Skip Navigation

The growth of the consumer price index published by the Bureau of Labor Statistics is widely used as a measure of the increase in the cost of living. Numerous researchers have argued, however, that changes in the consumer price index do not truly reflect changes in the cost of living, and the results of recent studies have intensified the debate. This Congressional Budget Office (CBO) paper--prepared at the request of the House Committee on the Budget--analyzes the arguments about the degree to which changes in the consumer price index may, on average, overstate the increase in the cost of living--that is, the degree to which the measure may be upwardly biased.

John F. Peterson of CBO's Macroeconomic Analysis Division wrote the paper under the supervision of Robert Dennis. The paper benefited from useful suggestions by Jack E. Triplett and William Randolph. Derek Briggs provided research support.

Sherry Snyder edited the paper, with the assistance of Chris Spoor, and L. Rae Roy and Dorothy Kornegay prepared it for publication.
 

Robert D. Reischauer
Director
October 1994
 
 

SUMMARY

I - INTRODUCTION

II - EMPIRICAL ESTIMATES OF MEASUREMENT BIAS

• Commodity Substitution Effect
• Sample Rotation Bias
• Empirical Studies of the CPI for Prescription Drugs
• Other Empirical Studies of Bias

III -OTHER POTENTIAL SOURCES OF MEASUREMENT BIAS

• New Goods
• Medical Care
• Nonmedical Services
• Special Problems in Selecting a Weight for Medical Care

IV - MEASURING THE COST OF LIVING FOR SUBGROUPS OF THE POPULATION

V - IMPLICATIONS OF MEASUREMENT BIAS

• Effect on Measures of Growth in GDP, Productivity, Wages, and Poverty
• Effect on Federal Outlays and Revenues
• Effect on Monetary Policy

VI - ISSUES RELATED TO REVISING THE CPI

• Plans for Revisions to the CPI and Development of an Experimental Index
• Should There Be Separate Indexes for Measuring the Cost of Living and Inflation?

SUMMARY

As a measure of the cost of living for the overall population, the consumer price index (CPI) is probably upwardly biased--that is, it tends to overstate the increase in the cost of living. Although the amount of bias is not known, the existing empirical evidence, which addresses many but not all of the potential areas of mismeasurement, indicates that the CPI has probably grown faster than the cost of living by between one-fifth and four-fifths of a percentage point in recent years. Other potential areas of mismeasurement that have not been subjected to empirical examination may offset or add to the bias that the empirical studies have found. Because the CPI is widely used as an indicator of both the cost of living and inflation, a bias in that index could distort economic decisions, government spending programs, and tax policies. Because of the importance of this issue, further study is warranted.

Users of cost-of-living indexes should be aware at the outset that numerous theoretical and practical difficulties are associated with measuring the cost of living and that all estimates are subject to a great deal of uncertainty. Even if all of the necessary data were readily available, fundamental problems would arise in deciding what should be included in a cost-of-living index and how best to measure the cost of living. For example, rapid changes in medical care and the potentially huge valuations that could be placed on new treatments or cures make medical care one of the thorniest problems in developing a cost-of-living index. Numerous practical problems also arise: not all of the data are readily available; and an immense effort is needed to ensure that consumers' shopping patterns are reflected in the data samples and that the value to the consumer of new and substitute goods and services is properly estimated.

The use of a price index that is based on the spending habits of one group of people to approximate the changes in the cost of living for another group introduces another potential for mismeasurement. For example, the spending patterns of the elderly are, on average, quite different from those that are used to construct the CPI. Medical care constitutes a greater percentage of the expenditures of the elderly than of the population as a whole; and the medical care component of the CPI, as currently measured, has been increasing more rapidly than the overall CPI. For that reason, an experimental index for the elderly increases more rapidly than the regularly published CPI. The current measure of medical care, however, may be biased upward. In addition, other nonmedical categories of the CPI may suffer from a similar upward bias. If so, the growth of the CPI may, in fact, be similar to or even overstate the true increase in the cost of living for the elderly.

An overestimate of the increase in the cost of living of even half a percentage point, if not generally recognized, may seriously distort private and public economic decisions. Overstatement would obviously mislead decisions about cost-of-living adjustments to wages and other contracts, entitlement programs, and income tax brackets. But overstating the growth in prices could also affect a number of other decisions because, for example, measures of trends in the growth of real gross domestic product, real wages, and productivity would be lower than they should be. In addition, mismeasurement may lead to unintended real transfers of wealth among various groups.

The Bureau of Labor Statistics, the agency that produces the CPI, is well aware of these issues and is examining ways to address them. In fact, the bureau will institute some changes in January 1995 that are likely to reduce the degree of bias.

This document is available in its entirety in PDF.