Department of Labor

Recommendations and Actions

DOL17: Revise and Update the Consumer Price Index


The Consumer Price Index (CPI) is the principal source of information concerning consumer price trends and inflation in the United States. As one of the most important economic indicators, the CPI affects many elements of everyday life.

Both the public and private sectors use the CPI extensively for economic analysis and policy formulation, as well as to escalate contract costs among individuals and organizations. The CPI also affects the finances of the federal government, because it is used to adjust payments to social security recipients; federal and military retirees; and certain entitlement programs, such as food stamps and school lunches. In addition, the CPI is used to adjust individual income tax brackets and personal exemptions for inflationary changes. Each 0.1 percent change in the CPI results in a net change of about $600 million in federal spending and revenues.

It is essential to maintain the accuracy of the CPI, which measures price changes for a fixed market basket of goods and services purchased in a historical base period. Periodic revisions of the CPI are necessary for it to remain an accurate, relevant, and timely measure of inflation. The CPI tends to misstate a true cost-of-living index, however, in part because a fixed market basket does not reflect changes in consumers' choice where relative prices on certain items change differently over time. Because the current market basket for the CPI is based on a 1982-1984 consumer spending pattern, experts believe it is overstating changes in current living costs.

Studies in the economic literature have shown that the introduction of a new market basket could reduce the CPI by 0.1 to 0.2 percent per year.(1) Further, recent research suggests that, had a 1987-1989 market basket been used last year, the CPI would have been about 0.7 percent lower.(2) Assuming that this difference accumulated equally over the 6 years since the last market basket update, the difference would be about 0.1 percent per year.

To ensure the CPI's accuracy, index revisions are scheduled every 10 years, after each decennial census. The revisions incorporate population shifts, changes in consumers' buying habits, improvements in technology and indexing methodology, and redesigned survey questionnaires and computer systems to make the index more accurate, reliable, and timely.

Revision is a multiyear effort, and the current revision had been planned as a six-year task. Work was slated to begin in 1994 with a revised index to be released in January 1998, and a revision to be completed in 1999. However, funding for the CPI revision has been deleted from the budget due to the projected costs of $56 million and the estimated peak of 277 personnel needed for the project's six-year life. Due to the critical role of the measure and its potential long- term positive fiscal impact, the funding to do the revision should be restored.


1. DOL should introduce a new geographic sample and new market baskets for the CPI.

The sample of geographic areas on which the index is based would be reselected to reflect population shifts since the 1980 census. These changes would be incorporated into the CPI, as well as the Consumer Expenditure Survey (CES) and the Point of Purchase Survey (POPS), both of which support the CPI. In addition, the index's expenditure patterns, which currently refer to 1982-1984, would be updated to the period of 1993-1995.

2. DOL should redesign the processing system for the CES.

The system presently used to process CES data was designed and programmed in the late 1970's. Since then, significant improvements in computing capabilities have been achieved, and numerous changes have been made to the CES questionnaires. A new processing system could take advantage of improvements in computing technology and allow efficient program changes and adaptation to future improvements.

3. DOL should revise the housing sample for the CPI.

Because of the changes in geographic sample, the housing sample must reflect the 1990 census data. The housing sample needs to represent accurately the housing cost patterns of owners and renters for the 1990's. Use of pertinent housing data from the Department of Housing and Urban Development may conserve time and resources.

4. DOL should redesign the computer system for processing housing data.

The current computer system for processing housing data was the first major system designed during the last CPI revision. Now it must be redesigned not only because it is the CPI program's oldest computer system, but also because the housing sample must be redesigned. The redesigned computer system would support the CPI more efficiently in this essential area of price measurement.

5. DOL should introduce Computer-Assisted Data Collection (CADC) technology.

Computer technology has advanced since the last revision to the point that collectors can efficiently capture data using small computers in the field and transmit the data electronically to Washington, thereby eliminating data entry processing in Washington. CADC would replace a system in which collected materials are mailed to regional locations and then mailed to Washington for data entry into the computer. The proposed new process, which will include automatic edits built into these small individual computers, would result in greater accuracy by enabling data collectors to correct errors at the source. In addition, the data transmission process would be more reliable and timely.

6. DOL should convert the POPS from personal visit to telephone interview.

The current POPS, used to select stores in which CPI prices are collected, is conducted by personal visit. Research shows that, because of the length of the interview, the survey suffers from high nonresponse levels. Further, testing has demonstrated that the survey can be restructured using telephone interviews in a manner that results in a shorter interview spread over more respondents.(3)

Using a telephone-based survey has three advantages over the current survey method. First, by including more places of business, the results are more comprehensive. Second, it is more flexible, enabling survey adjustment on short notice to replenish samples that experience unusually rapid changes. Finally, it would result in a more balanced workload for data collection; i.e., instead of revising all samples in 20 percent of the geographic areas every year, approximately 20 percent of the products and services priced would be resampled in all geographic areas every year.


The expected decrease in the CPI resulting from the revision and update will affect the finances of the federal government; that is, it will adjust payments to social security recipients, to federal and military retirees, and to recipients of food stamps and school lunches. It will also reduce the indexed deductions allowed to taxpayers. However, the net effect will be to bring these adjustments more nearly in line with actual price changes.

Fiscal Impact

Overstatement of the CPI has a significant detrimental budgetary impact on federal government outlays and receipts. To the extent the CPI does not account for market substitutions, social security and many other benefit adjustments are larger than they need to be.

The precise level of savings cannot be determined, but the impact on the national economy will be dramatic. The Office of Management and Budget (OMB) has recently shown that each 0.1 percent CPI increase in calendar 1993 would trigger an increase in the U.S. deficit of approximately $600 million in fiscal year 1995, with an even larger effect in subsequent years.(4) Thus, for an initial investment of about $56 million spread over six years, there is the potential to avoid costs of $600 million for every 0.1 percent slower rate of increase in the CPI due to the revised market basket. Using this assumption, implementation of the recommendations would result in:

--- $400 million savings in federal entitlement payments (Social Security payments would decrease by $290 million, and payments for all other indexed programs--such as U.S. civilian and military retirement programs--would decrease by $120 million); and

--- $200 million gain in potential federal income tax revenues (tax revenues would increase because IRS personal exemptions and tax brackets are adjusted annually by changes in the CPI).

The following table shows only the costs to implement the study because the exact savings cannot be projected until the study is completed. The cost levels are based on fiscal year 1994 budget estimates.

 Budget Authority (BA) and Outlays (Dollars in Millions) 
 Fiscal Year
        1994     1995     1996     1997     1998     1999     Total
 BA     $0.0     $4.0     $8.0    $15.0    $19.0    $10.0     $56.0

Out- lays $0.0 $4.0 $8.0 $15.0 $19.0 $10.0 $56.0

Change in FTEs 0 23 56 126 277 96 96


1. Braithwait, Steve D., "Substitution Effect in the Lasperyes Price Index," American Economic Review, vol. 70 (March 1980), pp. 64-77; and Manser, Marilyn, and Richard McDonald, "An Analysis of Substitution Bias in the Measurement of Inflation," Econometica, vol. 57 (July 1987), pp. 909-930.

2. BLS research article to be published in a forthcoming issue of the Monthly Labor Review Journal.

3. "Evaluation of the 1988 Current Point of Purchase CATI Feasibility Test" (paper presented at the August 1991 annual meeting of the American Statistical Association), p. 3.

4. Estimates provided by Doug Norwood, U.S. Office of Management and Budget, Economic Assumptions staff.

Return To Report Index

Who We Are |||Latest Additions |||Initiatives |||Customer Service |||News Room |||Accomplishments |||Awards |||"How To" Tools |||Library |||Web Links

Reinvention Comments
Technical Comments