A Survey of Measurement Biases in Price Indexes

Abstract

This paper reviews the literature on measurement error in the major US price indexes -the Consumer Price Index (CPI), the Producer Price Index (RPI), and the Gross Domestic Product (GDP) deflators. We take as our point of departure Triplett's, 1975. survey and focus on the studies of measurement error that have appeared since then. We review the problems of substitution bias, quality bias, new goods bias, and outlet substitution bias that are generally considered to be the main sources of error in price indexes. The bulk of the paper is devoted to problems in the CPI and PPI, as the GDP deflators tend to be based mainly on the components of these series. We find that there has been surprisingly little work on the problem of overall measurement error in any of these price indexes, and we conclude that there is very little scientific basis for the commonly accepted notion that measured inflation at 2 to 3 percent a year is consistent with price stability.

Measurement Biases in Consumer Price Indexes

“…A number of studies on the magnitudes of the quality bias for certain product categories are found in Wynne & Sigalla (1996). Comprehensive estimates of quality bias for Canada are limited: Crawford (1993) puts an upper limit on quality bias of 0.2% caused by consumer durables, basing his estimate on those found in Gordon (1990) for the U.S.; Fortin (1990) does not believe the quality bias problem to be serious for the Canadian CPI, though there is really no reason to suspect that the Canadian CPI suffers less quality bias than the U.S. CPI.…”

Measurement Bias in the HICP: What do we know and What do we need to know?

“…It has long been argued by practitioners that the difficulty of disentangling changes in 'fashion' from changes in quality in the apparel component of a consumer price index make it as likely that the consumer price index understates inflation as overstates inflation. Wynne and Sigalla (1996) cite an example pointed out by Jack Triplett in an unpublished conference paper wherein he noted that the infant's and toddler's component of the US CPI (which was presumably less influenced by fashion cycles) showed a much more rapid rate of increase than the men's or women's components. 45 The classification of goods and services in the HICP does not distinguish between men's, women's and children's apparel, but it does distinguish between 'clothing materials' (COICCOP/HICP code 03.1.1), 'garments' (COICOP/HICP code 03.1.2) and 'other articles of clothing/clothing accessory' (COICOP/HICP code 03.1.3).…”

Should the Cost‐of‐living Index Provide the Conceptual Framework for a Consumer Price Index?

“…In the United States, the central bank publicised shortcomings of the CPI as a COL index. Reviews by central and regional Federal Reserve staffs adopted the COL framework (Wynne and Sigalla, 1993;Lebow et al, 1994;Shapiro and Wilcox, 1996). None suggested in any way that the COL framework was inappropriate for monetary policy.…”