June 23, 1999
Most of the difference between the movements of the official Consumer Price Index and the CPI research series can be explained by two sources of adjustments: housing costs and index-number formulas. These sources produced downward adjustments to consumer prices as shown in the chart.
The research series calculates homeowner costs for 1978-82 using a method called "rental equivalence" that was first included in the official CPI in 1983. Extending the method back to 1978-82 in the research series reduced the rate of inflation by 0.86 percentage point in that period. With the rental equivalence method, the homeownership component of the CPI is based on a flow-of-services approach rather than on the current cost of the purchase of a home.
Also, the research series used index-number calculation formulas for 1978 forward that were first adopted in the official CPI during the mid-to-late 1990s. Using the improved formulas in the research series decreased the inflation rate by 0.28 percentage point in 1978-82, 0.26 percentage point in 1983-86, and 0.41 percentage point in 1986-97. All of the other changes incorporated in the research series had a net effect that was small and positive.
CPI data are produced by the BLS Consumer Price Index program. BLS has made numerous improvements to the CPI over the years, which have increased the accuracy of the index; however, the official historical price indexes are not adjusted to reflect the improvements. More information on the CPI research series can be found in "CPI research series using current methods, 1978-98," by Kenneth J. Stewart and Stephen B. Reed, Monthly Labor Review preprint, June 1999. It is important to note that the CPI research series has certain limitations and that it is subject to revision. Growth rates reflect December-to-December changes.