Medicaid Matching Formula: Effects of Need Indicators on New York's Funding

Published by the Government Accountability Office on 1997-06-09.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

      United States
GAO   General Accounting
                    D.C. 20848

      Health,   Education   and Human Services Division

      June 9, 1997
      The Honorable Daniel Patrick Moynihan
      United States Senate
      The Honorable Alfonse M. D’Amato
      United States Senate

      Subject: Medicaid Matching Formula: Effects of Need Indicators on New York’s
      This letter responds to your request of April 29, 1997,for an analysis of what
      New York’s federal Medicaid matching percentage and funding would have been
      if the current matching formula had included certain factors that affect state
      financing burdens, such as the number of people in poverty and the cost of
      health care services.
      The current Federal Medical Assistance Percentage (FMAP) is based on the per
      capita income of the state compared with the per capita income of all states.
      Under current law, no state can receive a matching percentage that is less than
      60 percent or more than 83 percent. In fiscal year 1996,Mississippi had the
      lowest income and qualified for a matching percentage of 78 percent. Because
      of their above averageper capita income, New York and 11 other states
      received the minimum SO-percentmatch.
      In testimony before the Finance Committee in July 1996,we noted that the
      legislative history of the matching formula suggeststhat higher matching rates
      were granted to low-income states in an effort to offset their greater financing
      burden compared with states with larger tax bases. However, state financing
      burdens continue to vary becausethe matching formula does not fully account
      for state poverty counts, the cost of health care services, and the concentration
      of high-cost recipients (for example, the elderly and people with disabilities).
      While data on these indicators were not available when the current matching
      formula was adopted into law, they are available today.
      As you requested, we calculated federal matching percentages based on four
      factors affecting state financing burdens: (1) state tax bases, (2) the number of
      low-income residents living in poverty, (3) the cost of delivering health care

                                        GAWHEHS-97-162R   Medicaid   Matching   Formula
services, and (4) the concentration of elderly and disabled recipients. However,
we were unable to take state differences in the cost of living into account in
measuring the number of low-income residents, as you had requested. Instead,
we used the official poverty counts reported by the Bureau of the Census.’

To measure the cost of health care services, we used two different indicators:
(1) an index of hourly wages paid to hospital workers compiled by the Health
Care Financing Administration (HCFA) that is used in the Medicare Hospital
Reimbursement program and (2) an index of wages per worker in the health
care industry that is compiled by the Bureau of Labor Statistics (BLS). Because
the two measures differ in the degree to which they show New York to be a
high-cost state, we did our analysis using both measures.2 Finally, we
developed an indicator to measure the extent to which the state’s caseload is
composed of those who are comparatively more expensive to serve-elderly and
disabled people.3 We conducted our work during May of 1997. Except that we
did not verify data obtained electronically from federal agencies, we did our
work in accordance with generally accepted government auditing standards.

In brief, we estimate that had the Medicaid matching formula reflected poverty
and other factors affecting state financing burdens between fiscal years 1989
and 1996, New York’s matching percentage would have fluctuated between 50
and 60 percent, depending on the particular year in question and the indexes
used. On the basis of these rates, we estimate that the state would have
received between $3.4 billion and $6.5 billion in additional federal assistance

‘Official poverty counts do not take into account state differences in the cost of
living, and experts disagree on how this might best be done. However, cost-of-
living measures examined in our report, Povertv Measurement: Adiusting for
Cost-of-Living Differences, (GAO/GGD-9564, Mar. 9, 1995), indicate that New
York is a high cost-of-living state. Consequently, our estimates of New York’s
matching percentage and increases in federal funding are understated compared
with what would prevail if such differences were reflected in the official
poverty counts.
‘We were able to obtain data for HCFA’s survey of hospital workers for the
years 1988 through 1993. For years prior to 1988, we used BLS wage data to
reflect the cost of services. See enc. I for a more detailed discussion of the
strengths and weaknesses of the HCFA and BLS wage cost factors.
 3See enc. II for a more detailed description of how we constructed this caseload

 2                          GAO/HEHS-97-152R       Medicaid   Matching   Formula
during this periodn4 Because recent trends in poverty and health care costs
have adversely affected the state’s financing burden, the state’s matching
percentage would continue to increase in fiscal years 1997 and 1998. With a
modified matching formula we estimate federal funding for New York’s
Medicaid program would be between $4.9 billion and $7.2 billion higher than
under the current formula based on projected spending for these years.


Trends in each of the factors affecting New York’s Medicaid financing burden
are summarized in table 1 for the years 1986 to 1995. Each factor was
expressed as a percentage of the corresponding U.S. average to facilitate a
comparison of state trends relative with other states.

4Spending for each year between 1989 and 1996 was adjusted to represent the
purchasing power of 1996 dollars for the United States using the medical care
component of personal consumption expenditures from the National Income
and Product Accounts published by the Department of Commerce.

3                         GAO/HEIHS-97-152R      Medicaid   Matching   Formula

Table 1: Trends in Factors Affecting New York’s Medicaid Financing Burden

  1990              119           102     107         119        101
     1989           119           102     106         NA         100
     1988           119           101     105         NA         100
     1987           117           105      105        NA          99
     1986           115           107      105        NA          99

Note: NA means data were not available.

“All variables are computed based on 3-year averages to improve the reliability
of the estimates and reflect the underlying trends in the data.

bBased on federal fiscal years.

New York’s tax base, as measured by its TTR, compared with the national
average has remained stable since 1988. Trends in New York’s poverty rate,
health care costs, and caseloads, however, have all worsened compared with
the national average. Relative poverty rates, after declining in the late 198Os,
have steadily increased, rising from 1 percent above the national average in
 1990 to 15 percent above average in 1995. Wage costs of workers in the health
care industry, compiled by BLS, have risen from 5 to 9 percent above the

 4                          GAOElEHS-97-152R       Medicaid   Matching   Formula

national average between 1986 and 1995.5 Finally, though less dramatic, there
has been a steady increase in high-cost recipients served by New York’s
program, relative to other states, increasing from just below the national
average in 1984 to about 4 percent above the national average.


If the federal matching formula had been designed to account for the financing
burden of states, New York’s matching percentage would have been higher than
the 50-percent minimum in all but 1 year since 1989. Using the information in
table 1, we have estimated what New York’s matching percentage would have
been had these factors been taken into account in calculating federal matching
percentages.6 Separate calculations were made using cost adjustments based
on the HCFA and BLS wage surveys. The results of these calculations are
shown in table 2.7

5The cost of hospital workers based on HCFA surveys is substantially higher
than the cost of health care industry workers as reported by BLS. Because of
the limited availability of data from the HCFA survey, we are unable to
compare the trend in this measure of the cost of health care services.
‘?itIe XIX, section 1905 of the Social Security Act requires that state FMAPs be
calculated 2 years prior to the fiscal year in which they will be used. The
legislation also requires that the most recently available per capita income data
be used in these calculations. Because of these lags, each year’s matching
percentage reflects data that are from 3 or more years prior to the period for
which they are used to calculate federal reimbursements of state Medicaid
spending. We included a comparable lag in our calculation of alternative
matching rates. For example, fiscal year 1996 matching percentages were
calculated based on Tl’R, poverty, cost, and caseload data that would have
been available in 1993.

7Data for health care costs based on the HCFA wage cost index were only
available for selected years as reported in table 1. We used the corresponding
BLS wage cost index for the years prior to 1988 for calculating matching
percentages based on HCFA wage data.

5                           GAOLHEHS-97-152R       Medicaid   Matching   Formula

Table 2: Current and Alternative Federal Medical Assistance
Percentages (F&al Years 1989-961

Numbers in percent

 Fiscal             Current            New FMAP           New FMAP
 Year                                  (BLS index)        (HCFA index)”
  1996                        50.0               55.2                 59.5
  1995                        50.0               53.7                 57.9
  1994          I             50.0 I             52.41                55.1
  1993          I             50.0 I             51.0 I               52.4
  1992                        50.0               50.5                 50.5
  1991                        50.0               50.0                 50.0
  1990                        50.0               51.7                 51.7
  1989                        50.0               53.2                 53.2
“BLS wage data were used for the years that HCFA data were not available (see
table 1).

If all factors affecting state financing burdens were reflected in the matching
formula, New York would have received higher matching percentages in most
years. Because of the upward trends in poverty, health care costs, and a more
expensive caseload, the matching percentage would have risen from 50 percent
in fiscal year 1991 to 60 percent in 1996, based on the BLS wage cost index. If
matching percentages were calculated on the basis of HCFA hospital wage
surveys, the matching percentage would have increased to nearly 60 percent.

If these higher matching percentages had been used to reimburse the state for
its Medicaid expenses, the state would have received additional federal
assistance as shown in table 3. We estimate that between 1989 and 1996, New
York would have received $3.4 billion in additional assistance in real dollars,

                              GAOIHEHS-97-152R            Medicaid   Matching   Formula
based on using the BLS health industry wage index.8 If wage costs were
measured using the data from the HCFA hospital workers’ wage survey, the
state would have received an additional $6.5 billion.

Table 3: Estimated Increase in New York’s Funding if the Matching Formula
Had Been Designed to Eaualize State Financing Burdens (Fiscal Years 1989-961

Dollars in millions

“I’his figure is negative because the 50-percent match was assumed to apply to
spending for both benefits and administrative costs. States, however, receive
an enhanced match for some administrative expenses that were not reflected in
our calculations.

*We used the medical care component of the personal consumption
expenditures from the National Income and Product Accounts published by the
Department of Commerce, expressed in terms of 1996 purchasing power. The
effect of this adjustment is to inflate spending in past years to reflect the fact
that a dollar purchases fewer health care services today compared with earlier
years when prices were lower.

7                           GAOIHEHS-97-152R      Medicaid   Matching   Formula

Because relative poverty rates have risen substantially in 1994 and 1995 and
because health care costs and more expensive elderly and disabled caseloads
have continued to rise, New York’s matching percentage would also rise if
these trends were taken into account. We estimate New York’s matching
percentage could rise to either 59 or 63 percent in fiscal year 1998, increasing
federal reimbursements for the 2 years by either $4.9 billion (based on BLS
wage costs) or $7.2 billion (based on HCFA hospital wage cost surveys).

Table 4: Estimated Increase in New York’s Funding if the Matching Formula
Had Been Designed to Equalize State Financing Burdens @%xaI Years 1997-98)

Dollars in millions

                   I  FMAP based on BLS wage
                      Increase           Cumulative
                                                           FMAP based on HCFA wage
                                                         I costs
                                                          Increase           Cumulative
  5-r                                    increase                            increase

  1998                       $2,859             $2,859           $4,033            $4,033
     1997                        2,008           4,867               3,203           7,236

If you have any further questions regarding this letter, or if we can be of further
assistance, please call Jerry Fastrup, Assistant Director, at (202) 512-7211 or me
at (202) 512-7114. Dick Horte, Greg Dybalski, and Mark Vinkenes also
contributed to this letter.

Sincerely yours,

 William J. ScanIon
 Director, Health Financing and Systems Issues

 Enclosures - 2

 8                                GAO/HEHS-97-152R        Medicaid    Matching    Formula
ENCLOSURE I                                                        ENCLOSURE I


Because most of the cost of providing health care services is directly or closely
related to the cost of personnel, we constructed a cost index based on wages
paid to health care workers in a state compared with wages paid to such
workers in all states. Specifically, the cost index was calculated on the
assumption that approximately 15 percent of health care costs do not
systematically vary across states (medical supplies purchased in national
markets, for example) and the remaining 85 percent is related to cross-state
differences in personnel costs.

There are two main sources of information on the cost of health care workers:
information on wages paid available through the unemployment insurance
system tabulated by the Bureau of Labor Statistics (BLS) and a survey of wages
paid to hospital workers under the Medicare Prospective Payments System
(PPS) conducted by the Health Care Financing Administration (HCFA). Both
sources of information have certain strengths and weaknesses. Because of
these offsetting strengths and weaknesses, we have not taken a position on
which is the superior indicator of health care costs.

The BLS data we used represent wages paid to health care workers under the
unemployment insurance system. Under this system, employers report both
total wages paid and the number of workers to whom those wages were paid.
The data include wages paid to workers in a broad array of settings, including
offices and clinics of physicians, dentists, and optometrists; nursing and
personal care facilities; hospitals; medical and dental laboratories; and home
health care providers.

The main advantage of the BLS data is their breadth of coverage across work
settings. Its primary disadvantage is that the unemployment system reports
only the number of workers rather than the number of hours worked.
Consequently, a wage index based on these data reflects differences in the mix
of part-time and full-time workers in the state as well as differences purely in
the cost of a unit of labor. For example, if both total wages paid to all workers
and total hours worked were the same in both state A and state B, the state
using more part-time workers would have a lower average wage per worker,
even though wages per hour worked were the same. To avoid this bias, it
would be better to have data on wages paid per hour worked.

The data collected by HCFA under the Medicare PPS represent wages paid to
hospital employees, including nurses, therapists, technicians, and administrative

9                           GAO/HEHS-97-152R       Medicaid   Matching   Formula
ENCLOSURE I                                                       ENCLOSURE I
staff. An advantage of these data is that they include the number of hours
worked per employee. Consequently, a wage index can be constructed that
avoids cross-state differences in part-time versus full-time workers. The
disadvantage of the HCFA data is that they only reflect the wages paid to
personnel who work in a hospital setting rather than wages to personnel
working in a broad array of health care settings. If there are systematic cross-
state differences in wages paid to workers in hospital versus nonhospital
settings, the HCFA wage cost index would not reflect these differences.

An additional issue relating to the use of the HCFA wage index is the time lag
in the availability of data for use in the matching formula. The HCFA data lag
about 2 years behind the BLS wage data. For example, the Medicaid formula
for fiscal year 1998 was calculated in late 1996. At that time, the latest
available BLS wage data would have been for 1995 wages, but the latest HCFA
wage data would have been for 1993 wages.

 10                          GAO/HEHS-97-152R       Medicaid   Matching   Formula
ENCLOSUREII                                                                        ENCLOSURE II


States that serve a higher proportion of elderly and disabled individuals incur a
higher cost of providing services to those in need. To compare cross-state
differences in costs associated with difference in the proportions of recipients
who are more and less expensive to serve, we constructed a caseload cost
index. In developing this index, we first calculated the proportion of each
state’s caseload in each of four eligibility categories: elderly, blind and
disabled, children, and adults and other recipients. We then weighted the
proportion of each state’s caseload by the national average cost per case for
each eligibility group. This weighted average cost was then divided by the
national average cost per case in all eligibility groups to arrive at an index that
measured the extra cost associated with the composition of the state’s actual

Using fiscal year 1995 data, the caseload index was calculated using the
following formula:

 Caseload   = $9265*P65'Shase   + $8535*B&DShare    + $1076*ChildShare    + $1814*0ther   Share
  Index                                            $3405


      $9,265 = Weight applied to the proportion of elderly recipients in a state
               (U.S. average spending per elderly recipient).

P65’ Share = The proportion of the state’s caseload eligible because they are
             aged 65 and over.

      $8,535 = Weight for the proportion of recipients who are blind or disabled
               (U.S. average spending per blind or disabled recipient).

B&D Share = The proportion of the state’s caseload who are blind or disabled.

      $1,076 = Weight for the proportion of recipients that are children
               (U.S. average spending per child recipient).

Child Share = The proportion of the state’s caseload who are children.

11                                GAOLHEHS-97-152R             Medicaid     Matching       Formula
ENCLOSURE II                                                         ENCLOSURE II

      $1,814 = Weight for the proportion of adults and “other” recipients
               (U.S. average spending per adult recipient and those in “other”
               categories of eligibility, mostly those receiving Aid to Families
               With Dependent Children).

Other Share = The proportion of the state’s caseload eligible as adults or in
              other categories.

      $3,405 = The U.S. average Medicaid spending per recipient.


For fiscal year 1995, 12.5 percent of New York’s recipients were elderly; 16.9
percent, blind and disabled; 44.6 percent, children; and 26.1 percent, adults and
those in other eligibility categories. Using these numbers, its caseload index
would be as follows:

    Index                                  $3405

  Caseload    = $1154+$1436+$480+$474
   Index                   $3405

  Caseload    = $3544
   Index        $3405

    cost      = 1.04


 12                           GAO/HEHS-97-152R         Medicaid   Matching   Formula
Ordering     Information

The first copy of each GAO report and testimony is free.
Additional   copies are $2 each. Orders should be sent to the
following  address, accompanied by a check or money order
made out to the Superintendent     of Documents, when
necessary. VISA and Mastercard      credit cards are accepted, also.
Orders for 100 or more copies to be mailed to a single address
are discounted 25 percent.

Orders by mail:

U.S. General Accounting Office
P.O. Box 6015
Gaithersburg, MD 20884-6015

or visit:

Room 1100
700 4th St. NW (corner     of 4th and G Sts. NW)
U.S. General Accounting      Office
Washington, DC

Orders may also be placed by tailing (202) 512-6000
or by using fax number (301) 2534066, or TDD (301) 413-0006.

Each day, GAO issues a list of newly available reports and
testimony.   To receive facsimile copies of the daily list or any
list from the past 30 days, please call (202) 512-6000 using a
touchtone phone. A recorded menu will provide information         on
how to obtain these lists.

For information on how to access GAO reports on the INTERNET,
send an e-mail message with “info” in the body to:


or visit    GAO’s World Wide Web Home Page at:

United States
General Accounting  Office                     BulkRatq
Washington, D.C. 20548-0001              Postage & Fee+ Paid
                                     I     Permit No. GlOO
Official   Business
Penalty    for Private   Use $300

Address    Correction    Reauested