Mismeasurement in the Consumer Price Index: An Evaluation

Abstract

A number of analysts have claimed recently that the Consumer Price Index overstates the annual increase in the cost of living. This paper develops a framework for studying measurement problems in the CPI and systematically analyzes the available evidence concerning the magnitude of these problems. The evidence suggests that the bias is centered on 1.0 percentage point per year. To take into account uncertainty about the extent of the bias, the paper presents a probability distribution for the bias rather than a point estimate or range. It is estimated that there is a 10% chance that the bias is less than 0.6 percentage point and a 10% chance that it is greater than 1.5 percentage points per year. CPI methodology overstates the price increase for medical procedures that are subject to technological improvement. To illustrate this point and to show how better to measure medical care prices, the paper presents a prototypical price index for cataract surgery. This index grows much more slowly than a price index for cataract surgery constructed using the CPI methodology.