Consumer Price Index: Cost-of-Living Concepts and the Housing and Medical
Care Components (Chapter Report, 08/26/96, GAO/GGD-96-166).

Pursuant to a congressional request, GAO determined: (1) whether changes
made to the housing component of the consumer price index (CPI) made it
more or less suitable as a cost-of-living measure; and (2) the
advantages and disadvantages of changing the current measurement of
medical care costs to a cost of living measurement.

GAO found that: (1) the CPI is not a cost-of-living index, but a measure
of the change in prices paid for a fixed market of goods and services;
(2) a comprehensive cost-of-living index is broader in coverage than an
index based on consumer expenditures and budgets; (3) the Bureau of
Labor Statistics (BLS) uses the rental equivalence method to better
measure housing costs within the CPI structure; (4) this method has made
the CPI more suitable for measuring cost of living; (5) two-thirds of
medical care expenses are excluded from the CPI, since they are paid by
third parties payers; (6) including third-party payments in the CPI
would move the CPI towards a cost-of-living index; (7) BLS excludes
third-party payments from the CPI to better represent direct
expenditures by consumers; (8) changing the medical care component of
CPI would improve the formulation of health-care-specific policies and
macroeconomic policies, but there is little technical feasibility in
making such changes; (9) the Stigler committee believes that CPI should
better reflect the cost-of-living index; (10) it is difficult to design
a cost-of-living index for the federal government because of the
additional uses of CPI; and (11) policymakers need to consider how the
CPI will be affected by changing the medical care component and whether
any single price index can account for such cost-of-living measurements.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  GGD-96-166
     TITLE:  Consumer Price Index: Cost-of-Living Concepts and the 
             Housing and Medical Care Components
      DATE:  08/26/96
   SUBJECT:  Price indexes
             Macroeconomic analysis
             Economic indicators
             Cost of living
             Health care costs
             Rental rates
             Health care cost control
IDENTIFIER:  Consumer Price Index
             Medicaid Program
             Medicare Program
             BLS Consumer Expenditure Survey
             
******************************************************************
** 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.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to the Ranking Minority Member, Committee on Banking and
Financial Services, House of Representatives

August 1996

CONSUMER PRICE INDEX -
COST-OF-LIVING CONCEPTS AND THE
HOUSING AND MEDICAL CARE
COMPONENTS

GAO/GGD-96-166

Consumer Price Index

(243061)


Abbreviations
=============================================================== ABBREV

  BLS - Bureau of Labor Statistics
  CBO - Congressional Budget Office
  CEX - Consumer Expenditure Survey
  CPI - consumer price index
  CPI-U - consumer price index representing all urban consumers
  CPI-W - consumer price index representing all urban wage and
     clerical workers
  FHLBB - Federal Home Loan Bank Board
  FHA - Federal Housing Administration
  GDP - Gross Domestic Product
  HCFA - Health Care Financing Administration
  HMO - health maintenance organization
  ILO - International Labor Organization
  OMB - Office of Management and Budget
  PPI - Producer Price Index
  VA - Department of Veterans Affairs

Letter
=============================================================== LETTER


B-261120

August 26, 1996

The Honorable Henry B.  Gonzalez
Ranking Minority Member
Committee on Banking and Financial Services
House of Representatives

Dear Mr.  Gonzalez: 

This report responds to your request that we use the opinions of
recognized experts to (1) determine if a change made in the early
1980s to the housing component in the Consumer Price Index (CPI) made
the CPI either more or less suitable for use as a cost-of-living
measure and (2) identify the advantages and disadvantages of changing
the current measurement of medical care costs to an approach that
more closely matches a cost-of-living measure.  As you requested, we
did not try to identify and address all of the policy issues that
would be relevant to determining whether the CPI should be moved
further toward a cost-of-living index. 

We are sending copies of this report to the Secretary of Labor, the
Commissioner of Labor Statistics, the Acting Director of the Office
of Management and Budget, and other interested parties.  Copies will
also be made available to others on request. 

The major contributors to this report are listed in appendix VII.  If
you have any questions about this report, please contact me on (202)
512-8676. 

Sincerely yours,

L.  Nye Stevens
Director, Federal Management
 and Workforce Issues


EXECUTIVE SUMMARY
============================================================ Chapter 0


   PURPOSE
---------------------------------------------------------- Chapter 0:1

The Consumer Price Index (CPI) is the measure of price changes that
was used in fiscal year 1995 to adjust for inflation $441 billion in
federal spending and $595 billion of federal tax receipts, thereby
affecting the lives of millions of individuals who received federal
benefit payments and paid federal taxes.  The CPI measures the price
of a fixed market basket of goods and services, organized into major
components, such as transportation and medical care.  Although it is
often referred to as a cost-of- living index, the CPI is not designed
for this purpose.  A comprehensive cost-of-living index does not
exist.  Members of Congress have questioned the use of the current
CPI for adjusting federal benefits and taxes.  One outcome has been
the Senate Committee on Finance's appointment of an expert commission
to study the CPI. 

The Ranking Minority Member of the House Committee on Banking and
Financial Services, concerned that taxpayers may be negatively
affected if the estimation problems of the CPI are not well
understood, asked GAO to (1) determine if a change made to the
housing component in the early 1980s made the CPI either more or less
suitable for use as a cost-of-living measure and (2) identify the
advantages and disadvantages of changing the current measurement of
medical care costs to an approach that more closely matches a
cost-of-living measure.  GAO surveyed recognized experts to obtain
their views on how the change affected the housing component and on
the advantages and disadvantages of changing the medical care
component.  As agreed with the Ranking Minority Member, GAO did not
try to identify and address all of the policy issues that would be
relevant to determining whether the CPI should be moved further
toward a cost-of-living index. 


   BACKGROUND
---------------------------------------------------------- Chapter 0:2

The CPI is a measure of the average change over time in the prices
paid by urban consumers for a fixed "market basket" of goods and
services that people buy for day-to-day living.  However, the CPI
does not measure consumers' actual cost of living.  When consumers
face rising prices, and especially when some prices rise faster than
others, consumers tend to alter their purchasing patterns to maintain
as high a living standard as possible.  Because the CPI holds the
market basket constant and does not account for what consumers would
pay when they change the amounts they buy or substitute one product
for another, it does not measure consumers' cost of living.  In
addition, the CPI does not represent the overall cost of living
because it does not include all goods and services that individuals
consume. 

An index measuring the cost of living is inherently broader than one
focused on consumer expenditures.  In theory, a cost-of-living index
would include everything that contributes to consumer satisfaction,
for example, market goods and services, environmental amenities, and
public goods provided from tax revenues.  However, the components of
an actual cost-of-living index may vary and there is no single,
comprehensive measure of the cost of living. 

In 1961, the Price Statistics Review Committee of the National Bureau
of Economic Research, chaired by George Stigler, recommended
modifying the CPI's conceptual framework to represent a
cost-of-living index because it was in fact being used in the private
sector as a cost-of-living index.  The Stigler committee specifically
recommended a change in the method used to measure homeownership.\1
The Bureau of Labor Statistics (BLS), which produces the CPI, first
changed the measure of the costs of homeownership in 1983.  The
current approach, which estimates the amount of rent that would be
paid for owner-occupied housing were it rented, is known as the
rental equivalence method.  This approach seeks to measure the costs
of consuming housing services over time rather than the value of
housing as an asset that might appreciate over time.  The latter
approach was used before 1983. 

The medical care component of the CPI is based only on out-of- pocket
medical expenses that consumers pay, including health insurance
premiums.  The CPI does not include payments by third- party
payers.\2 Under a comprehensive cost-of-living concept, all medical
care expenses, regardless of the source of payment, would be included
in the CPI because consumers receive benefits from the payments. 


--------------------
\1 Price Statistics Review Committee, The Price Statistics of the
Federal Government:  Review, Appraisal, and Recommendations, A Report
to the Office of Statistical Standards, Bureau of the Budget (New
York:  National Bureau of Economic Research, 1961).  The Stigler
committee was formed under a contract between the Bureau of the
Budget--the predecessor of the Office of Management and Budget--and
the National Bureau of Economic Research. 

\2 Third-party payers can include health insurance, including
employer-provided or subsidized insurance, or government-financed
health care programs, such as Medicaid or part A of Medicare. 


   RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3

According to BLS officials, BLS made the change to the rental
equivalence method to improve the measurement of the cost of
homeownership while adhering to the CPI's conceptual structure of
pricing a fixed market basket of goods and services.  It was not
done, BLS officials said, to move the CPI toward a cost-of-living
index.  However, according to 10 experts on the housing measurement
whom GAO surveyed, the change also had that effect.  The experts said
that the change, because it better measured consumer consumption,
made the CPI more suitable for use as a measure of the cost of
living. 

Nearly all of the experts GAO surveyed agreed that the rental
equivalence method adequately addressed issues that had been raised
about the method that was replaced.  Several experts said that new
additional methodological issues had emerged with the use of the
rental equivalence method. 

Since the 1960s, an increasing portion of medical care costs have
been paid by third-party payers, such as employers and governments. 
Today, about two-thirds of medical care expenses are excluded from
the CPI because third parties directly pay for these costs. 
Including third-party payments in the CPI would move it further
toward being a cost-of-living index.  However, according to BLS
officials, BLS excludes third-party payments from the CPI because BLS
constructs the CPI to represent only direct expenditures by consumers
and because BLS officials do not believe that adding such payments
would make the CPI a clearly better index for its most important
federal uses. 

Of the 10 medical care measurement experts responding to GAO's
structured interview survey, a majority offered advantages and all
identified disadvantages to changing the medical care component to
more closely match a cost-of-living measure.  Advantages cited by the
experts tended to focus on improved policymaking, including
health-care-specific policies and macroeconomic policies. 
Disadvantages they cited included, for example, the technical and
political feasibility of making the changes and the expense
associated with the changes. 

The Stigler committee said that the CPI and its uses should match and
that a cost-of-living based CPI would better match its uses at that
time.  GAO believes that there is a fundamental soundness to the
principle of the index matching its uses.  However, the federal
government uses the CPI in a variety of ways, some of which did not
exist when the Stigler committee did its work.  As a result, it is
increasingly difficult to design an index to match its uses, and it
is unclear whether these uses would all be better served by changing
the medical care component or by some other means of moving toward a
cost-of-living index. 


   GAO'S ANALYSIS
---------------------------------------------------------- Chapter 0:4


      HOMEOWNERSHIP CHANGE MADE
      THE CPI MORE SUITABLE AS A
      COST-OF-LIVING MEASURE
-------------------------------------------------------- Chapter 0:4.1

All of the housing measurement experts GAO surveyed said BLS' change
to the rental equivalence method made the CPI more suitable for use
as a measure of the cost of living.  The experts' comments indicated
that the CPI is more suitable because it measures the cost of housing
services that are used (rental equivalence method) rather than the
cost of buying a house or its value as an asset (asset-price
approach).  Although the change to the rental equivalence method made
the CPI more suitable as a cost-of-living index, BLS officials said
that the intention of the change was not to move the CPI toward a
cost-of-living measure.  Rather, they said the change was intended to
improve the measurement of housing costs. 

All of the experts reported that, in general, the rental equivalence
method adequately addressed the concerns that had been expressed
about the use of the asset-price approach.  For example, the rental
equivalence method measures the value of the use of the house rather
than the change in the investment capital of the house.  One expert
agreed that the concerns with the previous approach were addressed,
but noted that the new method still was inaccurate in its
representation of homeowners with mortgage payments or those with
very low housing costs. 

The experts' responses to whether concerns have emerged as a result
of using the rental equivalence method were mixed--some identified
additional methodological issues, such as difficulties in finding
rental units that match the characteristics of owner- occupied units,
while other experts did not mention such issues.  Of the issues
identified, however, none were mentioned by more than two of the
experts. 


      EXPERTS CITED ADVANTAGES AND
      DISADVANTAGES OF CHANGING
      MEASUREMENT OF MEDICAL CARE
      COMPONENT
-------------------------------------------------------- Chapter 0:4.2

According to Health Care Financing Administration (HCFA) data,
out-of-pocket expenses (which are counted in the CPI) represent less
than one-third of what was spent on medical care in 1991.  Including
in the CPI the approximately two-thirds of medical expenses currently
paid by third parties, such as employers and governments, would be
consistent with a comprehensive cost-of- living concept that would
incorporate all medical care expenses. 

The 10 medical care measurement experts GAO surveyed differed in
their opinions about whether the current measure of medical care
costs should be changed to better approximate a cost-of-living
measure.  A majority of the experts cited advantages to making such a
change.  For example, a few of the experts said that public and
private policymaking would be improved with such a change because
policymakers and researchers would, for example, have a better
understanding of what is happening in medical care costs.  These
experts said a change to a cost-of-living concept would support
implementation of appropriate health care policies and improved
macroeconomic policymaking.  Nonetheless, the experts all raised
disadvantages to making such a change.  For example, a few of the
experts noted that a change in conceptualization would mean a break
in the continuity of the price data, which would affect long-term
trend analyses involving medical care prices.  A few experts also
noted other concerns, such as how to measure medical care using a
cost-of-living basis, the cost of doing so, and the political
acceptability of changes that might result. 

A majority of the experts said that some types of third-party payment
should be added to the medical care component.  However, they did not
agree on which third-party payments to include.  The majority would
include medical care paid for by employer- or union- provided health
insurance, whereas one-half supported the inclusion of
government-provided care.  One expert did not support the addition of
any third-party paid expenses. 

BLS does not plan to include third-party payments in the CPI because
it views the CPI as an index that measures the changes in the prices
of goods and services that consumers purchase directly.  BLS
officials said they make changes to improve the representation of
out-of-pocket expenses rather than to move the CPI conceptually
toward a cost-of-living index.  BLS officials also said that they do
not believe that adding third-party payments would make the CPI a
clearly better index for its most important uses in adjusting Social
Security payments and income tax brackets. 


      UNCLEAR WHETHER MEDICAL CARE
      COMPONENT SHOULD BE CHANGED
-------------------------------------------------------- Chapter 0:4.3

The Stigler committee recommended that the CPI better reflect the
cost of living because of the uses that were being made of it. 
However, the ways in which the CPI is used have increased since that
recommendation was made in 1961.  For example, the federal government
now uses the CPI to adjust some federal benefit payments and the
income tax brackets. 

Because of the additional uses being made of the CPI, whether the
medical care component should be changed to better reflect the cost
of living is unclear.  To make this decision, policymakers would need
to consider several issues GAO did not address, such as (1) how these
uses would be affected if the medical care component were changed to
more fully reflect cost-of-living concepts, (2) whether any single
price index can completely meet all purposes, and (3) the inevitable
choices to be made between the cost of changing the CPI to be more
reflective of the cost of living and the scope and quality of such an
altered index.  While many of these choices would be on technical
issues, some would entail policy judgments. 


   RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:5

GAO is making no recommendations in this report. 


   AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:6

The Office of Management and Budget (OMB) and BLS commented on a
draft of this report.  In a meeting on July 15, 1996, OMB's Chief
Statistician said the draft was a fine report and that it would be
useful in educating laymen and policymakers about the CPI.  In her
written comments of July 11, 1996, the BLS Commissioner focused on
the report's treatment of the medical component.  The Commissioner
said that BLS had excluded employer-provided benefits for a variety
of considerations but had not rejected the cost-of-living concept. 
She said that one theoretically correct, comprehensive measure of the
cost of living does not exist and shaping a medical care component to
include employer-provided benefits would raise formidable measurement
problems.  The Commissioner suggested that a separate index might be
a better way to address different medical care cost policy concerns
and uses than changing the CPI. 

Given the limited scope of its work, GAO has not taken a position on
whether the medical care component should reflect the cost of living
or whether multiple indexes are needed to better fit the uses made of
the CPI. 

The Chief Statistician's and the BLS Commissioner's comments are
discussed further at the end of chapter 4.  A copy of the BLS
Commissioner's written comments is included in appendix VI. 


INTRODUCTION
============================================================ Chapter 1

Changes in prices as measured by the Consumer Price Index (CPI) were
automatically linked to $441 billion in federal spending and $595
billion of federal tax receipts and affected the lives of millions of
individuals who received federal benefit payments and paid federal
taxes in fiscal year 1995.\1 For example, when Congress legislated
the use of the CPI to automatically increase Social Security
payments,\2 it indicated that this indexation was to offset increases
in the cost of living. 

According to BLS, the CPI is not a cost-of-living index but measures
the change in prices of a fixed market basket of goods and services. 
However, the CPI has been used in various ways that are related to
the cost of living.  For example, the CPI is used as an escalator to
adjust income payments, tax brackets, and deductions for personal
exemptions.  Although some elements of the CPI reflect cost-of-living
concepts, the CPI was not designed as a cost-of-living index.  To
date, the federal government has not developed a comprehensive
cost-of-living index. 

The CPI tracks the change in prices of a fixed market basket of goods
and services purchased directly by urban consumers.  These purchases
are for food, clothing, shelter, fuels, transportation,
entertainment, medical services, and other goods and services that
people buy for day-to-day living.  Only expenditures made by
consumers are captured in the CPI. 

The CPI does not attempt to measure all changes in the cost of
consumption needed for an individual to maintain a constant level of
utility, that is, consumer satisfaction.\3 When consumers face rising
prices, and especially when some prices rise faster than others,
consumers tend to alter their purchasing patterns to maintain as high
a living standard as possible.  Because the CPI holds the "market
basket" constant and does not account for what consumers would pay
when they change the amounts they buy, or substitute one product for
another, it does not measure consumers' cost of living.  The CPI,
therefore, is not a cost-of- living index. 

A comprehensive cost-of-living index would be broader in coverage
than an index based on consumer expenditures or consumer budgets.  In
theory, a cost-of-living index would include purchased goods and
services; the use of semidurable and durable goods, such as houses
and automobiles, owned or rented; free goods of nature; and
government-provided goods and services.  However, the components of
an actual cost-of-living index may vary and there is no single,
comprehensive measure of the cost of living. 

Some government-provided goods and services, such as public mass
transit, that charge for the service are included in the CPI. 
However, other items that would be in a comprehensive cost-of- living
index, particularly public and free goods of nature, are excluded
from the CPI because they cannot be readily measured and consumers do
not directly pay for their use. 


--------------------
\1 For additional information about congressional mandated uses of
the CPI, see the report, Statistical Agencies:  Statutory
Requirements Affecting Government Programs (GAO/GGD-96-106, July 17,
1996). 

\2 42 U.S.C.  415(i). 

\3 In economics, the term "utility" is used to denote an individual's
overall satisfaction.  Constant utility is maintaining the same level
of satisfaction from one period to another and can entail
substituting one good or service for another in response to changes
in relative prices.  Utility for economic purposes, however, cannot
be measured.  Nonetheless, it is affected by various factors, such as
one's level of consumption, psychological attitudes, social
pressures, personal experiences, and cultural environment. 


   BACKGROUND
---------------------------------------------------------- Chapter 1:1

The Bureau of Labor Statistics (BLS), within the Department of Labor,
produces the CPI by measuring the average change over time in the
prices paid by urban consumers for a fixed market basket of consumer
goods and services.  The market basket is determined from detailed
records of purchases made by thousands of individuals and families. 
The items selected for the market basket, such as potatoes, are to be
priced each month at retail outlets, such as grocery stores, in urban
areas throughout the country.  According to BLS, in 1995,
approximately 30,000 outlets were visited each month, with prices
collected for 94,000 items. 

The CPI is used as a measure of price changes to make economic
decisions in the private and public sectors.  For example, landlords
use the CPI to adjust rental payments for the effects of inflation. 
According to BLS, the CPI has three major uses:  (1) indicator of
inflation for policymaking and economic decisionmaking; (2) escalator
for wages, income payments, and tax brackets to preserve the
purchasing power of people receiving government transfer payments and
to adjust the tax burden so that people pay in inflation-adjusted
dollars; and (3) deflator of selected economic statistical data
series to make adjustments to show real changes in the data over
time.  For additional information about the uses and construction of
the CPI, see appendix II. 

The CPI was initiated during World War I, when rapid increases in the
prices of goods and services, particularly in shipbuilding centers,
made such an index essential for calculating cost-of- living
adjustments in wages.  In 1921, BLS began regular publication of an
index representing the expenditures of urban wage and clerical
workers, which was then called the Cost-of- Living Index.  The name
of the index was changed to the CPI following controversy during
World War II over the index's validity as a measure of the cost of
living.  According to BLS, it has always been a measure of the
changes in prices for goods and services purchased for family living. 

Major revisions were made to the CPI about every 10 years to update
the fixed market basket; the next major revision is scheduled to be
released in January 1998.  Because people's buying habits changed,
new studies were made of what goods and services people were
purchasing and major revisions of the CPI were made in 1940, 1953,
1964, 1978, and 1987.  In the 1978 major revision, several changes
were made, including the publication of a new index for all urban
consumers--CPI-U.  According to BLS, the CPI-U, which represents the
expenditures of about 80 percent of the population, takes into
account the buying patterns of professional employees, part-time
workers, the self-employed, the unemployed, and retired people, as
well as those previously covered in the CPI.  BLS continued
publication of the original index, the CPI-W, which represents the
expenditures of urban wage and clerical workers, about 32 percent of
the population. 


   CONCEPTUAL CHANGE
   TO THE CPI RECOMMENDED IN 1961
---------------------------------------------------------- Chapter 1:2

In 1961, the Price Statistics Review Committee of the National Bureau
of Economic Research, chaired by George Stigler, identified
conceptual problems with the CPI and addressed issues concerning the
measurement over time of durable goods, such as housing.\4 The
Stigler committee acknowledged that the CPI's original purpose was to
measure average price changes of a fixed market basket of goods and
services over time, which could measure the change in consumers'
standard of living if the marketplace did not change.\5 However,
given that consumers' tastes change over time, or that higher quality
goods at lower prices may become available, the committee determined
that a fixed market basket of goods and services did not
realistically represent a consumer's standard of living.  The Stigler
committee recommended that the conceptual framework of the CPI be
modified to represent a cost- of-living index because the CPI was
being used in the private sector as a cost-of-living measure.\6
Specifically, the committee recommended that the asset-price approach
for measuring homeownership costs be replaced with an approach that
determined the cost of consuming a flow of services generated by
durable goods like houses.  A flow-of-services approach would measure
the cost of consuming housing rather than the change in the
investment value of a house that the asset-price approach measured. 

According to BLS, the Stigler committee's effort was the last
comprehensive review of price indexes.  At the time we were doing our
work, a CPI commission appointed by the Senate Finance Committee was
conducting a study on the CPI's accuracy as a measure of the cost of
living.  The commission issued an interim report in September 1995.\7
According to the interim report, the commission's formation and
charter were motivated by concern that the CPI misstates inflation
and leads to inappropriate changes in federal individual income tax
brackets and federal benefits.  The interim report discusses
categories of potential bias in the CPI, such as substitution and
quality change. 

The commission's interim estimate was that the CPI overstates
inflation by 1.0 percent per year, which fell within a range of 0.7
to 2.0 percent.  The commission expects its final report to include
recommendations for procedures to improve and/or complement the CPI. 
The commission's final report is scheduled to be issued by December
1996. 


--------------------
\4 The Stigler committee was formed under a contract between the
Bureau of the Budget--the predecessor of the Office of Management and
Budget--and the National Bureau of Economic Research.  The scope of
the committee's review was extremely wide and included not only the
study of the main price indexes compiled by the federal government,
but included price indexes that had not been developed.  The CPI was
one of three principal price indexes that the committee reviewed. 
The committee's work is still recognized as the preeminent study of
the CPI. 

\5 Price Statistics Review Committee, The Price Statistics of the
Federal Government:  Review, Appraisal, and Recommendations, A Report
to the Office of Statistical Standards, Bureau of the Budget (New
York:  National Bureau of Economic Research, 1961). 

\6 In 1961, neither Social Security payments nor income tax brackets
were adjusted for inflation.  Automatic adjustments of Social
Security benefits, which are based on increases in the CPI, began in
1975 (42 U.S.C.  415(i)); automatic inflation adjustments of federal
income tax brackets and deductions for personal exemptions were
required by law to begin with the 1985 tax year (26 U.S.C.  1(f)). 

\7 See "Toward a More Accurate Measure of the Cost of Living" Interim
Report to the Senate Finance Committee from the Advisory Commission
to Study the Consumer Price Index, Sept.  15, 1995. 


   OBJECTIVES, SCOPE, AND
   METHODOLOGY
---------------------------------------------------------- Chapter 1:3

The Ranking Minority Member of the House Committee on Banking and
Financial Services asked us to (1) determine if a change made to the
housing component in the early 1980s made the CPI either more or less
suitable for use as a cost-of-living measure and (2) identify the
advantages and disadvantages of changing the current measurement of
medical care costs to an approach that more closely matches a
cost-of-living measure.  We surveyed recognized experts to obtain
their views on how the change affected the housing component and on
the advantages and disadvantages of changing the medical care
component.  As agreed with the requester, we did not try to identify
and address all of the policy issues that might arise in moving the
CPI toward a cost-of- living index. 

We reviewed relevant literature and held discussions with experts in
the field to gain an understanding of the methodologies used in
computing the CPI.  These experts included individuals associated
with the CPI at BLS, as well as private organizations and academic
institutions.  We also obtained information from BLS officials on
their plans to revise and improve the CPI.  On the basis of these
reviews and discussions, we identified the major concerns that were
associated with the asset-price approach, which was used to measure
homeowners' costs before 1983, and the measurement of medical care. 
We recognize that these concerns and issues we identified are not
exhaustive. 

To obtain the views of experts, we selected two panels of experts and
surveyed them.  We chose 10 housing measurement experts from a
candidate list of more than 50 names; we also chose 10 individuals to
serve as medical care measurement experts from a candidate list of
more than 50 names.  To obtain diverse candidate lists, we conducted
a literature review and asked for nominations of potential experts
from those experts in the field and representatives of BLS that we
met with during our initial discussions.  We then contacted the
nominated individuals and asked for their nominations of experts.  To
avoid potential conflicts of interest, we excluded individuals from
the lists who were current political appointees, current BLS
employees, and previous BLS administrators responsible for making CPI
methodological changes.  In selecting the experts, we first selected
those who were nominated more frequently than the others and then
randomly chose thereafter.  We verified that these selections
included experts from academic and user communities, such as the
American Medical Association, and that the selections contained at
least one expert suggested by BLS.  The responses we received reflect
only the views of the experts included.  (See app.  IV for a list of
the selected experts.)

We surveyed 10 housing measurement experts.  The questionnaire we
sent to these experts contained a historical synopsis of housing cost
measurement in the CPI, including brief descriptions of the concerns
that stimulated BLS' adoption of the rental equivalence method in the
1980s and an overview discussion of the rental equivalence method in
terms of measuring the cost of living.  We asked the selected housing
experts if the rental equivalence method adequately addressed the
concerns expressed by critics about the use of the asset-price
approach and if any concerns emerged as a result of using the rental
equivalence method.  In addition, we asked them if the adoption of
the rental equivalence method made the CPI either more or less
suitable for use as a measure of cost of living.  (See app.  V for a
copy of the information and questionnaire sent to each expert.)

We interviewed 10 medical measurement experts.  Before holding the
interviews, we sent a letter to these experts in which we provided a
background paper that identified and briefly described measurement
issues that may result from cost shifting among medical care payers. 
This material was provided to the experts prior to our interviews as
a reference point from which to begin our in-person interviews.  At
the interviews, we asked the experts about the influence of cost
shifting on the medical care component.  We also asked the experts to
provide advantages and disadvantages of changing the current
measurement of medical care costs to an approach that more closely
approximates a cost-of- living measure.  (See app.  V for a copy of
the information sent to and the questions asked of each expert.)

We did our work in Baltimore, MD; Boston, MA; Chicago, IL;
Philadelphia, PA; Richmond, VA; and Washington, D.C., between July
1995 and January 1996 in accordance with generally accepted
government auditing standards. 

We requested comments on a draft of this report from the Acting
Director of OMB and the Secretary of Labor, or their designees.  The
comments are summarized and addressed in chapter 4.  A more detailed
account of our scope and methodology is contained in appendix I. 


METHODOLOGICAL CHANGES TO
HOMEOWNERSHIP MEASURE MADE THE CPI
MORE SUITABLE AS A COST-OF-LIVING
MEASURE
============================================================ Chapter 2

In the 1980s, BLS began using the rental equivalence method of
measuring homeowners' costs.  Recognized experts that we surveyed
viewed the change that BLS made as making the CPI more suitable as a
measure of the cost of living.  They noted, however, that some new
issues have emerged as a result of using the rental equivalence
method.  Although the new method may have made the CPI more suitable
as a cost-of-living measure, BLS officials said this was not their
objective.  Rather, they made the change to better measure housing
costs within the CPI's structure of measuring price changes of a
fixed market basket of goods and services. 


   HISTORICAL DEVELOPMENT OF
   HOUSING MEASURES IN THE CPI
---------------------------------------------------------- Chapter 2:1

Homeownership was not included in the original CPI that was designed
to represent 1917 through 1919 expenditures of wage earner and
clerical worker families in large shipbuilding and industrial
centers, for the main reason that these families typically did not
own homes.  Costs associated with homeownership were first included
in 1953 because homeownership among the urban wage earner and
clerical worker population increased following World War II.  The
homeownership measure in the CPI--the asset- price approach--was
designed to measure changes in the cost of acquiring, financing, and
maintaining houses.  This concept was used from 1953 to 1983.  For
more information on the historical development of the measurement of
housing, see appendix III. 


      BLS ADOPTED THE RENTAL
      EQUIVALENCE METHOD
-------------------------------------------------------- Chapter 2:1.1

Following publication of the Stigler committee report, BLS started in
the 1960s to explore measures of the flow of services received from
owner-occupied homes.  However, because methodologies had not yet
been developed on how to measure these services, BLS had to develop
new methodologies. 

BLS decided to explore two flow-of-services methods:  user cost and
rental equivalence.  BLS staff considered the two methods to be
equally powerful in concept; however, problems raised over
implementation of the user cost method outweighed its attractiveness. 
A user cost index measures total cost to owners living in their homes
by adding the various explicit costs, such as payments for mortgage
interest and property taxes, and implicit costs, such as
depreciation, which homeowners incur in providing shelter for
themselves.  The user cost approach was abandoned after much review
because it involved estimating the appreciation or depreciation value
of a house over time and the cost of not having access to equity in
the house.  BLS noted that calculating such estimates was difficult
because of substantial variations in housing data.  BLS also found
that the method sometimes provided some peculiar results. 

Rental equivalence, on the other hand, was easier to explain to the
public and the users.  The rental equivalence method attempts to
infer the income that homeowners forgo when they reside in their own
homes rather than rent them to others. 

In 1977, when it was time to implement methodological changes for the
1978 major revision to the CPI, BLS' advisory groups had not reached
a consensus on an appropriate flow-of-services approach, so the
asset-price approach was continued.  In addition, some users of the
CPI thought it should reflect the cost of purchasing a home because
most families lived in their own homes and did not rent.  Therefore,
it was not changed. 

BLS, however, decided to continue research and consultation and, in
1980, began publication of experimental indexes that represented
alternative homeownership concepts.  These indexes were variations of
the user cost method and the asset-price approach, as well as the
rental equivalence method.  By the early 1980s, however, changes in
real estate and mortgage markets--high prices of housing and high
interest rates--drew attention to the limitations of the approach
used to measure homeowners' costs.  For a detailed description of
issues associated with the asset- price approach, see appendix III. 
In 1981, the Commissioner of Labor Statistics announced that
beginning in 1983, BLS would use the rental equivalence method to
measure homeownership costs. 

In January 1983, BLS changed the measurement of homeowners' costs in
the CPI-U from an asset-price approach to a flow-of-services
approach.  The CPI-W was not changed until January 1985 because BLS
wanted to provide adequate notice of the conceptual change, since the
CPI-W was used to escalate long-term labor contracts and federal
programs. 


   EXPERTS SAID THAT THE RENTAL
   EQUIVALENCE METHOD MAKES CPI
   MORE SUITABLE AS A
   COST-OF-LIVING MEASURE
---------------------------------------------------------- Chapter 2:2

We asked 10 experts their views on whether the rental equivalence
method made the CPI more or less suitable as a cost-of-living index. 
All 10 were expert in measuring housing costs and were very familiar
with the CPI housing component. 

All of the housing measurement experts agreed that the adoption of
the rental equivalence method made the CPI more suitable for use as a
measure of the cost of living.  Our analysis of the experts' comments
showed that most of the experts responded that the CPI is now more
suitable because it measures the cost of housing services that are
used, rather than the cost of buying the house or its value as an
asset.\1 One of the experts said that the cost-of- living index
"concept is based on consumer utility theory which suggests that
utility comes from consumption, or use.  Since rent is the `price of
using,' rental equivalence enhances the use of the CPI as a proxy for
the cost of living." A few of the experts also noted that additional
improvements, such as including environmental costs and taxes in the
CPI, would make it more like a cost-of-living index.  A few of the
experts noted that the rental equivalence method was not perfect for
measuring the cost of living because renting is not the same as
owning a home. 


--------------------
\1 Our summary descriptions of experts' views are derived from our
compilations and content analysis of their comments and observations. 


      EXPERTS AGREED THAT CONCERNS
      ABOUT THE ASSET-PRICE
      APPROACH WERE ADEQUATELY
      ADDRESSED WITH THE RENTAL
      EQUIVALENCE METHOD
-------------------------------------------------------- Chapter 2:2.1

All of the housing measurement experts reported that in general, the
rental equivalence method adequately addressed the concerns that had
been expressed about the use of the asset-price approach, which had
been used prior to the early 1980s.\2 A few of these experts
commented that in comparison with the asset-price approach, the
rental equivalence method was better in the representation of
inflation and tracking the changes in the cost of occupying a home. 
The rental equivalence method was viewed by a few experts to be more
appropriate than the asset-price approach, especially if the CPI is
to approximate the cost of living.  A few of the experts also noted
that rental equivalence addressed long-standing concerns with the
mortgage and housing price data that were associated with the
asset-price approach (e.g., mortgage interest rates overstating
actual interest expenses).\3 According to a few of the experts,
rental equivalence could represent all sections of the housing
market, as long as rental housing of similar quality is available.  A
few of the experts also commented that the rental equivalence method
in comparison to the user cost index method, which was also proposed
in the 1970s as an alternative methodology, was easier to understand,
more stable, and was not subject to arguable assumptions. 

A few experts mentioned alternative methods to the rental equivalence
method.  These experts' general comments indicated some interest in
using variations of current mortgage payments, down payments, or
mortgage interest to measure homeowners' costs.  However, there was
no consensus on any one variation.  As one of these experts noted,
the "suggestion is not in any way intended to invite a return to the
asset-price approach used until 1982."


--------------------
\2 One expert provided a yes/no response.  This expert stated that
the rental equivalence method "still does not accurately represent
homeowners who are making a mortgage payment or who have paid off
their mortgages and have very low housing costs."

\3 See appendix III for a more detailed explanation of concerns
associated with the asset-price approach to measuring homeowners'
costs. 


      MIXTURE OF ISSUES WITH THE
      RENTAL EQUIVALENCE METHOD
-------------------------------------------------------- Chapter 2:2.2

Overall, the experts' responses to whether issues have emerged as a
result of using the rental equivalence method were mixed.  The
majority of experts said there were issues with the method, but none
of the specific issues was identified by more than two experts. 
Although the experts found the rental equivalence method to be a good
replacement for the asset-price approach, a few of the experts
expressed concern that the rental units used in the methodology may
not be similar to the owner-occupied housing units they are to
represent.  One expert said that this "could lead to errors, but in
both directions."

A few of the experts noted another concern that the CPI overstates
inflation because of the time at which the rental equivalence method
was implemented.  The change to rental equivalence occurred at a
historical peak in mortgage interest rates.  As a result of the
timing of the switch in methodology, the subsequent decline in
mortgage interest rates was not captured in the CPI.  One of the
experts noted that "government transfer payments would be .  .  . 
lower today if the switch had not been made at the interest rate
peak.  Moving to a better index, but at the wrong time, has been
extremely costly."


   BLS VIEWS RENTAL EQUIVALENCE AS
   CONSISTENT WITH FIXED MARKET
   BASKET DEFINITION
---------------------------------------------------------- Chapter 2:3

BLS officials told us that the adoption of the change to the rental
equivalence method was simply a change in the measurement of the
costs of homeownership rather than one intended to move the index
toward a cost-of-living index.  In making changes to the CPI, BLS
said it seeks to improve the presentation of out-of- pocket
expenditures, not to move the CPI conceptually toward a
cost-of-living index. 


ADVANTAGES AND DISADVANTAGES CITED
FOR CHANGING MEASUREMENT OF
MEDICAL CARE
============================================================ Chapter 3

Medical care expenses, including health insurance premiums, directly
paid by consumers have historically been included in the CPI. 
Medical care expenses paid by third parties, such as employers, which
make up about two-thirds of all medical care expenses, are excluded
from the CPI.  A CPI based on comprehensive cost-of-living concepts
would include expenses paid by third parties. 

The 10 medical care measurement experts responding to our structured
interview survey offered various advantages and disadvantages to
changing the medical care component to an approach that more closely
matches a cost-of-living measure.  A majority of the experts said
that some types of third-party expenses should be included in
determining the level of importance BLS would assign to medical care
price changes.  Medical care expenses paid by third parties are
excluded because, according to BLS, the CPI is designed to measure
only out-of-pocket expenses and given that the most important
purposes are probably Social Security and tax bracket indexation, it
is not clear that health insurance fringe benefits should be
included. 

All of the experts we interviewed said that prices that are actually
paid by consumers and third-parties should be used in the CPI.  Since
1987, BLS has been moving toward collecting more transaction prices
for medical care items. 


   DETERMINING MEDICAL CARE COSTS
   IN THE CPI
---------------------------------------------------------- Chapter 3:1

The CPI is constructed from two kinds of data.  One kind is used to
determine what items are to be included in the CPI components and the
relative importance of the components.  The second kind of data
reflects the prices paid for items in the CPI.  (See app.  II for
more information on the construction of the CPI.) These two kinds of
data are used to construct the medical care component; and over the
years, BLS has tried different methods of incorporating these data. 


      DIFFERENCE BETWEEN WEIGHTING
      AND PRICING
-------------------------------------------------------- Chapter 3:1.1

The goods and services that consumers purchase are collectively
referred to as items in the market basket.  These items are grouped
together into components.  For example, hearing aids and dental
services are items in the medical care component.  BLS assigns
weights to items within a component and to the components.  Weighting
is the proportionate emphasis given to price changes of one item or
component in relation to other items and components.  In the medical
care component, weighting is affected by the presence or absence of
third-party payers.  In addition, prices paid by consumers are
collected for the items in the market basket.  For the medical care
component, BLS may collect several different prices for the same item
from medical care providers.  For example, one hospital may have a
list price that is charged a patient who pays the fees directly,
while another hospital reports a discounted transaction price for the
same procedure that has been negotiated with third-party payers. 

To further illustrate how weighting and pricing differ, consider a
hypothetical example of a consumer who receives medical care at a
physician's office.  The consumer pays a $5 insurance co- payment and
the insurance provider pays the physician an additional $12 under a
negotiated fee arrangement.  The combined payment to the physician of
$17 is less than the price that the physician "lists" for the service
provided, $20.  In computing the CPI, such a transaction may have the
following effects: 

  -- The $5 co-payment is the "out-of-pocket" expenditure, which is
     used to set the amount of the weight.  To determine the share of
     consumer expenditures spent on medical care or other components,
     BLS added that $5 together with all other consumption
     expenditures in the Consumer Expenditure Survey (CEX), which
     includes the consumer's cost of premiums paid for health
     insurance.  This aggregate of medical expenses is compared with
     the aggregate of expenditures on all goods and services in the
     market basket.  The percentage of medical care expenses in
     relation to all expenditures becomes the expenditure weight
     assigned to the medical care component. 

  -- In its pricing surveys, BLS would price the medical service at
     $20, the list price.  The transaction price would be $17 in this
     hypothetical situation, assuming that this payment method were
     selected for pricing. 


      BLS TRIED DIFFERENT METHODS
      TO PRICE MEDICAL CARE COSTS
-------------------------------------------------------- Chapter 3:1.2

Medical care has always been in the CPI, and consumer-purchased
health insurance has always been included as a medical expense.  Over
the years, BLS tried different methods of pricing medical care,
including health insurance directly paid by consumers, but has not
altered the method used to determine the weight of the medical care
component. 

Originally, the price change rate for health insurance was assumed to
be equal to the average of other medical items.  In the 1950
revision, BLS deviated from this approach and began to directly price
health insurance policies.  In 1964, the approach was changed to an
approach that again based health insurance price changes on prices
observed for other medical goods and services, as well as the
insurance carriers' operating costs and profits.  In 1978 and 1987,
BLS made minor adjustments in measuring health insurance with most of
the changes occurring in publication of health insurance price
changes. 

In 1987, BLS began to collect medical care transaction prices, actual
prices paid, which include fees that have been negotiated between
medical care providers and third-party payers.  BLS plans to collect
transaction prices for all medical care items by January 1997.  (See
app.  III for a detailed historical description of the measurement of
medical care items.)


   TRENDS IN MEDICAL CARE
   EXPENDITURES
---------------------------------------------------------- Chapter 3:2

In 1965, households directly paid for about two-thirds of all medical
care expenses.  About 30 years later, these medical care expenses,
upon which the CPI is based, represented less than one-third of all
medical care expenses.  As shown in figure 3.1, the proportion of
total medical care represented in the CPI steadily declined since
1965. 

   Figure 3.1:  Percentage
   Distribution of Medical Care
   Expenditures by Type of Payer,
   for Selected Years

   (See figure in printed
   edition.)

\a A small proportion of medical care expenses is paid by
philanthropic sources and income received by medical care
institutions from assets, such as interest, dividends, and rents. 

Source:  Health Care Financing Administration (HCFA). 

The most recent data available, for 1991, indicate that about 28
percent of total consumption of medical care is represented in the
CPI.\1 Because the medical care component is based only on
out-of-pocket medical expenses and health insurance premiums reported
in the CEX, not all medical care expenditures are included in the CPI
(see fig.  3.2).  The CPI does not include medical care that
consumers receive through employer-provided benefits and
government-provided health care programs, such as Medicaid and part A
of Medicare.  Although employees and the self-employed make
contributions to the Medicare hospital insurance trust fund, BLS
considers these contributions as employment taxes and thereby
excludes expenses paid under these programs.\2

   Figure 3.2:  Percentage
   Distribution of Medical Care
   Expenditures by Type of Payer,
   1991

   (See figure in printed
   edition.)

Source:  BLS. 


--------------------
\1 The 28 percent includes out-of-pocket expenses directly paid by
households, employees' portion of employer-provided health insurance
premiums, Medicare part B premiums, and health insurance premiums
directly paid by households.  In the background paper we sent to the
medical care measurement experts, we incorrectly stated that
out-of-pocket medical expenses and health insurance premiums in the
CEX represent about 20 percent of total consumption of medical care. 

\2 Employers pay medical care expenses through their contributions to
Medicare, workers' compensation, temporary disability insurance, and
industrial in-plant health services. 


   RESEARCHERS' VIEW OF THE CPI'S
   MEASUREMENT OF MEDICAL CARE
---------------------------------------------------------- Chapter 3:3

Like other aspects of research on the CPI and cost-of-living indexes,
research on the medical care component has sought to determine the
appropriate weights and prices for medical care.  Research into the
issue of cost shifting between third parties and consumers has drawn
attention to the impact that the inclusion of third-party payments
may have on the weight given to the medical care component of the
CPI.  Similarly, research findings on medical care providers using
multiple prices for the same medical care service have led to an
examination of the CPI's use of list prices rather than transaction
prices. 


      CONCERNS ABOUT MEDICAL CARE
      WEIGHT
-------------------------------------------------------- Chapter 3:3.1

The appropriateness of the weight assigned for medical care in the
CPI has been questioned by some researchers who contend that the
current weight distorts price changes that result from cost shifting
among health care payers.  They contend that the current weight based
on expenditures directly paid by households can result in an
inaccurate level of importance being assigned to this component of
the CPI, as compared with a weight based on all medical expenditures. 
The inaccuracy can occur when costs shift between payers who are
included in the CPI and those who are excluded.  For example, if
employers decrease the amount that they pay of their employees'
medical care expenses, then the employees' direct expenses, which are
used to set the CPI medical care component weight, increase thereby
increasing the weight assigned to medical care.  An inappropriate
weight of a component in the CPI can lead to over- or understatement
of the rate of inflation, if the rate of price change for that
component differs from other components in the CPI.\3 A weight based
on all medical care expenses, however, would not be affected by cost
shifting over time between payers because all costs, regardless of
who paid for the care, would be represented in the CPI. 

Cost shifting over time can be illustrated by employers' efforts to
constrain increases in annual health insurance premiums by raising
deductibles and shifting more of the premium costs to employees.  BLS
reported that the proportion of families paying all or part of their
health insurance premiums has increased from 60 percent in 1984 to 67
percent in 1992.\4 BLS reported that between 1984 and 1992 average
household out-of-pocket medical care expenditures rose from $1,049 to
$1,634, a rise of about 7 percent per year.  These data suggest that
the weight for medical care, which is based on 1982 through 1984 CEX
data, is lower than one that would be derived from more recent
out-of-pocket medical expenses. 

A comprehensive cost-of-living index would include out-of-pocket,
government-provided, and employer-provided medical care costs to
incorporate all medical care expenses.  In a 1994 study, the
Congressional Budget Office (CBO) noted that, if the measurement of
all medical care expenses were more appropriate for measuring the
cost of living, the current CPI would have a downward bias relative
to the cost of living.\5 Using data from the National Income and
Product Accounts, CBO estimated in 1994 that the influence of medical
care price changes on the CPI would more than double if all medical
care expenses were incorporated. 


--------------------
\3 The change to the rental equivalence method of measuring
homeowners' costs changed its relative importance from 25 percent to
14 percent of all items in the CPI.  As explained in appendix II, a
change in relative importance of an item affects that item's
influence on the overall CPI.  The decrease in the relative
importance of homeownership in this instance corrected for
overstatement of the rate of inflation. 

\4 Bureau of Labor Statistics, "Health Insurance Premiums Dominate
Health Care Budget of Consumers," Issues in Labor Statistics
(Washington, D.C.:  U.S.  Government Printing Office, 1994). 

\5 John F.  Peterson, "Is the Growth of the CPI a Biased Measure of
Changes in the Cost of Living?" CBO Papers (Washington, D.C.:  CBO,
1994) pp.  22-23. 


      CONCERNS ABOUT HOW MEDICAL
      CARE EXPENSES ARE PRICED
-------------------------------------------------------- Chapter 3:3.2

In addition, researchers have noted the divergence in medical care
price indexes when one group of payers subsidizes another.  Cost
shifting within the marketplace also occurs when third-party payers
are charged prices that differ from prices paid by consumers.  BLS
does not use transaction prices for all medical care goods and
services included in the CPI; it has announced plans to do so.  Any
collection and use of inappropriate medical care prices could lead to
over- or understatement of the rate of inflation, if the rate of
change for transaction prices differs from the prices used in the
CPI.  (See app.  II for a description of the collection of medical
care prices for the CPI.)

In its 1994 study, CBO noted that a CPI based on out-of-pocket
medical costs fails to capture price distortions caused by cost
shifting.\6 In this case, cost shifting may occur when the government
does not reimburse health care providers for full cost of services to
Medicare patients, and providers try to recoup the difference by
increasing the costs to their private- pay patients (e.g., those
paying for services themselves).  CBO found that Medicare
reimbursement in the early 1990s paid for 88 percent of the costs of
covered services, compared with full reimbursement during the
mid-1980s.\7 CBO observed that the CPI for out-of- pocket medical
care costs increased faster than the price index for Medicare during
the mid-1980s to early 1990s because of cost shifting.\8 The study
suggests that the CPI overstates the rate of inflation because it
fails to capture the price paid by the federal government. 

However, HCFA disputes CBO's findings on cost shifting.  According to
an HCFA official, any findings that Medicare or Medicaid pays less
per day of hospital care than others is not evidence that these
programs are shifting their costs elsewhere. 

The medical care measurement experts we surveyed indicated in their
general comments that this issue is not resolved.  A few of the
experts said that cost shifting does not occur between
government-provided programs and other payers and cited research on
cost shifting in Illinois\9 and California\10 hospitals.  The study
of Illinois hospitals supported CBO's findings that hospitals offset
most of the rise in unreimbursed Medicare costs during the 1980s by
generating higher revenues from private payers, which were cited in
the background paper sent to the medical care experts.  The author,
however, suggests that as private sector pricing becomes more
competitive, the ability and willingness of hospitals to cost shift
will decline.  The study of California hospitals found no cost
shifting from publicly funded patients to privately insured patients. 


--------------------
\6 Peterson, pp.  22-23. 

\7 Congressional Budget Office, "Responses to Uncompensated Care and
Public-Program Controls on Spending:  Do Hospitals `Cost Shift'?" CBO
Papers (Washington, D.C.:  Congressional Budget Office, 1993). 

\8 Peterson, pp.  22-23. 

\9 David Dranove, "Pricing by Non-Profit Institutions:  The Case of
Hospital Cost-Shifting," Journal of Health Economics, Vol.  VII
(1988), pp.  47-57. 

\10 David Dranove and William D.  White, "Government Cutbacks in
Hospital Reimbursement:  Who Bears the Burden?" (Working paper under
revision for journal review July 1995.)


   EXPERTS' VIEWS ON CHANGING
   MEDICAL CARE MEASUREMENT WERE
   MIXED
---------------------------------------------------------- Chapter 3:4

We asked 10 experts for their views on the advantages and
disadvantages of changing the current measurement of medical care
costs to more closely match a cost-of-living measure.  We also asked
about the types of medical care expenses that should be included in
the weighting of the medical care component, as well as the types of
prices that should be incorporated in calculating the changes in
medical care costs. 


      ADVANTAGES AND DISADVANTAGES
      CITED BY EXPERTS OF MOVING
      TO A COST-OF-LIVING CONCEPT
-------------------------------------------------------- Chapter 3:4.1

All but one of the experts cited advantages to changing the current
measurement of medical care to an approach that more closely
approximates a cost-of-living measure.  Our analysis of the experts'
responses showed that a few of the experts said that policymaking
would be improved with such a change.\11 These experts said that a
change to a cost-of-living concept could support the implementation
of appropriate health care policies.  For example, one expert cited a
need for accurate information during a debate on pharmaceutical
drugs.  A few experts also cited each of the following advantages: 

  -- A change would improve macroeconomic policymaking; one expert
     noted that the Federal Reserve was currently guessing at the
     amount of overstatement of inflation in the CPI. 

  -- A change to cost-of-living concepts for the medical care
     component would allow private and public policymakers and
     researchers to have a better understanding of what is happening
     in medical care costs. 

  -- The change would improve the Gross Domestic Product's (GDP)
     implicit price deflator thereby improving research that used the
     implicit price deflator.\12

All of the experts cited at least one disadvantage to changing the
medical care component to more closely approximate a cost-of- living
measure.  A few of the experts were concerned about the measurement
of utility (as previously defined in footnote 3, p.  10) in the
medical area.  For example, one expert questioned how one would
measure a patient's satisfaction from a procedure that had a very
high mortality rate but also offered, when successful, a long-term
survival rate.  In addition, a few experts noted the following
disadvantages: 

  -- A change would mean a break in the continuity of the price data,
     which would affect long-term trend analyses. 

  -- Measures based on cost-of-living concepts were susceptible to
     manipulation because of the subjectivity of measuring
     satisfaction. 

  -- It would be expensive to switch to the new methodology, and it
     would also be more expensive than the current methodology to
     maintain.  For example, one expert noted that transaction price
     data would be regarded by health care providers as sensitive
     information and burdensome to provide to BLS.  More
     specifically, this expert said that physicians would have to go
     through each of their third-party contracts to obtain this
     information. 

  -- In general, the political environment is not conducive to making
     a change in medical care indexes.  For example, one expert noted
     that recognition of a previous overstatement in the CPI would
     anger those whose benefits are indexed with the CPI. 

In addition, a few of the experts were concerned about changing the
medical care component without changing other components at the same
time to more closely approximate a cost-of-living measure. 


--------------------
\11 Our summary descriptions of experts' views are derived from our
compilation and content analysis of their comments and observations. 
The 10 experts that we interviewed all had different reasons for
their opinions on changing the medical care component to be more
reflective of the cost of living.  Because the format used to record
their reasons was open-ended, the elaborations given by one expert
did not necessarily follow those expressed by another expert. 
Therefore, the information we present is a description of what they
said, as opposed to a tally of specific reasons.  See appendix V for
a copy of the instrument used to collect this information. 

\12 The implicit price deflator is a measure of inflation that is
obtained from GDP data.  The movement of the implicit price deflator
usually closely parallels the movement of the CPI but is rarely
identical to it.  In theory, the implicit price deflator reflects the
price trends throughout the economy, whereas the CPI represents price
trends at the retail level. 


      MOST EXPERTS SAID ADDITIONAL
      EXPENSES SHOULD BE INCLUDED
      IN THE WEIGHTING OF MEDICAL
      CARE
-------------------------------------------------------- Chapter 3:4.2

In addition to obtaining the experts' opinions on the advantages and
disadvantages of changing the current measurement of medical care
costs, we took the opportunity to ask the experts how cost-of-living
concepts would be implemented.  All but one of the experts said that
some types of medical expenses other than those already captured in
the CPI should be included in weighting the medical care component. 
The majority of the experts said that employer-provided and
union-provided medical care should be included in the weighting of
the medical care component.  One- half of the experts supported the
inclusion of government- provided care.  Fewer experts supported the
inclusion of medical expenses provided by charitable organizations
and expenses absorbed by health care providers.  One expert said that
no additional expenses should be included in the weighting of the
medical care component.  This expert said that if the CPI is used to
adjust wages or payments, then only out-of-pocket expenses paid by
consumers should be included in the CPI. 

Our analysis of the experts' responses showed that a few of the
medical care measurement experts who supported the inclusion of all
medical care expenses commented that it was logical to include all
expenses if the burden of payment fell upon the general population. 
Of the experts who supported the inclusion of additional expenses
other than those provided by charities or the government, a few said
that expenses that affect the buying power of consumers should be
included in the medical care component. 

There was no consensus on how to implement weighting that is based on
cost-of-living concepts.  The experts' observations on whether
public-provided health care should be included illustrate both the
diverse and occasionally contradictory comments of the experts.  A
few of the experts did not want such care included because other
government-provided services (e.g., national defense) were not in the
CPI.  Other experts expressly told us that government-provided care
should be included because consumers pay for this care through taxes
and lower wages.  These experts also supported the inclusion of taxes
in the CPI.  And still other experts were silent on this issue.  A
few of the experts expressed the opinion that the medical care
component should not be changed to measure cost-of-living concepts
unless all components were changed at the same time.\13


--------------------
\13 Although we did not inquire further into their reasons during the
interviews, these experts may have been referring to the issue of
including taxes in the CPI when referring to changing all components
at one time.  The Stigler committee recognized the difficulty in
including government services and taxes in the CPI.  The committee
said that more research was needed before a more comprehensive
cost-of-living index that included government services could be
constructed and did not recommend making any changes at that time. 
This issue has not been addressed nor has there been an evaluation of
the CPI in the context of its uses to adjust federal benefit payments
and income tax brackets. 


      EXPERTS AGREE THAT
      TRANSACTION PRICES SHOULD BE
      USED
-------------------------------------------------------- Chapter 3:4.3

All of the medical care measurement experts said that transaction
prices should be used in gathering price data for medical goods and
services.  A few of the experts observed that consumers are paying
transaction prices and that list prices should only be used in
instances when nothing else is available, or if list prices are
cheaper to collect.  A similar number advocated the pricing of
comprehensive health care packages, such as basic health maintenance
organizations' (HMO) plans.  However, one expert advocated using list
prices in geographic areas where HMOs had not penetrated the market. 
A few of the experts made the following additional comments: 

  -- List prices could be transaction prices in some instances. 

  -- Transaction prices are available for data collection. 

  -- BLS was not recording the appropriate transaction prices. 

  -- List prices usage in the CPI has led to the overstatement of
     inflation, especially in pharmaceutical drugs. 

  -- BLS' pricing of medical care items was inappropriate.  These
     experts told us that BLS should be pricing the cost of a
     treatment or cure of an illness.  They said that the current
     approach of pricing the cost of medical care items, such as
     x-rays, doctor visits, diagnostic tests, and hospital stays, is
     inappropriate for today's CPI. 

  -- BLS' methodologies are not capturing the substitution of new
     treatments for items in the CPI's medical care market basket. 
     For example, one expert said BLS should be pricing the cost of
     treating medical conditions, such as heart attacks, rather than
     hospital stays.  This expert stated that BLS' methodology used
     for the CPI indicates that the per day charges for hospital
     stays are going up for heart attack patients, when in reality
     new treatments allow patients to go home earlier.  According to
     this expert, by incorrectly pricing hospital stays, the current
     BLS methodology results in overstatement of inflation. 

In addition, a few of the experts also provided an example of the
difference in rate of price change between transaction and list
prices.  They noted the lower rate of increase for hospital rooms in
the Producer Price Index (PPI), which uses transaction prices.  The
difference between the CPI and the PPI for physicians fees that both
use transaction prices, however, is not as large as that for hospital
rooms. 

BLS began collecting and using medical care transaction prices in
1987.  Since then, it has expanded the collection of transaction
prices for additional medical care items.  It plans to collect
transaction prices for all medical items by January 1997.  (See app. 
III for further details.)

According to BLS, the incorporation of discounted transaction prices
was accomplished through a series of improvements in detailed data
collection procedures; therefore, the experts were unlikely to know
that BLS had already begun to incorporate hospital transaction prices
in the CPI.  According to a BLS official, about 15 percent of
hospital prices in the CPI are transaction prices. 

While BLS officials considered the capturing of medical transaction
prices to be an improvement, they said many problems remain in
measuring price change in medical care.  They agreed with our experts
who noted that measuring specific commodities and services used in
medical treatments does not capture changes in the approaches for
treating specific medical problems.  However, according to the
officials, every treatment is administered, not just to a medical
condition but to an individual with that condition.  Therefore,
according to these officials, the treatment administered in different
cases with a given condition, such as a heart attack, need not be the
same.  Using the same example of hospital stays cited by one of the
experts, BLS said that some patients will require shorter or longer
hospital stays, or different combinations of drugs or surgical
procedures, which further complicates defining what is to be priced
and calculation of expenditure weights.  According to BLS, in some
cases a shorter hospital stay might not be better (if, for example,
the patient were weaker and at greater risk for complications when he
or she left the hospital), while in other cases it might be better. 


      OTHER COMMENTS
-------------------------------------------------------- Chapter 3:4.4

When asked if they would like to make additional comments, the
medical care measurement experts identified several issues related to
changing the approach used to measure medical care.  In their general
comments, a few of the experts said that the United States measures
medical care better than other developed countries.  One expert said
that if the United States changes its approach to measuring medical
care, the other countries are likely to change their methodologies to
whatever the United States does.  Other comments stated by a few of
the experts included

  -- The CPI cannot be used for their work because the CPI uses list
     prices. 

  -- They were concerned about how a change to cost-of-living
     concepts would measure quality changes and noted that the
     implementation of these concepts would involve value judgments. 

  -- The distinction between prices and quantities had to be clear,
     implying that an increase in total expenditures cannot be easily
     translated into increases in prices or increases in quantities
     without the collection of additional data. 

Also, a few experts questioned whether the CPI should be used as a
measure by which to make cost-of-living adjustments.  These experts
suggested that BLS develop CPIs for specific demographic groups. 


   BLS DOES NOT PLAN TO INCLUDE
   THIRD-PARTY PAYMENTS
---------------------------------------------------------- Chapter 3:5

BLS does not plan to include third-party payments in the medical care
component.  BLS officials hold this position for several reasons. 

  -- BLS views the CPI as an index that measures the changes in
     prices of goods and services that consumers purchase directly--
     the fixed market basket.  Therefore, the CPI excludes payments
     made by private third parties.  The changes to the CPI that BLS
     seeks to make are to improve the representation of out-of-
     pocket expenditures, not to move the CPI conceptually toward a
     cost-of-living index. 

  -- BLS considers medical care provided through employment as a cost
     of doing business rather than a consumer expenditure. 

  -- BLS excludes income taxes, which pay for government-provided
     health care, because they are indirect payment for medical care. 
     The CPI only includes taxes that are paid as a result of
     consumption, such as sales taxes. 

  -- According to the Commissioner of Labor Statistics, methodologies
     have "not advanced to the point where anyone knows how to
     construct true cost-of-living measures" for medical care and
     other CPI components.\14

  -- According to BLS, the most important purposes for use of the CPI
     are probably the indexation of Social Security payments and
     federal income tax brackets.  BLS says it is not clear that
     health insurance fringe benefits should be included in the CPI
     because these benefits are not taxed. 


--------------------
\14 Consumer Price Index:  Hearings Before the Committee on Finance,
United States Senate.  104th Cong., 1st Sess., p.  108 (1995)
(statement by Katharine G.  Abraham, Commissioner, BLS, Department of
Labor). 


OBSERVATIONS
============================================================ Chapter 4

According to BLS, the CPI is not a cost-of-living index but a measure
of the change in prices of a fixed market basket of goods and
services.  But questions have surfaced from time to time as to
whether the CPI could and should be made into a cost-of-living index. 
The Stigler committee's landmark study in 1961 said the CPI should be
changed to better reflect the cost of living because of the uses that
were being made of it at that time.  However, additional uses have
been made of the CPI since 1961, most notably indexing Social
Security benefits and individual income tax brackets and deductions
for personal exemptions. 

Since the Stigler committee's report, BLS changed the way in which
the CPI measures homeownership.  It went from an asset approach to a
rental equivalency method.  For our review, we asked 10 housing
measurement experts whether the change made the CPI more or less
suitable for use as a cost-of-living measure.  They all said it made
the CPI more suitable as a cost-of-living index. 

BLS has said that it did not make the change to move the CPI closer
to a cost-of-living index, whether or not it had that effect. 
According to BLS, it made the change to improve the presentation of
consumers' out-of-pocket expenditures, which was in keeping with the
concept it follows to construct the CPI. 

BLS likely would be unable to remain faithful to that concept if it
were to make the medical care component truly reflective of the cost
of living.  The medical care component is not reflective in large
measure because it excludes payments made by third parties for
medical care that consumers receive.  BLS is opposed to adding
third-party payments to the CPI because the payments do not reflect
what consumers spend directly and because BLS officials do not
believe that adding such payments would make the CPI a clearly better
index for its most important uses. 

We discussed with 10 medical care measurement experts the advantages
and disadvantages to changing the medical care component to more
closely match a cost-of-living measure.  A majority of the experts
offered advantages and all identified disadvantages to making such a
change.  Also, we discussed with the experts the question of what
types of medical care expenses the CPI should include in determining
the weight of the medical care component.  Their answers were not
unanimous, and cautionary statements were made.  Most experts would
include some type of third-party payment, but there was no consensus
on how to implement weighting that is based upon cost-of-living
concepts.  A few questioned whether a single CPI should be used as a
measure by which to make cost-of-living adjustments. 

The overall impact of changing the medical care component of the CPI
is unknown.  In terms of the weighting of the component, a 1994 CBO
study suggests that the present system leads to an understatement of
the rate of inflation.  Regarding the pricing, a few of the medical
care measurement experts we interviewed stated that the use of list
prices leads to an overstatement of the rate of inflation.  Taken
together, the overall magnitude or direction of a possible
misstatement from the current weighting and pricing of medical care
items is unknown. 

Taking into account the views of our experts and the scope of our
work, we do not have a view as to whether the medical care component
should be changed to reflect the cost of living.  The Stigler
committee held that the CPI and the uses made of it should match. 
Although BLS cannot control the uses made of the CPI, we believe
there is a fundamental soundness to the principle of the index
matching its uses.  However, the federal government uses the CPI in a
variety of ways today, some of which did not exist when the Stigler
committee did its work.  Because the relationship between the current
CPI and these uses has not been assessed, it is not clear whether the
current CPI, a CPI based on cost-of-living concepts, or even multiple
new indexes would best meet all of the purposes for which the CPI is
now used.  Further, there would be inevitable technical and policy
choices to be made in any effort to change the CPI.  These choices
would reflect on the cost, scope, and quality of such an altered
index.  Because these issues were outside the scope of our review, we
are not making recommendations on whether BLS should work toward
making the CPI a comprehensive cost-of-living index. 


   AGENCY COMMENTS AND OUR
   EVALUATION
---------------------------------------------------------- Chapter 4:1

OMB and BLS commented on a draft of this report.  At a July 15, 1996,
meeting, OMB's Chief Statistician characterized the draft as a fine
report and said it had an educational quality that would make it
useful for laymen and policymakers.  She and her staff identified
several places where a technical change could be appropriate or the
wording of the report could be improved, and we made these
alterations where appropriate. 

The BLS Commissioner focused her comments on the medical care
component.  Appendix VI contains a copy of the Commissioner's July
11, 1996, letter and our additional comments.  The Commissioner said
the draft report asserted that incorporating expenditures on medical
care goods and services by third-party payers would move the CPI
toward "the cost-of-living concept." According to the Commissioner,
this argument implies that there exists one theoretically correct,
comprehensive measure of the cost of living and that the CPI deviates
from this measure because it lacks a cost-of-living concept as a
measurement objective.  Neither assumption is strictly correct, the
Commissioner said. 

Elaborating on this statement, the Commissioner presented information
to indicate that different index concepts are required to address
different policy concerns and uses, implying that it is infeasible to
change the CPI to conform with every possible use.  She said
developing a separate index measure might be a better way to address
the concerns with tracking medical care costs than changing the CPI. 
The Commissioner also identified conceptual and operational
difficulties, some of which she termed formidable or impossible to
overcome, that she associated with developing a comprehensive
cost-of-living measure.  Finally, concerning the exclusion of
employer-provided benefits from the medical care component, the
Commissioner said BLS' decision to exclude those benefits reflected a
variety of considerations about the scope and use of the CPI but not
a rejection of the cost-of-living concept. 

We did not intend to suggest that there was a single, correct, and
comprehensive measure of the cost of living or that there could only
be one measure.  Accordingly, we made this position clearer in the
executive summary and in chapter 1. 

As we said earlier in this chapter, because of the limitations of our
scope, we have not taken a view as to whether the medical care
component should be changed to reflect the cost of living or whether
multiple indexes should be developed.  We agree with BLS' contention
that designing a cost-of-living index is not an easy task.  However,
to the extent the government uses the CPI for significant purposes as
if it were a cost-of-living index, we believe there is fundamental
soundness to the principle of an index matching its purposes. 


OBJECTIVES, SCOPE, AND METHODOLOGY
=========================================================== Appendix I

The Ranking Minority Member of the House Committee on Banking and
Financial Services asked us to (1) determine if a change made to the
housing component in the early 1980s made the CPI either more or less
suitable for use as a cost-of-living measure and (2) identify the
advantages and disadvantages of changing the current measurement of
medical care costs to an approach that more closely matches a
cost-of-living measure.  We surveyed recognized experts to obtain
their views on how the change affected the housing component and on
the advantages and disadvantages of changing the medical care
component.  As agreed with the requester, we did not try to identify
and address all of the policy issues that would be relevant to
determining whether the CPI should be moved further toward a
cost-of-living index. 

We first reviewed relevant literature and held discussions with
experts in the field.  These experts included individuals associated
with the CPI at the Bureau of Labor Statistics (BLS), as well as
private organizations.  We also included individuals in congressional
and other federal government agencies who are researchers on the
topic of the CPI; these included several former BLS officials. 

To obtain the views of experts, we selected two panels of experts and
surveyed them.  The first panel, consisting of 10 housing measurement
experts, was provided a data collection instrument that included a
historical synopsis of BLS' housing costs measurement for the CPI. 
This instrument also included brief descriptions of the concerns that
stimulated BLS' adoption of the rental equivalence method in the
1980s, an overview discussion of the rental equivalence method in
terms of measuring the cost of living, and a set of questions about
housing costs measurement methodologies. 

We asked the selected housing measurement experts if the rental
equivalence method adequately addressed the concerns expressed by
critics about the use of the asset-price approach and if any concerns
emerged as a result of using the rental equivalence method.  We also
asked them if the adoption of the rental equivalence method made the
CPI more suitable for use as a measure of cost of living.  They were
asked to explain their answers. 

The second panel, consisting of 10 medical care measurement experts,
was provided the background paper that identified and briefly
described issues that several researchers view as creating
measurement limitations for the CPI as a result of cost shifting
among medical care payers.  This material was provided to these
experts prior to our interviews as a reference point from which to
begin the interviews. 

In the interviews, we asked the selected medical experts whether the
influence of cost shifting on medical care pricing and associated
weighting of the medical care component are measurement problems or
limitations that need to be addressed.  We inquired about the types
of health care expenditures that they thought should be included in
the weighting of the medical care component of the CPI, as well as
the types of prices that should be incorporated in calculating
changes in medical care costs.  We also asked the experts to provide
advantages and disadvantages of changing the current measurement of
medical care costs to an approach that more closely approximates a
cost-of-living measure. 

We used the following process to identify and select two panels of
experts.  The 10 housing measurement experts were chosen from a
candidate list of more than 50 names.  The 10 individuals who served
as medical care measurement experts were selected from a candidate
list of more than 50 names.  To obtain diverse candidate lists, we
conducted a literature review and contacted experts identified
through this review.  We asked for nominations of potential experts
from the recognized experts and representatives of BLS that we met
with during our initial discussions.  We then contacted the nominated
individuals and asked for their recommendations of experts.  To avoid
potential conflicts of interest, we excluded individuals from the
lists who were current political appointees, current BLS employees,
and previous BLS administrators responsible for making CPI
methodological changes.  In selecting the experts, we first selected
those who were nominated most frequently and then randomly chose from
those with equal frequency of nomination.  We verified that these
selections included experts from academic and user communities and
that the selections contained at least one of the experts suggested
by BLS.  We recognize that the responses we received reflect only the
views of the experts included. 

Several of the experts initially selected were unable or unwilling to
participate.  We replaced these individuals with alternates from the
remaining respective candidates lists. 

We provided the selected housing measurement experts a package
containing a letter of introduction, an instruction sheet, and
response sheets.  The packages were sent on September 27, October 18,
and October 24, 1995.  We received responses from all 10 housing
experts by November 21, 1995. 

We provided the selected medical care measurement experts a package
containing a letter of introduction and background information.  The
packages were sent between October 19 and November 17, 1995.  We
interviewed the 10 medical experts between October 31 and December
12, 1995. 

We took steps to safeguard the privacy of the experts' responses. 
All categorical responses were reported in summary form.  When
specific comments were discussed in the report, we did not include
any information that could be used to identify individual
respondents.  All identifying information was removed from the
responses so that they could not be matched with individual experts. 

We tabulated the responses for each question to obtain an overall
assessment of the experts' opinions.  We also did a content analysis
of the medical care measurement experts' identification of advantages
and disadvantages of changing the current measurement of medical care
costs to an approach that more closely approximates a cost-of-living
measure.  From an initial reading of the responses, we developed a
list of cited advantages and disadvantages.  We used this list to
code the responses of all experts.  The coding of the responses was
verified by a second coder, and a third person checked coding
reliability.  As a method of focusing our analysis on the recurring
advantages and disadvantages identified by the experts, we adopted a
decision rule to report only those advantages and disadvantages cited
by two or more experts. 

Experts' comments made throughout the data collection instruments
were identified separately and were used in our descriptions of the
experts' responses.  We did a content analysis of their comments. 
From an initial reading of the responses, we developed lists of
comments made by the housing measurement experts; lists were also
developed from the medical care measurement experts' comments.  We
used these lists to code the responses of all experts.  The coding of
the responses was verified by a second coder, and a third person
checked coding reliability.  As a method of focusing our analysis on
the recurring comments made by the experts, we adopted a decision
rule to report only those comments made by two or more experts. 
Unless we explicitly state otherwise, when we refer to the experts'
opinions, at least two of the experts expressed the same view that is
being cited.  We used these comments to illustrate our results. 

To gain an understanding of the methodologies used in the CPI, we
reviewed relevant literature and held discussions with experts in the
field.  These experts included individuals associated with the CPI at
BLS, as well as private organizations and academic institutions.  We
also obtained information from BLS officials on their plans to revise
and improve the CPI.  On the basis of our literature reviews and
discussions, we identified the major concerns that were associated
with the asset-price approach that was used to measure homeowners'
costs before 1983.  We also identified the conceptual limitations of
the current medical care costs methodology in terms of its
measurement of the cost of living.  We recognize that these concerns
and limitations are not exhaustive; however, we believe that the most
significant concerns and limitations were identified. 


CONSUMER PRICE INDEX
========================================================== Appendix II

A CPI is an economic statistic that measures changes over time in the
general level of prices of goods and services that a reference
population acquires, uses, or pays for consumption.\1 Ralph Turvey
suggested that the design of a CPI rests upon how it is used and
identified six uses:  (1) general measure of inflation; (2)
indexation by government--adjusting the burden upon people paying
taxes, fines, or fees so that they pay in inflation-free currency,
and preserving the purchasing power of people receiving government
transfer payments; (3) prices and wage and salary adjustment in
contracts; (4) current cost accounting--revaluation of fixed assets
and stock for accounting purposes; (5) national accounting
deflation--obtaining constant- price estimates of consumption
expenditures; and (6) retail sales deflation.\2


--------------------
\1 In this report, consumption is defined as purchased goods and
services; the use of semidurable and durable goods owned or rented;
free goods of nature, such as air and water; and public goods. 

\2 Ralph Turvey, Consumer Price Indices:  An ILO Manual (Geneva: 
International Labor Office, 1989), pp.  4-6. 


   DEFINITION AND USES OF THE U.S. 
   CPI
-------------------------------------------------------- Appendix II:1

In the United States, the Bureau of Labor Statistics (BLS) of the
Department of Labor produces the CPI, which is a measure of the
average change over time in the prices paid by urban consumers for a
fixed market basket of consumer goods and services.  BLS has
identified the following three major uses of the CPI: 

  -- economic indicator of inflation.  The administration, Congress,
     and the Federal Reserve use trends in the CPI as an aid in
     formulating fiscal and monetary policies.  Business and labor
     leaders, as well as private citizens, use the CPI as a guide to
     making economic decisions. 

  -- escalator for wages, income payments, and tax brackets.  The CPI
     is used by collective bargaining units to adjust wages of
     approximately three million workers.  It is used to adjust some
     federal benefit payments for inflation.  For example, in March
     1995, as a result of changes in the CPI, 43 million Social
     Security beneficiaries; 6 million Supplemental Security Income
     recipients; 6 million railroad, military, and federal civilian
     retirees and survivors; and about 26 million food stamp
     recipients had their benefits adjusted for inflation.  The CPI
     is also used to adjust the federal individual income tax
     structure to prevent bracket creep (i.e., increases in real tax
     rates due solely to inflation).  Benefit payments, deductions
     for personal exemptions, and tax brackets are adjusted
     automatically by the CPI, rather than on the basis of
     discretionary policy decisions. 

  -- deflator of selected economic statistical data series.  The CPI
     is used to adjust selected economic statistical series for price
     changes and to translate these series into inflation-free
     dollars.  Examples of data series adjusted by the CPI include
     retail sales, hourly and weekly earnings, and components of the
     National Income and Product Accounts. 


   CONSTRUCTION OF THE U.S.  CPI
-------------------------------------------------------- Appendix II:2

Construction of the CPI begins by selecting a group of goods and
services that are usually bought by the reference population in the
index.  The collection of goods and services, called items, is known
as the market basket.  The CPI market basket is developed from
detailed expenditure information provided by families and individuals
who participate in the Consumer Expenditure Survey (CEX). 
Altogether, about 29,000 individuals and families provide expenditure
information for use in determining the importance, or weight, of each
item in the index structure.  These data are also used to select the
categories of items from which specific unique commodity and service
items are selected to be priced for the CPI. 

BLS measures price changes each month by checking the prices of the
items in the market basket and then comparing the aggregate costs of
the market basket with those for the previous month.  BLS obtains
prices for most items through personal visits by its field
representatives to approximately 30,000 retail establishments.  BLS
staff also sample about 60,000 housing units to obtain information on
housing costs. 


      COMPONENTS
------------------------------------------------------ Appendix II:2.1

BLS classified all CEX expenditure items into over 200 categories,
arranged into seven major components:  (1) food and beverages; (2)
housing; (3) apparel and upkeep; (4) transportation; (5) medical
care; (6) entertainment; and (7) other goods and services, such as
haircuts, college tuition, and bank fees.  Taxes that are directly
associated with the prices of specific goods and services, such as
sales and excise taxes, are also included.\3


--------------------
\3 The CPI excludes taxes not directly associated with the purchase
of consumer goods and services, such as income and Social Security
taxes.  The CPI does not include investment items, such as stocks,
bonds, real estate, and life insurance, because they relate to
savings and not daily living expenses. 


      EXPENDITURE WEIGHTS
------------------------------------------------------ Appendix II:2.2

Expenditure weights are used to give proportionate emphasis for price
changes of one item (or component) in relation to other items
(components) in the CPI and are derived from the expenditure on those
items as recorded in the CEX.  Expenditure weights allow the CPI to
distinguish between items that have a major impact on consumers and
to provide appropriate emphases to price changes associated with
them.  For example, the weight for airline fares is larger than the
weight given to white bread because a larger proportion of aggregate
consumer expenditures is made on airline tickets than on white bread. 
Therefore, if the price of white bread dropped more than one-half of
its previous price, the CPI would experience a smaller amount of
change than if the price of airfares experienced a slight decline. 

Based on average expenditures during the reference period,
expenditure weights remain fixed or constant until the next major
revision of the CPI and serve as a benchmark from which price
comparisons are calculated.  The weights of the components for the
last major revision in 1987 are those as derived from the 1982
through 1984 CEX (see fig.  II.1). 

   Figure II.1:  Expenditure
   Weights for 1987 CPI Revision

   (See figure in printed
   edition.)

Source:  BLS. 


      RELATIVE IMPORTANCE
------------------------------------------------------ Appendix II:2.3

A concept related to expenditure weights is the relative importance
of an item, which can be used to show the direct effect an item has
on the overall CPI price change.  The relative importance shows the
share of total expenditure that would occur if consumed quantities of
the items remain constant.\4 Although the expenditure weights remain
fixed until the CPI is rebased in a major revision about every 10
years,\5 the relative importance changes over time reflecting the
effect of price changes. 

Expenditure weights proportions equal the relative importance
percentages at the time of a major revision.  But since BLS maintains
the quantities of the items as the same amounts that were consumed in
the base period, the relative importance percentages change as a
result of changing prices.  Items registering a greater-than-average
price increase become relatively more important.  Conversely, items
registering a smaller-than-average price increase become relatively
less important.  Therefore, as the time between major revisions
increases, items with higher-than-average rates of inflation have
increasing rates of influence upon the CPI.\6 As shown in figure
II.2, the relative importance of medical care in the index for all
urban consumers, which was 6.0 in December 1988, increased to 7.4 in
December 1995 because medical prices increased at a greater rate than
the rate for the all items CPI--the overall CPI.  During the same
period, the relative importance of housing fell from 42.3 percent to
41.3 percent because housing prices increased at a lower rate than
the all items CPI. 

   Figure II.2:  Relative
   Importance of Components in the
   CPI-U, 1988 and 1995

   (See figure in printed
   edition.)

Source:  BLS. 


--------------------
\4 The CPI is estimating the change in the cost of a constant level
of consumption by using the same quantities of items in the fixed
market basket.  This methodology, however, does not maintain
consumers' utility with this fixed-quantity, fixed-item market
basket. 

\5 Major revisions were made to the CPI in 1940, 1953, 1964, 1978,
and 1987; a major revision is planned for 1998. 

\6 "Substitution" bias is directly related to this feature of the
CPI.  Because the quantities remained fixed, the CPI fails to reflect
consumer behavior in which consumers purchase larger amounts of
cheaper items that can be substituted for items that have become more
expensive.  This results in an upward bias in the CPI, relative to a
comprehensive cost-of-living index, as it gives increasing importance
to items with higher-than-average price increases.  A downward bias
can also occur when consumers are driven to purchase more expensive
substitutes, which occurs during wars and mandatory price controls. 


      PRICING OF MARKET BASKET
      ITEMS
------------------------------------------------------ Appendix II:2.4

Each month, BLS field representatives visit or call thousands of
retail stores, service establishments, rental units, and doctors'
offices all over the United States.  For the entire month, they
record the prices of about 94,000 items.  To determine which retail
outlets its representatives should visit to obtain its monthly price
quotations, BLS sponsors the Point-of-Purchase Survey, conducted by
the Bureau of the Census.  The survey respondents are asked by outlet
categories, such as doctors, whether they made specific purchases
and, if so, the names and locations of all places of purchases and
the expenditure amounts.  BLS uses the results from the survey to
select outlets for pricing.  This survey is conducted in
approximately 20 percent of a sample of urban areas each year and, as
a result, the entire nonshelter sample is updated every 5 years. 


      PRICE COLLECTION FOR
      NONSHELTER ITEMS
------------------------------------------------------ Appendix II:2.5

BLS field representatives visit each selected outlet to select items
that will be priced either monthly or bimonthly.  For each outlet,
categories of items are selected for pricing.  Using probability
selection methods based on revenues and volume information provided
by the retail outlet, BLS field representatives use a random numbers
table to select a unique item within the specified categories for
pricing.  The monthly price changes for the same item (e.g.,
cigarettes) that are collected by BLS field representatives in urban
areas throughout the United States are averaged, weighted according
to their relative importance, and published for selected items and
the all items CPI. 


         PRICE COLLECTION FOR
         SHELTER ITEMS
---------------------------------------------------- Appendix II:2.5.1

BLS uses monthly price changes of units in its housing survey for the
residential rent and owners equivalent rent items in the CPI housing
component.  Homeowners' equivalent rent is estimated from
approximately 36,000 rented units and 26,000 owned units in the BLS
housing survey.  To obtain values of implicit rent each month, BLS
assigns a set of renter units to each owner unit based on similar
housing characteristics and tracks the rent paid for these units. 
Each month, BLS field representatives obtain information from renter
units on the rent for the current month, the previous month, and what
services are provided.  From owner units, field representatives
obtain an estimated implicit rent-- what the owners think they could
rent the house for monthly, not including utilities.  This collected
implicit rent is only used to weight the owner-occupied housing unit;
it is not used to estimate the movement of the owners' equivalent
rent. 


         PRICE COLLECTION FOR
         MEDICAL CARE ITEMS
---------------------------------------------------- Appendix II:2.5.2

Medical care prices are collected in a unique manner by BLS because
the medical care component of the CPI is based on out-of-pocket
expenses rather than on total health care expenditures.  Therefore,
pricing for health insurance premiums and some other medical care
goods and services, which are paid indirectly through
consumer-purchased health insurance, is approached in a different
manner than the items previously described. 

Only the portion of health insurance premiums paid by the consumer is
included in the CPI; premiums that are paid by others, such as
employers, are excluded.  The health insurance item in the CPI
includes only that portion of the premium that is retained by the
insurance carrier for administrative costs and profits (about 10
percent of the premiums for the CPI) and is generally referred to as
retained earnings. 

The portions of the premium that are paid as benefits have been
assigned to relevant medical care categories, such as hospital rooms. 
BLS calculates the price for the health-insurance- represented
portion of the medical care category as a product of the price of the
item and the retained earnings of the health insurance carrier.  For
example, the CPI's index value for health insurance hospital room is
affected by both the insurance companies' retained earnings and
hospital room prices; whereas, the hospital room paid-out-of-pocket
and health insurance benefit payments for rooms are affected by
hospital room price changes. 

Unrelated to the problems associated with third-party payers for
medical care, in January 1995, BLS implemented changes in its
collection of prescription drug prices.  These changes addressed
problems associated with the introduction of generic drugs when the
patent protection for the brand drug expires.  The methodology that
BLS field representatives used to price prescription drugs was
altered.  When the patent expires for a brand drug that has been
consistently priced in the CPI, BLS field representatives repeat the
selection process for the pharmaceutical drug to allow for the
possible selection of the generic substitution to represent that
drug.\7

Like many other items, there are multiple prices paid for medical
care provided by a single provider.  For example, a medical care
provider may have a list price that is charged to a patient who pays
the fees directly, but the provider is also likely to have discounted
fees that have been negotiated with third-party payers, such as a
commercial health insurance company.  Efforts on the part of service
providers to segment their markets and to offer discounts have made
the measurement of price change difficult. 

In recent years, BLS has been attempting to obtain actual transaction
prices from medical care providers.  These prices represent the total
amount of payment received from third-party payers as well as
individuals who make coinsurance payments.  Medicaid or other public
assistance discount rates are not included because the government
makes direct payment for these medical care goods and services.  In
1987, BLS field representatives began to select discounted fees other
than list prices for physician services.  Field representatives began
selecting discounted fees from dentists in 1991, eye care providers
in 1992, hospitals in 1993, and from pharmacies in 1995. 


--------------------
\7 For a discussion of this methodological change in relation to the
Producer Price Index, see Prescription Drug Prices:  Official Index
Overstates Producer Price Inflation (GAO/HEHS-95-90, Apr.  28, 1995). 


HISTORICAL DEVELOPMENT OF HOUSING
AND MEDICAL CARE COMPONENTS
========================================================= Appendix III

The following sections describe the development of the various
methodologies used by the Bureau of Labor Statistics (BLS) to measure
homeowners' costs and medical care costs in the CPI. 


   PREVIOUS METHODS USED TO
   MEASURE HOUSING
------------------------------------------------------- Appendix III:1

Before 1953, there was no separate housing component within the CPI. 
Purchases of new homes were not included in the market basket, but
several types of housing expenses, such as rent, utilities, and
current maintenance costs were included.  In the mid-1930s, a
majority of families rented their homes; only about 30 percent were
homeowners.  Current maintenance payments-- mortgage interest,
property taxes, insurance, repairs, and financing charges connected
with buying and selling a house--of homeowners in the index
population were included in the expenditure weight for rent.  Because
home purchase costs and payments on mortgage principal were
considered as savings rather than expenditures on consumer goods,
they were not included in the CPI market basket.  Price changes for
homeowners' current maintenance items were assumed to be consistent
with the rate of price changes in rent, so this information was not
separately gathered or surveyed. 


      1953 CPI REVISION CREATED A
      HOUSING INDEX THAT MEASURED
      CHANGES IN THE COST OF
      ACQUIRING AND MAINTAINING
      HOUSES--AN ASSET- PRICE
      APPROACH
----------------------------------------------------- Appendix III:1.1

Changes in housing trends, including a large number of new units and
a continuing shift from rental to owner occupancy, caused BLS to
create a separate housing index in the 1953 major revision.  BLS
altered the way housing costs were included in the index because of
the increase in homeownership among urban wage earner and clerical
worker families after World War II.  According to BLS, by 1950, 49
percent of the families surveyed owned their own homes.  Extension of
credit on easy terms also made consumers less willing to defer
purchasing a house.  BLS dropped the assumption that home purchases
should be viewed as saving, while it broadened the definition of
housing to include all expenses connected with acquiring and
operating a home. 

In 1953, homeownership became an item in the housing component of the
CPI.  Consistent in concept and measurement with the overall CPI as a
measure reflecting price changes, the asset-price approach to
measuring homeowners' costs was designed to measure the price change
of all costs associated with acquiring and operating a home during
the current reference period.  Thus, there was no distinction between
the consumption of durable versus nondurable goods.  If a home was
purchased during the reference period, the CPI included it as if it
were consumed that year.  If a home was purchased before or after the
reference period, it was not included in the CPI.  The purchase of a
home and its mortgage was measured the same way as other items in the
CPI. 

The asset-price approach consisted of five parts:  property taxes,
property insurance, property maintenance and repairs, housing prices,
and contracted mortgage interest costs.  Expenditure weights for
property taxes, insurance, and maintenance and repairs were based on
the average expenditures of all households in the reference period.\1
Expenditure weights for contracted mortgage interest and housing
prices were based only on individuals who purchased a home during the
reference period.  These weights were based on the total price that
these individuals paid for homes purchased during the survey year,
minus the price of homes that they sold during that year, plus costs
associated with the purchase and selling of homes.  The index weights
included only newly purchased homes, or homes entering the market as
owner occupied for the first time and the total amount of interest
expected to be paid over the first one- half of the stated life of
the mortgage.\2


--------------------
\1 For a description of expenditure weights and relative importance
of items in the CPI, see appendix II. 

\2 From 1953 through 1964, the entire mortgage payment for a house
was included.  By the 1964 revision, BLS determined that mortgages
were either paid off or refinanced at about one-half of the term of
the mortgage. 


      CONCERNS EXPRESSED ABOUT THE
      ASSET-PRICE APPROACH
----------------------------------------------------- Appendix III:1.2

In 1981, we reported that the asset-price approach to measuring
homeowners' costs was subject to criticism and documented several of
the concerns in our report Measurement of Homeownership Costs in the
Consumer Price Index Should be Changed (GAO/PAD-81-12, Apr.  16,
1981).  For this study, we categorized these and other concerns we
located in our literature search or that were expressed to us in our
discussions with experts.  We also sent the categorized concerns to
our recognized housing measurement experts (see app.  V). 

1.  The asset-price approach included the value of the asset of the
home.  The 1961 Stigler committee report criticized the asset-price
approach because it measured the change in the prices of assets,
rather than the change in the user cost of consuming the flow of
services provided by durable goods.  By doing so, the asset-price
approach failed to separate the investment aspect of homeownership
from the immediate consumption (flow-of-shelter services) aspect of
homeownership. 

2.  The asset-price approach misstated inflation.  From 1959 through
1980, the homeownership item rose at a higher rate than other items
and the all items CPI.\3 The homeownership item of the CPI did not
track the same as other housing price measures from real estate
experts, and it did not track the same as the personal consumption
expenditures price deflator.  Also, during the period of double-digit
inflation, it was affected by the volatility of housing prices and
mortgage interest rates. 

3.  The asset-price approach misrepresented the cost of living.  This
approach was not based on the cost-of-living concept.  Notably, it
did not measure actual outlays for shelter by failing to incorporate
concepts such as income tax deductions for homeownership.  Because of
the methodologies used to determine the weighting and price changes
of homeownership, critics charged that the measure distorted the cost
of living for the basic necessities of food, housing, fuel, and
medical costs for the elderly and the poor, who were least likely in
the population to buy new homes. 

4.  The asset-price approach failed to represent all homeowners.  The
expenditure weights used for house prices and mortgages were based on
6 percent of the respondents in the consumer expenditure survey who
purchased a home during the survey period.  The weighting
calculations effectively excluded survey respondents who were making
a mortgage payment during the survey period and those who had paid
off their mortgage prior to the survey.  The asset-price approach was
affected by the rate of new home purchases in the reference period,
which could vary widely for a number of reasons between revisions. 
By including the total net purchase price and the first one-half of
the mortgage costs for homeowners who recently purchased their homes,
the weights for these two items were viewed to be too large.  For
example, the ratio between homeownership and residential rent in the
CPI was double the same ratio in the National Income and Product
Accounts. 

5.  The asset-price approach used limited home sales and mortgage
data.  BLS could not obtain consistent house price data in local
areas.  As a result, it used Federal Housing Administration's (FHA)
203b data to calculate monthly price changes in house prices.  BLS
also relied on data on conventional mortgages supplied by the Federal
Home Loan Bank Board (FHLBB) and interest rate ceilings on FHA and
Department of Veterans Affairs (VA) loans for mortgage interest
rates.  As a result of the data limitations, the pricing information
used to move the homeownership item was based on a small
unrepresentative sample of the housing market.  For example, by 1980,
monthly changes in the index were based on homes with FHA-insured
mortgages that had a low cap ($67,500) and low interest rates (2
percent below market rates), which caused buyers to make large down
payments on high- priced homes to get the lower FHA interest rates. 
This consumer behavior raised the housing prices of FHA-financed
homes.  Because FHLBB, FHA, and VA data did not contain information
about new mortgage instruments, such as adjustable rate mortgages and
owner- held financing, FHA data were becoming less representative of
the mortgage markets.  Also, during this time, FHA data were
declining in quality because FHA was changing the methods used to
collect its data. 

6.  The asset-price approach included interest costs.  Unlike some
other durable goods, contracted interest paid for homes was included
in the CPI.  Government action to raise or lower interest rates was
thereby reflected directly in the CPI through the inclusion of
mortgage interest.  Because the executive branch could determine the
interest rate ceiling for FHA-financed homes, it could have some
control over inflation as measured by the CPI. 


--------------------
\3 In the questionnaire we sent to the housing measurement experts,
we stated that the homeownership item rose at more than double the
rate of other items and of the all items CPI.  BLS analysts cannot
confirm this magnitude of difference for these years; the data are no
longer available to calculate the amount of difference. 


      ALTERNATIVE METHODOLOGIES
      CONSIDERED
----------------------------------------------------- Appendix III:1.3

BLS considered several alternative methods of measuring homeowners'
costs and, in 1980, published five experimental measures that
represented alternative homeownership concepts.  Two of these
measures used a user cost approach, which included prices for
property taxes, property insurance, property maintenance and repairs,
and mortgage interest rates, and made adjustments for the interest
equity and appreciation or depreciation of the house's value.  Two
measures applied an outlays approach, which used an average of the
interest rates paid over time instead of interest paid in current
dollars, for a house purchased during the reference period.  The
fifth measure used a rental equivalence method, which was designed to
capture the value of shelter services for owner-occupied units by
determining what rent would be charged for those units had they not
been owner occupied. 

In January 1983, BLS changed the measurement of homeowners' costs in
the CPI-U from an asset-price approach to the rental equivalence
method (the CPI-W was changed in 1985).  Even before the
implementation of the rental equivalence method, concerns emerged
about how BLS would identify rental units that were comparable in
structure and location to owner-occupied dwellings.  To use a rental
equivalence method, BLS developed a rent index that could estimate a
rental price for owner-occupied units.  Rental housing, however, is
generally located in different neighborhoods and may not be easily
matched with owner-occupied housing.  Some researchers question the
conceptual basis of using rental data to represent homeowners' costs,
since the motivations to buy and rent are different.  Also, questions
have been raised by researchers about how BLS treats vacant units,
units under rent control restrictions, and remodeling of
owner-occupied units. 

BLS implemented adjustments in 1988 and 1995 to address some concerns
with the rental equivalence method.  In 1988, BLS began making
adjustments for the depreciation of housing to help separate quality
changes from price changes.  These technical adjustments are applied
to the change in rent for each shelter index in each geographic area. 
The same adjustment was made to both the renters' and owners'
equivalent rent indexes.  In January 1995, two technical changes were
made to improve the validity and reliability of the residential rent
and homeowners' equivalent rent indexes.  According to BLS officials,
major enhancements to the housing component are planned for the
upcoming major revision.  These improvements are to be implemented
beginning with data for January 1999. 


   PREVIOUS METHODS USED TO
   MEASURE MEDICAL CARE
------------------------------------------------------- Appendix III:2

Medical care has always been in the CPI, and consumer-purchased
health insurance has always been included as a medical expense. 
Originally, the rate of price change for health insurance was assumed
to be equal to the average of other medical items.  BLS, in the 1950
revision, deviated from this approach and began to directly price
health insurance policies.  The approach was changed in the 1964
revision to an approach that again based health insurance price
changes on prices observed for other medical goods and services, as
well as the insurance carriers' operating costs and profits.  In the
1978 and 1987 revisions, BLS made minor adjustments in measuring
health insurance. 


      HEALTH INSURANCE DIRECTLY
      PRICED IN 1950
----------------------------------------------------- Appendix III:2.1

In 1950, BLS began direct pricing of health insurance by pricing the
gross premium rate for the most widely purchased family group
contract for Blue Cross hospitalization.  In December 1958, the Blue
Shield surgical insurance rates were similarly included in the index. 
Two concerns, however, emerged with the CPI's approach of direct
pricing of health insurance policies:  (1) defining constant quality
for making adjustments in premium rate changes and (2) lack of
geographic delineation within industrywide health insurance policies. 

Quality adjustment problems centered on changes in (1) utilization of
health care services, which is based on the total number of paid
insurance claims and (2) benefit coverage.  According to BLS, changes
in utilization presented problems both in concept and application. 
Rate changes due to increased utilization of health care services
were treated as price changes, rather than as changes in benefit
coverage.  An increase in utilization was assumed to mean an increase
in the number of claims, which translated to greater costs to the
insuring company and the consumer.  Some critics in the medical care
field objected to using changes in utilization as price increases
because doing so failed to eliminate from the CPI the effect of a
change in the quantity of medical services provided.  Health
insurance rate changes resulting from changes in benefit coverage
were excluded from the CPI; however, isolating such changes became
increasingly difficult.  For example, charges for hospital rooms or
physician fees can increase at the same time that a benefit plan
increases its coverage of these services.  In addition, Blue Cross
and Blue Shield reported data quality problems and a lack of uniform
reporting from various geographic areas, which led to difficulties in
discerning the utilization and cost factors affecting a benefit plan. 

Another problem BLS had with direct pricing of health insurance
policies was that many commercial insurance carrier contracts were
written on an industrywide basis, rather than by geographic areas. 
Data needed to calculate the CPI by various locations were not
available.  BLS, therefore, had to estimate prices by locations from
the industrywide insurance data. 

In response to the concerns about direct pricing of health insurance
policies, BLS engaged in a series of conferences with the Health
Insurance Association of America to discuss the uncertainties in
pricing premium rates.  BLS concluded that pricing medical care
services directly, rather than direct pricing of health insurance
policies, would be the best method for measuring health insurance
rate changes. 


      1964 CPI REVISION DROPPED
      DIRECT PRICING OF HEALTH
      INSURANCE
----------------------------------------------------- Appendix III:2.2

In the 1964 revision, the methodology BLS used to directly price Blue
Cross and Blue Shield premiums was dropped from the CPI.  In its
place, BLS adopted a method that indirectly priced a number of
hospital and professional services on the basis of claims data\4 plus
retained earnings--the operating costs in administering the insurance
plan and any remaining profit. 

Another change BLS made to the CPI in 1964 reversed the approach
taken to reflect changes in the utilization of health care services
as price movements.  Beginning in 1964, such changes were viewed as a
redefinition of the risk being covered, or benefit coverage, thus a
change in the quality of health care.  Under this interpretation,
premium changes resulting from utilization were not to be reflected
in the CPI because they were no longer viewed as price changes. 

Again, in this revision, the weight, or proportionate importance, for
the health insurance component was based on the average expenditure
for health insurance reported in the Consumer Expenditure Survey
(CEX).\5 The share of insurance paid by the employer was excluded
from the weight; employers' contributions for health care were viewed
as income and not as expenditures for employees.  The weight was
divided into two subweights:  (1) claims data, which reflect the
benefits companies pay out; and (2) retained earnings--the funds
companies have left over after paying for benefits.  Although
consumers do not pay directly for them, consumers' insurance premium
payments are affected by changes in retained earnings. 


--------------------
\4 Prices were already being collected for medical services as part
of the BLS pricing program for out-of-pocket medical expenses. 

\5 The CEX is a survey of the spending habits of American families. 
Each year approximately 4,800 families are interviewed. 


      MEDICAL CARE COMPONENT
      PUBLISHED SEPARATELY FOR THE
      FIRST TIME
----------------------------------------------------- Appendix III:2.3

The 1978 revision of the CPI introduced a new process that priced a
broader range of services and, for the first time, medical care was
separated from recreation and published as a separate component of
the CPI.  Also, a new sample design and collection methodology was
incorporated into the 1972 through 1973 CEX, which improved accuracy,
and provided more complete data for the selection and weighting of
items for the CPI market basket.  For example, the number of basic
types of priced medical services increased from 33 to 258 under the
1978 revision. 

In 1987, BLS made definition changes in the representation of health
insurance premiums; the changes, however, did not alter the index
results.  BLS viewed these changes as clearer definitions of health
insurance's role in the CPI.  The changes were in the presentation of
medical subcategories.  Instead of presenting information separately
for items by those paid by out-of-pocket and those paid by insurance,
in the 1987 revision, the expenditures were combined for a single
medical item. 


      BLS BEGAN TO INCORPORATE
      DISCOUNTED TRANSACTION
      PRICES IN 1987
----------------------------------------------------- Appendix III:2.4

Until 1987, all medical care price changes were based on the list
prices that providers charged a patient who pays the fees directly. 
As described in appendix II, BLS began incorporating discounted
transaction prices--fees that have been negotiated with third-party
payers--for physicians' services in 1987.  BLS made this change
because it recognized that up until this time, price changes for both
out-of-pocket expenses and insurance benefit payments were based on
list prices.  In 1991, BLS began collecting transaction prices for
dentists and in 1992 for eye care providers. 

During 1993, BLS began including discounted transaction prices for
hospital services, and about 15 percent of the hospital prices in the
CPI are discounted prices.  The newest procedure for obtaining
hospital transaction prices is for BLS field staff to select from
sampled hospitals a payer type and then ask for a patient's bill
associated with that payer.  BLS obtains from the bill the diagnosis
for the hospitalization and then identifies the eight largest items
on the bill and finds out what prices the health plan has negotiated
with the hospital.  BLS records whatever the plan has negotiated.  A
BLS official said, however, not all hospitals cooperate in providing
this information, thus, published charges are still collected for
some outlets. 


EXPERTS AND THEIR AFFILIATIONS
========================================================== Appendix IV


   HOUSING MEASUREMENT EXPERTS
-------------------------------------------------------- Appendix IV:1

David Carrier
Regional Financial Associates

Robert Gillingham
Department of the Treasury

Pat Hendershott
The Ohio State University

Steve Malpezzi
University of Wisconsin at Madison

Joel Popkin
Joel Popkin and Company

William Randolph
Congressional Budget Office

Thomas Thibodeau
Southern Methodist University

Robert VanOrder
Federal Home Loan Mortgage Corporation

John Weicher
Hudson Institute

David Wyss
DRI/McGraw-Hill


   MEDICAL CARE MEASUREMENT
   EXPERTS
-------------------------------------------------------- Appendix IV:2

Ernst Berndt
Massachusetts Institute of Technology

David Cutler
Harvard University

Patricia Danzon
University of Pennsylvania

David Dranove
Northwestern University

Charlie Fisher
Jung Xing Associates

Mark Freeland
Health Care Financing Administration

Thomas Getzen
Temple University

Robert Graboyes
Federal Reserve Bank of Richmond

Joseph Newhouse
Harvard University

James Rogers
American Medical Association


DATA COLLECTION INSTRUMENTS AND
BACKGROUND INFORMATION OF THE CPI
COMPONENT METHODOLOGIES
=========================================================== Appendix V

This appendix contains copies of the cover letters,
instruction/introduction sheets, and questionnaires that we sent to
the experts on the housing and medical care components of the CPI. 



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)




(See figure in printed edition.)Appendix VI
COMMENTS FROM THE BUREAU OF LABOR
STATISTICS
=========================================================== Appendix V

See comment 1. 



(See figure in printed edition.)

See comment 2. 



(See figure in printed edition.)

Now on p.  28. 

Now on p.  27. 

See comment 3. 

Now on p.  31. 



(See figure in printed edition.)



   GAO COMMENTS
--------------------------------------------------------- Appendix V:1

The following are GAO's comments on BLS' letter dated July 11, 1996. 

1.  BLS commented that although we make no recommendations, we hold
the current CPI up against the "cost-of-living concept." We held the
housing and medical care components of the current CPI up against the
cost-of-living concept in order to carry out the work objectives that
the Ranking Minority Member, House Committee on Banking and Financial
Services requested. 

2.  BLS commented that it is unclear what interpretation could be
given to an index that included an expanded scope only in the medical
care component.  As discussed in chapter 4, we agree that the overall
impact of changing just the medical care component of the CPI is
unknown.  This contributed to our not taking a position on whether
the medical care component of the CPI should be changed to reflect
the cost of living. 

3.  We added BLS' views on the remaining problems in measuring price
change in medical care to the discussion of medical care transaction
prices in chapter 3. 


MAJOR CONTRIBUTORS TO THIS REPORT
========================================================= Appendix VII

Kathleen K.  Scholl, Senior Economist
Anthony Assia
Thomas M.  Beall
Edward J.  Laughlin
James M.  McDermott
Pamela R.  Pavord
Loren Yager





RELATED GAO PRODUCTS
=========================================================== Appendix 1

Economic Statistics:  Status Report on the Initiative to Improve
Economic Statistics (GAO/GGD-95-98, July 7, 1995). 

Economic Statistics:  Measurement Problems Can Affect the Budget and
Economic Policymaking (GAO/GGD-95-99, May 2, 1995). 

Prescription Drug Prices:  Official Index Overstates Producer Price
Inflation (GAO/HEHS-95-90, Apr.  28, 1995). 

Developing a Consumer Price Index for the Elderly (GAO/T-GGD-87-22,
June 29, 1987). 

Stabilizing Social Security--Which Wage Measure Would Best Align
Benefit Increases With Revenue Increases?  (GAO/IMTEC-85-13, Aug. 
27, 1985). 

Funds Needed to Develop CPI Quality Control System (GAO/GGD-83-32,
Apr.  1, 1983). 

A CPI for Retirees Is Not Needed Now But Could Be in the Future
(GAO/GGD-82-41, June 1, 1982). 

A Consumer Price Index for Retirees and Alternatives for Controlling
Indexing (Testimony, Apr.  20, 1982). 

Measurement of Homeownership Costs in the Consumer Price Index Should
Be Changed (GAO/PAD-81-12, Apr.  16, 1981). 

Alternatives for Modifying the Indexation of Federal Programs
(Testimony, Mar.  10, 1981). 

*** End of document. ***