By Mark J. Warshawsky

The current bout of rapid price inflation, most recently
estimated at 9.1 percent at an annual rate, is based on a standard measure of consumer
prices calculated monthly by the Bureau of Labor Statistics (BLS) since 1978:
the CPI-U. “U” means it applies to urban consumers, representing 88 percent of
the US population. In addition to its consideration in public discussions and monetary
and fiscal policy actions, the CPI-U is used by the federal government to pay
interest on Treasury Inflation Protected Securities, and to adjust
retirement-related parameters in the tax code.

There are, however, other BLS measures of prices. The CPI-W,
“W” meaning urban wage and clerical workers, representing 28 percent of the
population, is the original measure of consumer prices produced by the BLS
going back to 1913. It is still used by the federal government to annually
adjust Social Security and federal employee pension benefits for a cost-of-living
adjustment (COLA). Although the CPI-U and CPI-W generally track closely, CPI-W
inflation, with a higher expenditure share on food and transportation, is
coming in at 9.8 percent, above the headline number of 9.1 percent, far above
the 3.8 percent projected for 2022 in the Trustees’ Reports released in June
2022, and even higher than the 8 percent indicated by the Social Security
actuary in remarks given almost immediately following the release of the Reports.
Yet another measure, the C-CPI-U, “C” meaning chained, published by the BLS
since 1999, is thought by many experts to be the most accurate and current of the
price measures to calculate inflation because it holds consumer welfare
constant, quickly recognizing the natural changes, “substitutions,” in consumer
buying practices as relative prices of goods and services change. It is now
used to adjust parameters in the income tax code such as brackets and has been
proposed by many experts as the best measure to use in determining COLAs. Most
recently, it has increased 8.4 percent; as we will see, its rate of change is
generally somewhat lower than the other measures.

The R-CPI-E, “E” meaning elderly and “R” meaning research,
has been calculated by the BLS since 1982 as an experimental index, to measure prices
based on the broad purchasing patterns of the population age 62 and older,
mainly reflecting a higher expenditure share on housing and medical care. In
the past, medical care inflation was quite high, but has cooled lately. Housing
inflation mainly reflects the change in the owner’s equivalent rental, which is
a conceptual measure, and does not reflect actual transactions. R-CPI-E has
been proposed in legislation, for example, “Social Security 2100,” to replace
the CPI-W for the Social Security COLA because it is perceived to measure a
higher rate of inflation. As we will see, this is barely true over the last
twenty years, and for the most recent period, its measure of inflation, at 8.5
percent, is notably below the indexes used in current law.

The BLS has described the particular inadequacies of the R-CPI-E.
The consumer expenditure survey used by the BLS for the CPI is designed in its
sampling properties for the large population of the U measure; smaller and
different populations such as W and E will have higher sampling errors. The
geographic distribution of the E population is different than that of the U
population, the retail outlets and specific items priced are likely to be
different among the two populations, and the E measure likely understates the
significance of senior discounts. It should also be noted that the E measure
per se would not be a good representation of inflation for non-elderly populations
getting Social Security benefits, such as children, some spouses and widows,
and the disabled, representing about a quarter of the total Social Security
population. In the opposite direction, one-fifth of the elderly population are
not Social Security beneficiaries and their spending patterns may differ. It
would be confusing for one agency to employ and explain to the public several
different measures for different populations, even if it were possible to be
more precise.

In Part II, we will look more precisely at how much these inflation metrics differ and the policy implications of these differences.

The post Inflation and the Social Security COLA, Part I: Introduction to Different Measures of Price Changes appeared first on American Enterprise Institute – AEI.