In 1983, in response
to rapidly rising costs associated with retrospective
cost-based reimbursement, Congress established
the Medicare Prospective Payment System (PPS)
to pay for inpatient services. In important respects,
the PPS is analogous to an experience-rated insurance
scheme with mandatory reinsurance. In exchange
for assuming financial risk for caring for a patient
during a hospital stay, the provider receives
a prospectively determined payment. This payment
may be thought of as an experience rated "premium,"
determined on the basis of patient diagnosis at
discharge, based on Diagnosis Related Groups (DRGs)
and selected provider and geographic characteristics.
Providers are also eligible for retrospective
"outlier" payments based on the actual
costs of caring for unusually costly patients.
These retrospective payments may be thought of
as a form of reinsurance (Keeler et al. 1988)
in which 1) participation is mandatory (i.e.,
opting out of reinsurance provisions is not an
option); and 2) there is an implicit premium imposed
on providers through a downward adjustment in
PPS-based payments to finance outlier payments
(Federal Register 1983).
The underlying notion in the PPS is that, by
placing providers at financial risk for the cost
of the services they provide incentives can be
created to increase efficiency across the board
(Schliefer 1985). Two related problems exist with
this approach. First, providers may respond strategically
to financial risk in addition to or in lieu of
attempting to increase efficiency, for example,
by manipulating quality and/or denying access
to care for unprofitable patients. Second, providers'
exposure to risk may vary for reasons beyond their
control. Thus, there may be systematic differences
in providers' expected costs of treating patients
due to differences in input costs or inter-DRG
case mix. Providers' costs may also vary because
of nonsystematic (random) fluctuations in the
treatment needs of patients within DRGs (Dranove
1987, Allen and Gertler 1991, Ellis and McGuire
1988).
Outlier payments have been justified on the grounds
that they mitigate incentives for providers to
avoid unusually costly patients or to forego appropriate
treatments, and that they systematically redistribute
revenues to providers who face a greater risk
of treating high-cost patients (ProPAC 1997).
However, even with outlier payments and other
existing adjustments, an important public policy
issue remains: Can current PPS rules can be modified
to make the distribution of financial risk more
equitable and/or efficient? In the face Of concerns
regarding fairness and strategic behavior, two
major types of modifications have been proposed
in PPS payment rules to redistribute providers'
exposure to financial risk. The first is refinements
in experience rating; for instance, in DRG categories
or in adjustments to payments on the basis of
provider characteristics (Edwards et al. 1994).
The second is refinements and/or expansion of
mandatory reinsurance provisions; for example,
through use of "mixed" payment systems
increasing the retrospective component of payments.
(Goodall 1990) Both refinements in experience
rating and in reinsurance provisions can simultaneously
alter providers' exposure to financial risk associated
with systematic and nonsystematic variations in
costs. Efforts to evaluate the financial risk
implications of proposed modifications have typically
examined the impact on the mean and variance of
patient profitability. Changes in expected profits
provide a direct monetary measure of changes in
systematic financial risk associated with alternative
proposals. The mean and variance of profits do
not, however, provide a comparable measure of
either the impact on nonsystematic (random) risk
or on the total risk faced by a provider in a
given period (i.e., the combined financial risk
associated with systematic and nonsystematic variations).
In addition, it is important to consider implications
of random variations for provider as well as patient-level
profitability. This is because the provider, rather
than the patient, is often the relevant unit for
analysis. The impact of random variations at the
provider level will depend not only on the degree
of variation at the patient level, but also on
total volume at the provider level; as the number
of discharges increases, financial risk associated
with random variations will approach zero.
Serious concerns about the rising costs of health
care have led to numerous proposals for reform.
Many of these recommendations (e.g., capitated
payments, managed care systems) focus on cost
containment strategies that constrain growth through
the prospective reimbursement for specific services.
Consequently, they represent a form of price control
designed to reduce revenues of health care providers.
While the cost reducing benefits of these price
controls have been promoted to the public, little
attention has been paid to the negative consequences,
particularly the effect on capital accumulation
in the health care industry. The impact of price
controls on hospitals' ability to raise capital
is an important issue. Gray (1993) notes, that
the move of the industry toward managed care and
large, integrated health care systems has put
a premium on those organizations having access
to capital. However, just as rent controls have
led to a reduction in the investment in rental
properties, price controls on hospital services
would be expected to lead to reduced investment
in the industry as well. In addition, hospitals
traditionally have required access to both long-term
and short-term capital to upgrade technology,
support new capital improvements, and maintain
the existing physical plant. Thus, corporate hospitals
need to issue new stock and/or accumulate retained
earnings to generate needed equity capital for
such future needs (Grossman, Goldman, Nesbitt,
and Mobilia, 1993). Lastly, both corporate and
not-for-profit (NFP) hospitals have relied heavily
in the past, and most likely will rely in the
future, on the bond market for their capital needs
(Grossman et al, 1993, Prince and Ramanan, 1994,
and Wedig, Sloan, Hassan, and Morrisey, 1988).
Regulatory experience with price controls in the
health care industry is limited; however, Congress
enacted over a decade ago the Tax Equity and Fiscal
Responsibility Act (TEFRA) that established target
rates per case for each hospital. The diagnosis-related
group-based (DRG) Medicare Prospective Payment
System (PPS), enacted one year later, set price
controls on Medicare hospital reimbursements.
Although these events have generated considerable
research, the focus of most studies has been limited
to the effect of PPS on hospital utilization patterns
or expenditures. Few studies have focused on the
effect on capital markets in the hospital sector.
Background
The Episode of PPS Legislation
Rapid growth in Medicare expenditures from rising
medical care costs prompted Congress to pass TEFRA,
signaling the end of retrospective cost reimbursement
from Medicare and unrestricted growth in the hospital
industry. TEFRA constrained rates of increase
in Medicare payments to hospitals by setting target
rates per case based on an inflation factor applied
to the hospital's base-year (1981) cost (Feder,
Hadley, and Zuckerman, 1987). Also, as part of
this legislation, Congress required that a proposal
for a prospective payment system for Medicare
be presented by the Secretary of Health and Human
Services (HHS) by the end of the year (1982).
Unlike TEFRA, the proposed system would offer
significant cost-reducing incentives for hospitals
by establishing rates based on averages per case
among hospitals (Feder et al, 1987).
After the December PPS proposal, Congressional
response was immediate (Congressional Quarterly
Almanac, 1983). In less than three months, both
the House and Senate passed PPS legislation as
part of the Social Security Amendment of 1983.
The President signed the bill one month later.
See Table 1 for a synopsis of Congressional actions.
The objective of PPS was to slow the growth in
hospitals' costs while providing access to quality
health services for Medicare beneficiaries (Guterman
and Dobson, 1986). Under PPS, payments are made
at predetermined fixed rates, representing the
average cost nationally of treating a Medicare
patient according to the patient's classification
into one of more than 470 DRGs.
Beginning October 1, 1983, the implementation
of PPS was to take place over a three-year period;
however, in 1986, this was extended one additional
year (Sloan, Morrisey, and Valvona, 1988). To
ease the pain of implementation, this phase-in
allowed a declining proportion of the total PPS
rate to be based on each hospital's historical
costs (Guterman and Dobson, 1986). Medicare payments
for capital costs and medical education were excluded
along with certain hospitals, such as psychiatric,
rehabilitation, children's, and long-term care.
By excluding capital costs from PPS, capital payments
continued to be paid on an allowable-cost basis,
but this policy generated much debate and uncertainty
throughout implementation of the law.
Thus, PPS legislation and its implementation
occurred over a long period with three specific
events: the passage of TEFRA (September 3, 1982);
the enactment of PPS (March 25, 1983); and, the
promulgation of PPS regulations (October 7, 1983).
While the phase-in of PPS was to take place over
an extended period, much of the rule promulgation
had occurred by the end of 1985, and the direction
that PPS was taking the Medicare reimbursement
system was readily apparent. Consequently, the
period of analysis in this paper runs from January
1982 through December 1985. The period is sufficiently
long to capture the full effect of PPS legislation
yet sufficiently parsimonious to limit the number
of other confounding events.
Previous Research--Market Effects of
PPS
Previous work has analyzed the reaction of the
stock market to PPS legislation. However, these
studies focus primarily on the impact of the passage
of PPS immediately around the event and do not
address the capital acquisition issue. This paper's
focus on the effect of PPS over time differentiates
it from earlier market-evaluation papers (Asper
and Hassan, 1993, Folland and Kleiman, 1990, and
Jacobson, 1994).
Folland and Kleiman (1990) employ an X-inefficiency
model suggesting that price regulation increases
the cost of non pecuniary benefits, thereby, reducing
their use. Using daily return data, they find
a significant positive market response for a four-stock
portfolio around the passage of PPS (March 24,
1983). Since not all firms had significant abnormal
returns around that date, Folland and Kleiman
(1990, p 64.) ". . . conclude overall that
the market did not respond decisively to the DRG
legislation." Their results provide only
weak support for their hypothesis that these hospitals
were not cost efficient. Moreover, it is difficult
from their results to reach a definitive answer
to the question of how PPS affected hospital systems'
ability to generate capital.
Jacobson (1994) also used daily return data but
tested portfolios that contained not only direct
providers of health care but also distributors
of hospital equipment, hospital supplies, and
manufacturers of medical equipment. She did not
find significant abnormal returns for any of the
event dates tested, including March 25, the day
the Senate approved the conference report. Her
results suggest that the impact of PPS on capital
generation by hospitals was neutral, yet Folland
and Kleiman (1990) show significantly positive
residuals around March 25. Jacobson's insignificant
results appear to be due to the inclusion of a
much more diverse group of firms in her portfolio
while the portfolios used in this study and Folland
and Kleiman (1990) contain only acute-care hospital
firms.
While Jacobson (1994) and Folland and Kleiman
(1990) focus on returns, Asper and Hassan (1993)
focus on the change in riskiness for publicly
traded hospital stocks due to the passage of PPS.
They test for changes in risk by comparing betas
for periods before and after the passage of PPS.
Their results show a decrease in systematic risk
after passage of PPS, but they find changes in
risk for some control industries as well.
Conclusion
They conclude that while the acute-care hospital
portfolio became less risky after PPS, the decrease
in risk may not have been unique to health care.
However, Asper and Hassan (1993) use monthly return
data and are forced to use very long periods to
generate sufficient observations. Their estimating
equation covers January 1976 to August 1982, while
the post period covers August 1983 to December
1988. These long periods could easily contaminate
the data by including many other events that affect
returns. Health care reform is a continuous time-dependent
process; therefore, the episode of reform needs
to be analyzed thoroughly over the full period
if there is to be an accurate understanding of
the effect on the market and capital accumulation
in the health care industry. The next section
describes the methodology used.
|