The Consumer Price Index (CPI) and the unemployment rate are the two most important economic statistics produced by the Bureau of Labor Statistics (BLS) in the U.S. Department of Labor. The CPI--a weighted measure of the average monthly change in the price of a fixed market basket of goods and services2 --is important because it is a central target of monetary policy, and because changes in the CPI will significantly affect government data and the public policy debate surrounding such issues as real wage and income growth.3 For example, overstating inflation by 1.1 percentage points per year from 1973 to 1995 would mean that real inflation-adjusted hourly wages have actually increased 13 percent instead of falling by 13 percent as is currently reported.4
An inaccurate CPI has major implications for the federal budget. In the current fiscal year, approximately 30 percent of total federal spending is indexed to movements in consumer prices. Social Security, Supplemental Security Income (SSI), veterans pensions, military retirement, and civilian pensions account for the bulk of this spending. Tax receipts are also affected, since individual income tax brackets, personal exemptions, and the standard deduction are adjusted--that is, indexed--according to the CPI. Individual income taxes account for about 45 percent of federal receipts.
There is a widespread consensus that the CPI significantly overstates the rate of increase in the cost of living. This problem received a great deal of attention following the release of the Boskin Commission Report--the Final Report of the Advisory Commission to Study the Consumer Price Index--in 1996. The Boskin Commission Report concluded that the CPI overstates inflation by 1.1 percent a year.5 In addition, the weight of evidence emanating from the academic community, the Federal Reserve, the Congressional Budget Office (CBO), and even the BLS that the CPI has an upward bias is overwhelming.6 The upward bias has, and will continue to have, significant implications for public policy decisions. Even after the next set of revisions in the CPI is completed in 1999, significant upward bias will remain.
The significant overestimate of the increase in the cost of living results because of a variety of methodological problems that occur when the CPI is calculated.7 To its credit, the BLS has a program of research and development activities aimed at improving the accuracy of the CPI. And since 1990, the BLS has implemented a number of improvements in addition to the soon-to-be-completed six-year CPI revision program.8 The combined effect of all these improvements is expected to reduce the rate of increase in the CPI by 0.7 percentage points per year compared to the early 1990s.9 These changes alone will raise federal revenues by $35.4 billion and reduce federal spending and interest payments by $61.6 billion between now and fiscal year (FY) 2003 (see Chart 1).10
The increased focus on the CPI that followed the release of the Boskin Commission Report caused the BLS to begin a new multiyear CPI Improvement Initiative in 1998. The BLS received $2.1 million for the project in its first year; it has asked for $9.1 million for FY 1999. According to the BLS, these funds will expand the Consumer Expenditure Survey to facilitate more timely CPI market basket updates and the production of an official superlative index that is more accurate than the current CPI. The new funding will also enable the BLS to expand the use of quality adjustment methods.11 These are positive--but long overdue--steps that will further minimize bias in the CPI.
How to Improve the CPI
Improving the Consumer Price Index will involve both the executive branch and Congress. Research and development activities that address the existing methodological problems must be undertaken by those in the executive branch who have relevant expertise. At the same time, by exercising its oversight responsibilities, Congress should ensure that the federal resources dedicated to such an effort are used effectively to achieve the desired goal of improving the quality of the CPI. The measures that can be taken include the following:
-
The Bureau of Labor Statistics should establish a new index, to be published and updated annually and revised historically as far back as possible. This index should take into account all of the changes made in the CPI since 1990 and any future improvements that are made in the official CPI. It could then be used much as the current "CPI-U-X1" index (which measures housing costs the same way from 1967 to the present) is used--as a deflator in economic research.12
-
The President and the Secretary of Labor have authority to direct the BLS to correct the CPI as part of the current improvement initiative. One of the Secretary's primary goals in improving the CPI should be to create an index that more accurately reflects changes in the cost of living. A perfect cost-of-living index is not possible in a complex economy, but the BLS should continually strive for that goal.
-
The Secretary of Labor should establish a permanent and rotating independent commission of experts to review progress, conduct research, and make recommendations for further improving the CPI. The commission should be appointed by either the National Bureau of Economic Research or the National Academy of Sciences and should report to Congress once a year. If the Secretary of Labor does not establish such a commission this year, Congress should implement one through the appropriation process with the reprioritization of existing resources. The 30-year span between the last two major commissions--the Stilger and Boskin Commissions--was far too long for comprehensive evaluation of the CPI. The BLS needs a more permanent, ongoing mechanism for bringing in outside information, expertise, and research to improve the CPI on a more timely basis.
-
Congress should establish an oversight process and appropriate additional funding for the CPI improvement initiative and the permanent commission from within the existing discretionary budget. Even though legislative action, such as modifying the inflation adjustment for Social Security benefits,13 is not necessary at this time, congressional oversight is critical. Congress should establish an annual oversight process that includes a General Accounting Office (GAO) review of how well the BLS strategic plan and performance plan meet the criteria in the Government Performance and Results Act of 1993 (GPRA). The most recent GAO report found that the BLS plan "did not fully portray how BLS' strategies and resources would help achieve the performance goals for improving CPI quality."14 Annual oversight hearings should be held to determine whether the BLS is fulfilling its GPRA requirements.
Conclusion
A major problem with the federal budget is not the Consumer Price Index itself, but how the CPI is used. Although Congress explicitly intended to insulate program beneficiaries and taxpayers automatically from the effects of inflation, it chose a mechanical process that does not accurately account for changes in the true cost of living. The Secretary of Labor and the BLS can and should make improvements in the CPI, including striving to achieve the goal of developing a more realistic cost of living index. At the same time, Congress must conduct effective oversight and make it a priority to keep such an index up to date. Even if the upward bias in the CPI is only a fraction of a percentage point per year, the relentless compounding of such a bias ultimately will have significant budgetary consequences.
D. Mark Wilson is the former Labor Economist at The Heritage Foundation.
Endnotes
1. This paper is based on testimony given by the author before the Subcommittee on Human Resources, Committee on Government Reform and Oversight, U.S. House of Representatives, on April 29, 1998.
2. Despite the best efforts of the BLS, many people mistakenly refer to the CPI as a "cost-of-living" index. As BLS-published descriptions make clear, the CPI as currently constructed will tend to rise more rapidly than a true cost-of-living measure.
3. The CPI also affects, among other things, the real rate of economic growth, productivity, and the poverty rate.
4. Michael J. Boskin, "Prisoners of Faulty Statistics," The Wall Street Journal, December 5, 1996, p. A20.
5. Final Report of the Advisory Commission to Study the Consumer Price Index, Toward a More Accurate Measure of the Cost of Living, Committee on Finance, U.S. Senate, 104th Cong., 2nd Sess., December 5, 1996. Also known as the Boskin Commission after Chairman Michael J. Boskin.
6. Michael J. Boskin, Ellen R. Dulberger, Robert J. Gordon, Zvi Griliches, and Dale W. Jorgenson, "Consumer Prices, the Consumer Price Index and the Cost of Living," The Journal of Economic Perspectives, Vol. 12, No. 1 (Winter 1998).
7. The four major problems are (1) lower-level and upper-level product substitution bias (when the price of beef goes up, people buy more chicken); (2) an outlet substitution bias (as prices rise, consumers may switch to new discount stores not sampled by the BLS); (3) a new product bias (new products are not introduced into the CPI, or are included only with a long lag time); (4) a quality change bias (the price increases related to medical services are not adjusted for increases in the quality of care). There is also a temporal substitution bias (consumers purchase items on the weekends when products are on sale).
8. The BLS has revised its methods for making seasonal adjustments; introduced procedures to capture the effect of generic drugs in reducing prices; changed the way shelter prices are estimated for renters and homeowners; corrected a flaw in the weights attached to new items in the CPI sample; and reclassified the services provided to hospitals to capture patterns of treatment better.
10. Heritage Foundation calculation based on Scott A. Hodge, ed., Balancing America's Budget: Ending the Era of Big Government (Washington, D.C.: The Heritage Foundation, 1997). Federal revenues rise because personal income tax brackets will be adjusted upward by a smaller amount. Federal outlays are reduced because Social Security benefits will not rise as fast, and interest payments on the debt will be smaller because the surplus will be larger.
11. Brent R. Moulton, "Bias in the Consumer Price Index: What Is the Evidence?" Bureau of Labor Statistics, Working Paper No. 294, October 1996, pp. 18-21; available on the Internet at http://stats.bls.gov/.
12. Beginning in 1983, the BLS changed the way it estimated housing costs in the official CPI, but it did not historically revise the official index. Instead, the BLS published a new index, the CPI-U-X1, which measures housing costs the same way from 1967 to the present. Economic analysts prefer the CPI-U-X1 to the official index because of its consistent treatment of housing costs over time. The CPI-U-X1 is published along with the official CPI in the Economic Report of the President.
