Title: Measurement Errors and Quality-Adjustment Methodology: Lessons Form the Japanese CPI
Abstract: Introduction and summary The Consumer Price Index (CPI) is widely used as a measure of inflation. However, there is a growing consensus that the CPI substantially overstates changes in the true cost of living. Table 1 summarizes recent studies on upward bias in the CPI of major industrial countries; the effect of this upward bias ranges from 0.5 percentage points to 1.1 percentage points per year. In the case of Japan, the effect is estimated at 0.9 percentage points per year. Upward bias in the CPI arises because the current CPI fails to account for the dynamic nature of economic activity, such as changes in consumers' behavior in response to relative price fluctuations between goods, the introduction of new goods, and the disappearance of old goods. Upward bias in the CPI has a direct implication for monetary policymakers, whose major mandate is to maintain price stability. Although biases in inflation measures do not matter when inflation is high, they do matter when policymakers are considering whether to bring down an already low inflation rate. In this sense, as economies approach price stability, accurate measurement of inflation is especially challenging. The importance of accurate price measurement is apparent in a country like Japan where there is controversy as to whether the country is on the verge of deflation.(1) Without upward bias, the Japanese CPI would have shown even stronger evidence of deflation in recent years. Moreover, accurate price measures are necessary to interpret economic developments, not only involving inflation but also real output and productivity. If measured inflation is rising more rapidly than actual inflation, measured real economic growth is simultaneously being understated. This implies that real incomes and living standards are rising faster than the published data suggest. In addition, the overstatement of inflation creates an automatic and unintended real increase in social security and other indexed federal benefits and a real cut in indexed individual income taxes each year.(2) In examining the problems of price measurement, a distinction must be made between the measurement of individual prices and the aggregation of those prices into the overall price index. Aggregation may introduce biases, because the CPI assumes that households purchase the same basket of goods and services over time, although, in reality, they substitute some goods for others as relative prices change and new goods are introduced. However, the problems of aggregation are well understood by economists, and workable solutions are within reach.(3) The most important remaining problems relate to the measurement of individual prices. Observing and measuring individual prices is quite difficult both conceptually and practically. This is because it is very hard to divide the nominal value of a good into quantity and price on a quality-adjusted basis. As the characteristics of products and services are changing rapidly, it is becoming increasingly difficult to define the unit of output and adjust an item's price for improvements in quality. These problems are pervasive in modern economies. For example, automobiles, refrigerators, TV sets, VCRs, camcorders, personal computers, winter jackets, and sports shoes have all changed in ways that make them surprisingly hard to compare with their counterparts in the past. Indeed, as shown in table 1, quality change/new product bias is identified as the largest source of upward bias in the CPI in major industrial countries. [TABULAR DATA FOR TABLE 1 OMITTED] In this article, I investigate the problems inherent to the quality changes/new goods bias in the CPI, taking the Japanese case as an example. I review the sources of measurement errors in the CPI and examine the problems inherent in the methodology used for quality adjustment in the Japanese CPI. I describe the basic framework of the hedonic approach (methodology to analyze the price-quality relationship by regressing prices on numerous characteristics of a product) and propose a practical way to improve the accuracy of quality adjustment by introducing this approach to the conventional procedure to compile the CPI. …
Publication Year: 1999
Publication Date: 1999-06-22
Language: en
Type: article
Access and Citation
Cited By Count: 18
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot