reported the executives' compensation.
We also obtained data on hospital characteristics such as size and
financial performance supplied to the Department of Health and Human
Services on Medicare hospital cost reports and to the American
Hospital Association (AHA) as part of their annual survey. We did
not verify the accuracy of information supplied by hospitals to the
IRS, other federal agencies, or AHA.
We furnished a list of the hospitals included in our survey to
officials charged with monitoring hospitals for compliance with
Medicare regulations, including regional administrators of the Health
Care Financing Administration (HCFA), and Medicare contractors that
administer the program. We asked these officials whether audits or
other field reviews conducted since 1989 revealed concerns about
excessive compensation for executives. All HCFA officials and all
but 6 contractors (covering 18 hospitals, or 4 percent of our sample)
responded to our request for a review of their records related to the
429 hospitals.
To study the effect of various factors on compensation levels, we
used multiple regression analysis--a standard statistical technique
that quantifies the relationship between a dependent variable and a
set of independent variables. Among the factors we included in our
analysis were the annual number of inpatient days and patients
discharged, the hospital's financial performance and ownership type,
the number and relative size of nearby hospitals, the hospital's
involvement in medical education, membership in health systems or
alliances, geographic location, and whether the chief executive was
new to the position. Appendix II contains more detail on our
regression analysis.
Our work was conducted between December 1992 and July 1994 in
accordance with generally accepted government auditing standards.
--------------------
\11 We excluded federally operated hospitals.
\12 While our sample was randomly selected, we chose not to project
our results to the universe of hospitals because of a potential bias.
Specifically, hospitals whose executives receive such payments may be
less willing to provide copies of their IRS reports than hospitals
whose executives do not.
ECONOMIC ANALYSIS OF CHIEF
EXECUTIVE COMPENSATION
========================================================== Appendix II
We developed an econometric model to quantify the impact of various
factors on the level of compensation hospitals pay their chief
executives. In building this model, we assumed that the basic market
forces of supply and demand influence executive compensation.
However, there are many reasons to believe that a simple neoclassical
labor-market model is inadequate to capture the complex conditions
that characterize the market for chief executives. Some of these
features include the difficulty in measuring a chief executive's
marginal contribution, variations in the workload and responsibility
associated with the title, differences in market conditions that
affect firms' demand for executive human capital, and disparities in
firms' ability to pay. Therefore, our model was augmented with a
number of variables intended to control for these conditions.
We estimated our model using multiple regression--a standard
statistical technique that quantifies the relationship between a
dependent variable and a set of independent variables. The
construction of the model and the data are described below. Data
sources and descriptive statistics for the analysis variables are
summarized in tables II.1 and II.2. The econometric results are
presented in table II.3. (These tables are at the end of this
appendix).
ECONOMETRIC MODEL
-------------------------------------------------------- Appendix II:1
Our model not only permits a hospital's output and fiscal performance
(see discussion later in this section) to influence the chief
executive's compensation, it also allows for the possibility that the
compensation amount may affect the hospital's output and fiscal
performance. If, for example, higher compensation represented the
purchase of "more" executive human capital, then output and fiscal
performance would depend on the compensation paid. Although this
construction seems obvious, most research on executive compensation
has ignored this simultaneity.\13 We used two-stage least squares--a
statistical technique designed to account for the simultaneous nature
of some of the relationships--to measure the effect of various
factors on the compensation of chief executives.\14
We considered an extensive list of factors that could affect chief
executive compensation. Even so, no model could reasonably be
expected to explain all of the variation in compensation. Some
variation is random; if the contributions of individual chief
executives are relatively difficult for hospitals to measure, there
may be greater random variation in chief executives' compensation
compared with that for other professional positions.\15 Also, many
unavailable or unquantifiable factors may influence compensation
amounts: personalities, friendships, institutional rigidities, and
other noneconomic factors. For example, trustees may consider
factors (such as service to the community) on which we have no data
when they set compensation amounts.
--------------------
\13 Most empirical research in the field of executive compensation
has treated firms' output and financial performance as independent,
or exogenous, variables. If, as we believe, the chief executive's
actions affect output, then such treatment produces biased estimates
of output's influence on compensation. For example, when we
intentionally misspecified the model and considered output exogenous,
our results erroneously suggested that executive compensation and
fiscal performance are inversely related. That is, all else equal,
executive compensation would tend to be highest at hospitals that
lose the most money.
\14 See Jan Kmenta, Elements of Econometrics, 2nd ed. (New York:
Macmillan Publishing Co., 1986), for a discussion of the problem of
estimating simultaneous relationships. We used a number of
variables, in addition to those listed in table II.1 to estimate the
first-stage regressions: Medicare and Medicaid discharges, Medicare
and Medicaid inpatient days, HCFA's case mix index, whether the
hospital had a contract with a health maintenance organization or a
preferred provider organization, the hospital's market share in the
county (in terms of the number of beds), county unemployment rate,
average income per capita in the county, county population per county
hospital bed, percentage of population aged 65 or older, and whether
the hospital was an eye or kidney transplant center.
\15 While the chief executive's actions may affect operations to a
degree infinitely greater than any other single employee's actions, a
hospital's success is also influenced by many factors outside the
chief executive's control. Consequently, hospitals may estimate the
marginal contributions of their chief executives with considerable
error. If those estimates are used to determine appropriate
compensation amounts, then equally productive individuals may be
compensated at different rates.
DESCRIPTION OF ANALYSIS
VARIABLES
-------------------------------------------------------- Appendix II:2
The compensation of each hospital's chief executive was the dependent
variable in our model. It was computed as the total of the
executive's hospital salary, taxable benefits, and allowances.\16
This amount was adjusted for inflation using the Consumer Price Index
and reported in 1991 constant dollars. Following common empirical
practice, we entered compensation measured in natural logarithms.\17
From economic theory and empirical research on chief executive
compensation, we compiled an extensive list of factors that could
influence compensation.\18 Each of these factors can be grouped into
one of the five categories listed below.\19
--------------------
\16 Executives may receive additional compensation from related
businesses. (See discussion on pp. 11 and 12.) However, because our
data on this source of compensation was limited and because we had no
information on the extent of any additional responsibilities, we did
not include compensation from related businesses in our econometric
analysis. Therefore, our results should be interpreted as measuring
the impacts of factors associated with running a hospital (strictly
defined) on chief executive compensation.
\17 To understand the theoretical origins of this practice, see Jacob
Mincer, Schooling, Education, and Experience (New York: National
Bureau of Economic Research, 1974). A double-log construction
conveniently allows the coefficients to be interpreted as
elasticities.
\18 We did not include personal characteristics of executives because
these data were unavailable. (This type of information is rarely
obtainable. For one study that did include this information, see
Timothy Hogan and Lee McPheters, "Executive Compensation:
Performance Versus Personal Characteristics," Southern Economic
Journal, 46(4): 1060-1068.) This omission is probably not serious
because a chief executive's recent performance may be a much better
predictor of future performance than traditional measures of human
capital, such as years of education acquired long ago.
\19 For a complete list of the variables included, their definitions,
and data sources see table II.1.
HOSPITAL OUTPUT AND FISCAL
PERFORMANCE
------------------------------------------------------ Appendix II:2.1
The level of hospital output may indicate the enormity of the chief
executive's position or, alternatively, the executive's skill at
running the hospital. If either is the case, then higher
compensation should be associated with greater output. We included
two measures of output: the annual number of inpatient days and
patients discharged.
Chief executives of financially healthier hospitals may receive
greater compensation--either as a reward for their role in producing
fiscal health or simply because of their hospitals' greater ability
to pay--than chief executives at fiscally weaker hospitals. To allow
for this possibility, we included a hospital's "operating margin"
(net operating profit as a percentage of net patient revenues) as a
measure of fiscal performance.
LOCAL MARKET COMPETITIVENESS
------------------------------------------------------ Appendix II:2.2
The degree of competitiveness in the local hospital market could
affect executive compensation for two reasons. First, hospitals in
highly competitive markets may feel compelled to hire more
experienced or skilled--and thus more expensive--individuals for the
chief executive position than similar hospitals in less competitive
markets. Second, highly competitive markets generally have more
hospitals--and thus a greater local demand for chief executives--than
less competitive markets. If the supply of executives is relatively
inelastic, compensation would tend to be higher in those areas with
greater demand. We measured competitiveness with the Herfindahl
index calculated at the county level with market shares determined by
the number of beds in each hospital relative to the county total.
HOSPITAL CHARACTERISTICS
------------------------------------------------------ Appendix II:2.3
Economic theory suggests that the working environment, scope of
responsibility, and level of nontaxable (and therefore unreported)
fringe benefits together will affect the level of compensation a
hospital must offer to attract and retain a suitable chief executive.
Although direct measures of these factors do not exist, we allowed
for the possibility that their net effect may vary between different
types of hospitals. We did this by including a number of variables
that identify hospital characteristics such as number of beds,
ownership type (e.g., for-profit), involvement in medical education,
status relative to other hospitals in the county, and membership in
health systems or hospital alliances.
HOSPITAL LOCATION
------------------------------------------------------ Appendix II:2.4
Compensation amounts may vary between areas because of differences in
the cost of living, amenity levels, or executive labor market
conditions. We introduced a set of dummy variables that allowed
average compensation to vary among the nine Census regions and
between rural and urban areas. We included a hospital wage index,
computed by HCFA, to control for local hospital labor costs. For the
subsample of urban hospitals, we included a measure of urban
population to capture the net effect of factors that vary by city
size.
OTHER CONTROLS
------------------------------------------------------ Appendix II:2.5
In some cases, the individual serving as chief executive changed
midyear. We allowed for the possibility that--because of severance
payments, moving allowances, or other one-time expenses--compensation
could be abnormally high for the hospital in that particular year.
We also included a variable to test whether, as some have suggested,
newly hired chief executives are compensated at a higher rate than
their predecessors. A third variable controlled for those cases
where we annualized the part-year salary reported by the hospital.
Lastly, we added a set of dummy variables to test whether real
compensation--that is, adjusted for inflation--grew between 1989 and
1991.
DATA SOURCES AND SAMPLE
DESCRIPTION
-------------------------------------------------------- Appendix II:3
To amass the information necessary to estimate our model, we tapped
several sources. The executive compensation data we received from
each hospital were matched to data in HCFA's Hospital Cost Report
Information System Minimum Data Set (HCRIS) Minimum Data Set for that
particular hospital in the same year.\20 Each record was then further
augmented with information from AHA's Survey of Hospitals 1989 on
selected characteristics of that hospital not contained in HCRIS, the
1992 area resource file (ARF) on county demographics, and the 1990
Census for urban population counts.
In total, 368 hospitals reported executive compensation data to us.
Most of the respondents, but not all, provided the 3 years of data we
requested. From this group we selected cases where the chief
executive reported working full-time and where we could determine
whether the same individual held the same position the previous
year.\21 The sample was further reduced when records could not be
matched to the AHA or ARF data.\22 Finally, four cases with extreme
financial performance values were excluded from the analysis.\23
The full data set used in the econometric analysis contained 550
observations (i.e., it included multiple years from 247 separate
hospitals). Because the mechanisms that determine the compensation
for chief executives at the very smallest hospitals may not be
identical to those that determine compensation at larger hospitals
(our data included hospitals from 16 beds to 1,365 beds), we analyzed
a subsample that included only those hospitals with at least 50 beds
(498 observations). To allow for the possibility that market
mechanisms in urban markets may differ from those in rural markets,
we analyzed a sample that contained only urban hospitals regardless
of size (402 observations).\24
--------------------
\20 The fifth cycle of HCFA's HCRIS was used to construct some
variables, including the Herfindahl index, for every hospital and
year.
\21 Although the literature suggested that newly hired chief
executives are compensated at a much higher rate than their
predecessors, our econometric results did not support this view. We
also tried excluding this variable (thereby increasing the sample
size) and found that the basic qualitative results remained the same.
However, the estimated coefficient on the operating margin, while
positive, was not statistically significant at conventional levels.
\22 Approximately 15 percent of the short-term acute care hospitals
included in the HCRIS database are not contained in the 1989 AHA
survey. Alaskan hospitals were not included because the ARF does not
provide county-level data for that state.
\23 Cases with an operating margin of less than -2.0 (three cases) or
greater than 0.99 (one case) were considered to be extreme outliers
and were excluded to prevent them from exerting a disproportionate
influence on the regression estimates.
\24 Limited degrees of freedom precluded the separate analysis of
small, medium-size, and large hospitals and also of rural hospitals.
SUPPORTING TABLES AND
ECONOMETRIC RESULTS
-------------------------------------------------------- Appendix II:4
Table II.1 describes each of the variables entered in the econometric
chief executive compensation equation and identifies the data
sources. Table II.2 lists the mean values and standard deviations
for these same variables. These statistics are presented separately
for the three samples used in the econometric analysis: all
hospitals, hospitals with at least 50 beds, and urban hospitals of
any size.
The estimated coefficients from our two-stage least squares
econometric analysis are provided in table II.3. Each coefficient
estimates the effect on (the natural logarithm of) chief executive
compensation resulting from a change in that variable, holding all
other factors constant. The standard errors of these estimates are
shown in parentheses.
Because inpatient days, patients discharged, and operating margin are
measured in natural logarithms, each coefficient estimates the
percentage change in chief executive compensation associated with a
1-percent increase in that variable's value. For example, if two
hospitals were identical in all respects except that the number of
patients discharged was 1-percent higher in one hospital, the
estimates from the all-hospital sample suggest that compensation
would be 0.236-percent higher in the hospital with the greater number
of patients discharged. For continuous variables not measured in
logarithms (i.e., the Herfindahl index, Herfindahl interacted with
county population and wage index for all samples, and metropolitan
population in the urban sample), each estimated coefficient
approximates the percentage change in compensation associated with a
one-unit increase in the variable.\25
Although the remaining variables are dichotomous (i.e., they take on
the value "1" if the observation possesses the described
characteristics and "0" otherwise)\26 the interpretation of their
coefficients is quite similar to the above. These coefficients
estimate the effect of that characteristic on compensation, holding
all other factors constant. This effect is relative to a reference
(or omitted) group.\27 For example, using the formula in footnote 25,
the estimates from the all-hospital sample indicate that a chief
executive of a government hospital would earn approximately 8.9
percent less than the chief executive of an otherwise identical
not-for-profit hospital.
Table II.1
Variable Descriptions and Data Sources
Variable Description Source\a
------------------------- ------------------------- --------------------------
Dependent variable: chief executive compensation
--------------------------------------------------------------------------------
Compensation Reported total of salary, GAO survey
taxable benefits, and
allowances for the
hospital's chief
executive. Adjusted for
inflation using the
Consumer Price Index and
reported in 1991 constant
dollars. Measured in
natural logarithms.
Hospital output and fiscal performance variables
--------------------------------------------------------------------------------
Inpatient days Annual number of HCRIS
inpatient days. For
example, a hospital that
cared for 100 patients,
each of whom stayed in
the hospital for 5 days,
would have produced 500
inpatient days.
Discharges Annual number of patients HCRIS
discharged.
Operating margin Profitability on patient HCRIS
care operations, measured
as net patient revenues
less patient care costs
as a proportion of net
patient revenues.
Local market competitiveness
--------------------------------------------------------------------------------
Herfindahl index Measure of the HCRIS PPS-V,
concentration of computed
hospitals within a
county. Computed by
summing the squared
market shares (based on
the number of beds) of
each hospital within a
county. Index ranges from
near 0 (highly
competitive hospital
market--many small
hospitals) to 1
(monopoly--only one
hospital in the
county).\b
Hospital characteristics
--------------------------------------------------------------------------------
Hospital size Hospitals were classified HCRIS
into 1 of 3 categories
based on the number of
hospital beds: small (1-
100 beds), medium-size
(101-500 beds), and large
(501 or more beds).
"Small" and "large"
measure compensation
received at those
institutions relative to
compensation received at
a medium-size hospital.
For-profit Equals 1 if operated as a HCRIS
for-profit hospital, 0
otherwise. Measures
compensation relative to
that received at an
otherwise similar not-
for-profit hospital.
Government Equals 1 if operated as a HCRIS
state or local government
hospital, 0 otherwise.
Measures compensation
relative to that received
at an otherwise similar
not-for-profit hospital.
Medical education Equals 1 if member of AHA
Council of Teaching
Hospitals of the
Association of American
Medical Colleges or
reports medical school
affiliation to the
American Medical
Association, 0 otherwise.
Biggest Equals 1 if the largest HCRIS,
hospital in the county, 0 computed
otherwise.
System Equals 1 if member of a AHA
member health care system, 0
otherwise.
Alliance Equals 1 if member of an AHA
alliance, 0 otherwise.
Management contract Equals 1 if the hospital AHA
is contract managed, 0
otherwise.
Holding company Equals 1 if the hospital AHA
is a division or
subsidiary of a holding
company, 0 otherwise.
Subsidiaries Equals 1 if the hospital AHA
itself operates
subsidiary corporations,
0 otherwise.
Hospital location
--------------------------------------------------------------------------------
New England Equals 1 if located in HCRIS
Connecticut, Maine,
Massachusetts, New
Hampshire, Rhode Island,
or Vermont; 0 otherwise.
Note: This and each of
the other regional
variables measure
compensation in that
region relative to
compensation received at
otherwise similar
hospitals in the South
Atlantic States, our
geographic reference
region.
Mid-Atlantic Equals 1 if located in HCRIS
New Jersey, New York, or
Pennsylvania; 0
otherwise.
South Atlantic The geographic reference HCRIS
group for our analysis.
It includes Delaware,
Maryland, West Virginia,
Virginia, North Carolina,
South Carolina, Georgia,
Florida, and the District
of Columbia.
E. N. Central Equals 1 if located in HCRIS
Illinois, Indiana,
Michigan, Ohio, or
Wisconsin; 0 otherwise.
E. S. Central Equals 1 if located in HCRIS
Alabama, Kentucky,
Tennessee, or
Mississippi; 0 otherwise.
W. N. Central Equals 1 if located in HCRIS
Iowa, Kansas, Minnesota,
Missouri, Nebraska, North
Dakota, or South Dakota;
0 otherwise.
W. S. Central Equals 1 if located in HCRIS
Arkansas, Louisiana,
Oklahoma, or Texas; 0
otherwise.
Mountain Equals 1 if located in HCRIS
Arizona, Colorado, Idaho,
Montana, New Mexico,
Utah, or Wyoming; 0
otherwise
Pacific Equals 1 if located in HCRIS
Alaska, California,
Hawaii, Oregon, Nevada,
or Washington; 0
otherwise.
Urban Equals 1 if located in a HCRIS,
metropolitan statistical Census
area (MSA), 0 otherwise.
For the sample of urban
hospitals this variable
measures 1990 urban
population of the MSA (or
consolidated MSA if one
exists) in natural
logarithms.
Wage index Hospital labor cost HCFA Hospital Wage Index
index, by county. Survey
Other controls
--------------------------------------------------------------------------------
Midyear Equals 1 if chief GAO survey
executive changed
midyear, 0 otherwise.
New chief executive Equals 1 if individual GAO survey
was not the chief
executive in the previous
year, 0 if he or she was
chief executive.
Annualized Equals 1 if hospital GAO survey
reported chief executive
compensation for a period
of less than 12 months
(amounts were annualized
in these cases), 0
otherwise.
Year=1990 Equals 1 if data were GAO survey
Year=1991 from that year (1990 or
1991), 0 otherwise.
Measures change in
inflation-adjusted
compensation from 1989
levels.
--------------------------------------------------------------------------------
\a Key to data sources: GAO survey = GAO Survey of Hospitals; AHA =
American Hospital Association Survey of Hospitals, 1989; HCRIS = HCFA
Medicare Hospital Cost Report Information System (PPS-V, VI, VII, and
VIII)
\b Counties vary considerably in size, and their political boundaries
may not always coincide with the relevant market area for some
hospitals. Large counties will tend to contain many hospitals and,
consequently, have a low Herfindahl index. Conversely, small
counties will tend to have few hospitals and a low Herfindahl index.
To partially control for this, we also interacted county population
as of 1990 with the Herfindahl index.
Table II.2
Variable Means and Standard Deviations
Standa Standa Standa
rd rd rd
deviat deviat deviat
Variable Mean ion Mean ion Mean ion
-------------------------------- ------ ------ ------ ------ ------ ------
Chief executive earnings $124,3 $266,0 $145,4 $241,8 $162,8 $240,5
02 32 51 52 36 32
Small 0.44 1.72 0.27 1.41 0.18 1.13
Large 0.05 0.77 0.07 0.80 0.10 0.88
Inpatient days 48,058 178,97 58,581 173,77 71,123 175,36
8 5 6
Discharges 6,695 26,531 8,501 25,237 10,196 25,820
Operating margin -0.06 0.67 -0.04 0.50 -0.03 0.35
Herfindahl 0.55 1.22 0.47 1.08 0.35 0.86
Herfindahl X 54.11 148.87 65.36 134.55 81.99 112.73
county population
For-profit 0.12 1.11 0.15 1.15 0.16 1.09
Government 0.33 1.63 0.21 1.29 0.18 1.14
Sole community hospital 0.12 1.13 0.07 0.81 N/A N/A
Medical education 0.20 1.39 0.26 1.39 0.34 1.42
Biggest 0.54 1.72 0.48 1.59 0.36 1.43
System member 0.26 1.52 0.31 1.47 0.35 1.42
Alliance member 0.26 1.51 0.34 1.50 0.35 1.42
Management contract 0.06 0.81 0.04 0.63 0.07 0.74
Holding company 0.24 1.48 0.30 1.45 0.32 1.39
Subsidiaries 0.16 1.25 0.20 1.28 0.23 1.26
Urban 0.54 1.72 0.66 1.50 1.00 0.00
Urban population N/A N/A N/A N/A 3,093, 14,025
105 ,320
New chief executive 0.14 1.19 0.13 1.05 0.12 0.96
Annualized 0.01 0.36 0.01 0.29 0.01 0.29
Midyear 0.03 0.58 0.03 0.56 0.03 0.54
New England 0.05 0.75 0.07 0.78 0.06 0.72
Mid-Atlantic 0.10 1.03 0.13 1.06 0.16 1.09
E. N. Central 0.19 1.36 0.23 1.35 0.24 1.26
E. S. Central 0.05 0.77 0.06 0.77 0.04 0.61
W. N. Central 0.11 1.09 0.05 0.66 0.09 0.86
W. S. Central 0.13 1.18 0.11 1.00 0.08 0.82
Mountain 0.10 1.04 0.05 0.72 0.04 0.55
Pacific 0.08 0.95 0.11 0.99 0.11 0.93
Wage index 9.20 5.29 9.45 4.92 10.08 4.12
Year=1990 0.40 1.69 0.41 1.56 0.41 1.47
Year=1991 0.38 1.67 0.38 1.55 0.40 1.46
--------------------------------------------------------------------------------
N/A = not applicable
Table II.3
Two-stage Least Squares Estimates of
Effects of Factors on Chief Executive
Compensation
(Dependent variable = ln (Earnings))
More
than
50 Urban
All beds \a
------------------------------------ ------ ------ ------
Small (fewer than 100 beds) - - -
0.284 0.127 0.455
\b \c \b
(0.050 (0.060 (0.072
) ) )
Large (more than 500 beds) 0.055 - 0.058
(0.064 0.020 (0.062
) (0.069 )
)
Inpatient days \d,e 0.004 0.054 0.056
(0.027 (0.044 (0.051
) ) )
Discharges \d,e 0.236 0.279 0.218
\b \b \b
(0.031 (0.049 (0.043
) ) )
Operating margin \e - 0.523 0.644
0.051 \c \f
(0.220 (0.250 (0.341
) ) )
Herfindahl - - -
0.199 0.140 0.160
\b \f \f
(0.074 (0.082 (0.094
) ) )
Herfindahl X Population - -
0.001 0.001
(0.001 \f
) (0.001
)
For-profit 0.111 0.145 0.193
\c \b \b
(0.045 (0.051 (0.058
) ) )
Government - - -
0.093 0.076 0.099
\c (0.047 \f\
(0.043 ) (0.052
) )
Sole community hospital 0.178 0.131
\b \c
(0.047 (0.063
) )
Medical education 0.054 0.017 0.056
(0.037 (0.040 (0.038
) ) )
Biggest 0.108 0.107 0.014
\b \c (0.048
(0.042 (0.045 )
) )
System member 0.085 0.070 -
\b \c 0.001
(0.032 (0.034 (0.037
) ) )
Alliance - - -
0.018 0.042 0.087
(0.033 (0.035 \c
) ) (0.036
)
Management contract - 0.000 0.014
0.056 (0.077 (0.073
(0.057 ) )
)
Holding company 0.037 0.005 0.029
(0.033 (0.036 (0.036
) ) )
Subsidiaries 0.015 0.008 0.054
(0.034 (0.036 (0.037
) ) )
Urban \g 0.012 0.025 0.037
(0.043 (0.047 \c
) ) (0.016
)
New chief executive - - -
0.026 0.070 0.057
(0.042 (0.047 (0.052
) ) )
Annualized 0.112 0.233 0.207
(0.111 (0.151 (0.152
) ) )
Midyear 0.420 0.509 0.502
\b \b \b
(0.078 (0.094 (0.105
) ) )
New England \h 0.076 0.113 0.127
(0.074 (0.081 (0.085
) ) )
Mid-Atlantic 0.031 0.043 -
(0.055 (0.059 0.037
) ) (0.060
)
E. N. Central - - -
0.036 0.023 0.088
(0.041 (0.046 \f
) ) (0.051
)
E. S. Central 0.028 0.069 0.032
(0.059 (0.066 (0.078
) ) )
W. N. Central 0.039 0.141 0.140
(0.057 \f \f
) (0.079 (0.080
) )
W. S. Central - 0.100 -
0.026 (0.067 0.173
(0.057 ) \b
) (0.063
)
Mountain - - -
0.042 0.086 0.196
(0.057 (0.075 \c
) ) (0.086
)
Pacific - - -
0.280 0.364 0.221
\b \b \b
(0.067 (0.075 (0.079
) ) )
Wage Index 0.063 0.087 0.009
\b \b (0.022
(0.016 (0.019 )
) )
Year=1990 0.009 0.039 0.048
(0.031 (0.037 (0.039
) ) )
Year=1991 0.084 0.114 0.095
\b \b \c
(0.032 (0.038 (0.039
) ) )
Intercept 9.147 7.976 8.839
\b \b \b
(0.334 (0.427 (0.440
) ) )
Adjusted R \2 0.80 0.70 0.73
N observations 550 498 402
F-statistic 71.22 36.45 36.69
------------------------------------------------------------
Note: Standard errors are in parentheses. Although the equations
contain a large number of variables, diagnostic techniques indicated
that collinearity is not a serious problem. These techniques are
discussed in Belsley, Kuh, and Welsh, Regression Diagnostics (New
York: Wiley, 1980).
\a No sole community hospitals are in the urban sample; hence, this
variable is excluded. The county population variable interacted with
the Herfindahl index is omitted because the "urban" variable measures
the urban population.
\b Significant at the 1-percent level.
\c Significant at the 5-percent level.
\d Variable measured in natural logarithms.
\e Endogenous variable.
\f Significant at the 10-percent level.
\g Measures MSA population (or consolidated MSA, if appropriate) in
natural logarithms for urban population. Otherwise, indicates
urban/rural area.
\h Reference region is South Atlantic. See table II.1 for list of
states in each region.
--------------------
\25 This approximation is closest when the estimated coefficient is
near zero. The actual percentage change is calculated by the formula
e\b - 1
where b is the estimated coefficient. The untransformed
coefficients, reported in Table II.3, are approximations of the
characteristics' effects.
\26 Also known as "dummy" variables.
\27 The reference group for "small" and "large" is hospitals with
between 100 and 500 beds. The reference group for "for-profit" and
"government" is not-for-profit nongovernment hospitals. The
reference group for the set of regional variables is hospitals
located in the South Atlantic States. The reference group for the
other dichotomous variables is hospitals that do not possess the
indicated characteristic, e.g., hospitals that are not part of health
systems is the reference group for "system member."
BOARD ROLE IN EXECUTIVE
COMPENSATION
========================================================= Appendix III
A hospital governing board has responsibility for the hospital's
overall strategy and policies, its mission and financial performance,
and its compensation strategy. In deciding compensation levels and
policies for chief executives, board members may collect data on
compensation levels prevalent in the area where the hospital is
located and among executives at comparable hospitals.\28 Boards may
also consider the executive's contribution to achieving the short-
and long-term goals of the hospital when setting compensation levels.
Increasingly, board decisions regarding executive compensation have
come under scrutiny from the public, state legislators, the IRS, and
shareholders. Partly in response to the increased scrutiny, more
boards link executive compensation, through incentive or merit
compensation plans, to the individual's performance in helping the
hospital meet its goals. Performance measures can be financial, such
as revenue targets or increases in net income; service related, such
as new programs or market share; or human resource related, such as
productivity increases or recruitment goals.
More recently, some boards have begun to include performance measures
that focus on the community's health status, such as the chief
executive's efforts to address community health care needs and
improve the health status of area residents. In this regard, AHA and
the American College of Healthcare Executives have outlined various
health care criteria that could be considered in evaluating executive
performance.\29 These criteria include how well an executive
contributes to health promotion and disease prevention, implements
processes to provide high-quality health care, plans for the
hospital's future, ensures compliance with regulations, and prepares
future leaders of health care organizations.
--------------------
\28 For example, compensation data can be obtained from the sources
described below. The annual Hay Hospital Management/Professional
Compensation Survey sponsored by the American Society for Healthcare
Human Resources Administration and the Hay Group. The 1992 survey
contains information from about 1,300 U.S. hospitals on compensation
for executive management, nursing, and professional/technical
positions. Study results show 1991-92 pay levels in hospitals
increased more rapidly than in general industry; hospital executives
and managers are now paid more competitively with their counterparts
in other fields than previously; and hospital chief executives
received average cash compensation of $151,000.
Another source is the Report on Compensation in Hospitals, Governor's
Task Force on Public Sector Compensation, April 1993. The task force
requested information on salary and other forms of compensation for
the 3 highest paid executives from New York's nearly 250 licensed
hospitals; all but 5 responded. Cash compensation, including bonuses
and payments from related businesses but not the value of benefits,
ranged from $54,000 in northern New York to $810,000 in New York
City. Additional results showed compensation was highest in New York
City and lowest in the less populated upstate areas; larger hospitals
and teaching hospitals paid more than smaller hospitals and
nonteaching hospitals; and not-for-profit hospitals paid more than
government hospitals. Regarding benefits, the study showed 6 percent
of hospitals provided housing or housing allowances; most provided
automobiles, life insurance, and retirement benefits; 33 percent
provided severance packages and educational benefits to executives
and their families; and 14 percent offered relocation benefits.
\29 For a more detailed discussion of health status as an element in
evaluating executive performance, see AHA and the American College of
Healthcare Executives, Evaluating the Performance of the Hospital CEO
in a Total Quality Management Environment, 1993.
MANAGEMENT SERVICES CONTRACTS
========================================================== Appendix IV
Between 6 and 15 percent of hospitals paid their chief executives
through management services contracts. These contracts can include
items other than chief executive compensation, e.g., salaries and
benefits for other top management executives, services such as data
processing and collections and billings, and overhead. For instance,
one not-for-profit hospital with 117 beds that reported no direct
payment of compensation to its chief executive paid $295,193 in 1991
to an unaffiliated management services firm. This contract covered
not only the chief executive but also at least two other top
management positions and the organization's management fee.
Hospitals with management contracts in 1991 reported that their
contract amounts ranged from $57,200 to $10.2 million. About
one-fourth of the management services contracts were with businesses
directly related to the hospital. Hospital management contracts with
related businesses were generally used by not-for-profit hospitals.
Contracts with related businesses were generally for higher dollar
amounts than management service contracts with unrelated businesses,
ranging from about $780,000 to $10.2 million. Because hospitals were
not required to itemize the content of payments for contract
services, we do not know if these higher contract amounts reflect
higher levels of compensation for chief executives or simply include
payments for additional services.
CHIEF EXECUTIVE COMPENSATION BY
HOSPITAL SIZE, 1989-91
=========================================================== Appendix V
(Amounts expressed in 1991 dollars)
25 25
percen percen
t t
receiv receiv
Sampli ed ed
Hospital Ye ng more less
size ar Mean Error Median than than
------------ -- ------ ------ ------ ------ ------
Small 19 $63,61 $6,293 $62,71 $76,70 $48,61
89 7 9 7 2
19 65,290 7,063 61,837 75,547 49,474
90
19 66,606 7,442 61,936 79,179 50,316
91
Medium 19 159,20 10,038 142,55 184,23 108,97
89 8 1 7 2
19 169,52 17,680 150,02 190,80 115,82
90 0 1 6 2
19 171,19 8,505 157,85 207,95 128,09
91 2 1 8 2
Large 19 273,84 28,673 232,09 300,30 193,40
89 6 5 1 0
19 280,85 31,595 243,48 301,01 195,24
90 5 6 3 4
19 277,35 21,182 253,51 326,33 205,31
91 2 7 5 4
--------------------------------------------------------
Note: Small=1 to 100 beds, medium=101 to 500 beds, large=over 500
beds.
CHIEF EXECUTIVE COMPENSATION BY
HOSPITAL OWNERSHIP TYPE AND SIZE,
1991
========================================================== Appendix VI
25 25
percen percen
t t
receiv receiv
Sampli ed ed
Si ng more less
Ownership ze Mean Error Median than than
------------ -- ------ ------ ------ ------ ------
Not-for- S $75,74 $14,65 $65,82 $94,65 $58,19
profit 3 8 8 5 0
M 175,63 8,275 169,18 212,03 136,15
1 6 0 2
L 288,77 24,278 264,79 332,25 214,35
4 6 9 0
For-profit S 72,129 30,319 72,518 101,29 45,300
1
M 181,54 37,133 157,85 214,99 130,09
1 1 3 0
L \a \a \a \a \a
Government S 55,132 5,369 57,190 64,512 47,967
M 138,53 23,195 129,25 162,93 102,25
6 3 7 0
L 240,05 51,116 222,55 279,45 164,24
0 1 0 8
--------------------------------------------------------
Key: Small (S)=1 to 100 beds, medium (M)=101 to 500 beds, large
(L)=over 500 beds.
\a Too few cases reported to develop representative figures.
Note: Among hospitals that are similar in all respects except
ownership, for-profit hospitals tend to pay higher compensation to
chief executives than not-for-profit hospitals do. This fact is not
clearly evident above because the table does not account for other
important characteristics (besides size) that vary between for-profit
and not-for-profit hospitals.
(See figure in printed edition.)Appendix VII
GAO SURVEY INSTRUMENT
========================================================== Appendix VI
(See figure in printed edition.)
MAJOR CONTRIBUTORS TO THIS REPORT
======================================================== Appendix VIII
Frank C. Pasquier, Assistant Director, (206) 287-4861
Victoria C. Marcella
Michael J. O'Dell
Patricia A. Padilla
Alfred R. Schnupp
Stanley G. Stenerson
Evan L. Stoll