The Experimental CPI Using Geometric Means
An experimental Consumer Price Index (CPI) for All Urban Consumers,
using a geometric mean formula to combine individual price quotations
at the lower level of aggregation while keeping the current Laspeyres
arithmetic mean formula for higher level aggregation, is being
introduced by the Bureau of Labor Statistics, U.S. Department
of Labor. This experimental index, called the experimental
CPI using geometric means
, or CPIUXG, is a supplement
to the price information available from BLS and does not replace
the published CPIU or CPIW indexes. BLS currently
is evaluating the full or partial adoption of the geometric mean
formula in the CPIU and CPIW.
The current CPI, which is calculated using a fixed-weight Laspeyres
formula, does not reflect the fact that consumers can and do change
spending patterns as relative prices change. Under certain assumptions,
a measure of change in consumer prices that uses a geometric mean
formula will successfully account for this consumer behavior.
To the extent that those assumptions are accurate, the index using
geometric means will provide a closer approximation to a cost-of-living
Each month in the CPI, prices of individual items are averaged
together to produce indexes for 9,108 item-area categories. It
is at this first or lower level where the CPIUXG uses
geometric means instead of the Laspeyres method (arithmetic means)
in the estimation of price relatives. At the second or upper
level, these basic indexes are combined to produce the All Items
CPI as well as major group and other intermediate indexes. In
both the official CPIU and the CPIUXG, basic
indexes are aggregated at the upper level using the arithmetic
means and not geometric means.
The CPIUXG indexes are available for the period from
December 1990 through February 1997 and will be issued in the
future according to the schedule set out below. Historical differences
between the CPIU and CPIUXG arise not only because
of the different formulas used, but also because of methodological
changes made in the CPIU since 1990 that are reflected in
the CPIUXG for the entire series. For historical comparison
to the CPIUXG, therefore, BLS also is issuing an experimental
Test Laspeyres series called the CPIUXL, which
differs from the CPIUXG only in the use of the Laspeyres
formula for aggregation of price quotations.
Over the period from December 1990 to February 1997, the CPIUXG
rose 16.2 percent, compared with 18.6 percent for the CPIUXL.
The average annual rate of growth in the CPIUXG over
this period was 2.46 percent, 0.34 percentage points lower than
the 2.80 percent annual growth rate of the CPIUXL.
Among major item groups, the largest differences between the two
indexes were in food and beverages, apparel and upkeep, medical
care, and entertainment.
The January 1995 methodological changes in the CPI treatment of
food at home appear to have reduced the difference between the
CPIUXG and CPIUXL. We expect to observe
a further narrowing of the gap as a result of additional changes
made in June and July of 1996.
Beginning on April 22, the Geometric Mean and Test Laspeyres indexes
will be released monthly one week after release of the CPIU.
These series will not require special embargo procedures and are
expected to be available at approximately 8:30 a.m. on the morning
of their release.
Each month, the most recent data will be added to the historical
tables for the National All-Items, Major Groups, and Special Indexes
(food, energy, and all items less food and energy) that are available
on the Internet at http://www.bls.gov/cpi/. This
information also can be obtained from the Consumer Price Index
information staff at (202) 606-7000. Indexes will be included
quarterly in the CPI Detailed Report, beginning with publication
of first-quarter 1997 CPI-U-XG data in the April Detailed Report.
The CPIUXG and CPIUXL will not be available
on a seasonally adjusted basis. Calculation and publication of
the CPIUXL will be discontinued after December 1997,
because that index is primarily of value for historical comparisons.
The relative movements in the CPIUXG will be studied
as part of a BLS research program to evaluate, item category by
item category, adoption of the geometric mean formula in all or
some components of the official CPI. Scanner data, studies of
substitutions between brands, and other information will be used
to assess the propensity of consumers to substitute across items
within individual item categories as the relative prices of those
By the end of 1997, BLS will announce the findings of its research,
including its determination of which CPI basic indexes are best
calculated with the geometric mean formula and when the implementation
of any change will take place. The likely date for that implementation
is with the release of January 1999 CPI data.
BLS's best estimate is that use of the geometric mean formula
in all CPI basic indexes would lower the growth rate of
the index by approximately one-quarter of one percent per year.
Partial adoption of the geometric mean formula would be expected
to have a downward impact of between zero and one-quarter of one
percent per year, depending on how many, and which, indexes continue
to be based on the Laspeyres approach.
For current CPIUXG and CPIUXL data, or
for further information, write to the Bureau of Labor Statistics,
Division of Consumer Prices and Price Indexes, Room 3615, 2 Massachusetts
Ave., NE, Washington, DC 20212-0001, or call Kenneth J. Stewart
at (202) 606-7000.
Last Modified Date: October 16, 2001