Mismeasuring Inflation

"Measuring Inflation in a High-Tech Age" by Leonard I. Nakamura, in Business Review (Nov.–Dec. 1995), Federal Reserve Bank of Philadelphia, Dept. of Research and Statistics, 10 Independence Mall, Philadelphia, Pa. 19106–1574.

Assessments of the economic state of the Union almost always revolve around the "fact" that Americans’ wages, corrected for inflation, have declined, falling from an average of more than $8 an hour (in 1982 dollars) in 1975 to less than $7.50 last July. Nakamura, an economic adviser in the research department of the Federal Reserve Bank of Philadelphia, contends that the decline is, in all likelihood, an illusion.

The culprit, he contends, is the Consumer Price Index (CPI), which measures changes in the cost of living by tracking the price of a fixed "basket" of goods. The CPI basket currently holds items selected in the early 1980s. But today’s actual consumer basket is different. Improvements in the quality of goods (e.g., personal computers and cars) and services (e.g., cable television and medical care) increase the standard of living yet are largely missed by the CPI. The result: at least a half-point overestimate of annual inflation.

There are other problems with the measure. If clothes go up in price, for example, while computer supplies go down, the consumer may buy more of the latter and fewer articles of clothing. The consumer is better off, but, again, the CPI, with its fixed basket, takes no notice, distorting the index further by two-tenths of a percentage point.

New products, such as CD-ROMs, that have come out since the basket’s contents were fixed, are largely ignored. The Bureau of Labor Statistics, which collects the basic data for the CPI, tries to keep abreast of new products by gradually fitting them into an existing category of goods and rotating part of the sample of stores and goods it surveys each year. But that procedure not only fails to capture all the improvements in the standard of living, Nakamura maintains; it itself pushes the inflation index further upward—by an estimated two-tenths to three-tenths of a percentage point a year. The reason: it gives greater weight to goods whose prices are likely to rise after their initial "sale price" introduction on the market.

All in all, Nakamura calculates, the CPI probably has been overstating inflation by more than one percent a year. If the index is revised downward by that amount, the post-1975 decline in real wages becomes an increase (to about $9.50, in 1982 dollars). If Nakamura is right about this politically charged subject—which leaves economists, as usual, divided—then other items tied to the CPI, including Social Security payments and personal income tax brackets, have also been distorted.

Rapid advances in computers and telecommunications are responsible for many of the quality-of-life improvements that go unmeasured by the CPI. But such new technologies may also be part of the solution. The Bureau of Labor Statistics (which, it should be noted, has been taking steps to improve the measurement of inflation) still does its work the old-fashioned way, sending people out to stores to gather data. If the bureau instead could electronically tap into the detailed information on sales that many retailers, whole-collect, a more accurate picture of inflasalers, and manufacturers now routinely tion might emerge.