The National Bureau of Economic Research released a paper, Comparing Past and Present Inflation (No. w30116), by Marijn Bolhuis, Judd Cramer and Lawrence H. Summer on Monday, June 6, 2022. There have been important methodological changes in the Consumer Price Index (CPI) over time. These distort comparisons of inflation from different periods, which have become more prevalent as inflation has risen to 40-year highs. To better contextualize the current run-up in inflation, this paper constructs new historical series for CPI headline and core inflation that are more consistent with current practices and expenditure shares for the post-war period. Using these series, we find that current inflation levels are much closer to past inflation peaks than the official series would suggest. In particular, the rate of core CPI disinflation caused by Volcker-era policies is significantly lower when measured using today’s treatment of housing: only 5 percentage points of decline instead of 11 percentage points in the official CPI statistics. To return to 2 percent core CPI today will thus require nearly the same amount of disinflation as achieved under Chairman Volcker. The full dataset can be found here.
Prior to 1983 measurement of shelter inflation was mechanically responsive to Federal Reserve interest rate policy via its inclusion of mortgage rates. Starting in 1983, home prices no longer enter the CPI, the BLS measures owners’ equivalent rent (OER). To allow for better comparison before and after this change, we constructed two new measures of inflation. In the first, we adjust CPI inflation pre-1983 by estimating (OER) using the CPI rent series. We backcast what we think that measures of OER would have been pre-1983 had the post-1983 method been used. We do this by regressing OER on rental inflation post-1983. Our dataset also includes series that use constant expenditure weights from other time periods to adjust for changing consumption habits. The CPI has become significantly more “sticky” over time, these constant weights allow the analyst to compare similarly transitory measures of inflation from different periods. With today’s expenditure shares, we estimate that the peak of headline CPI inflation in 1951 would have been 3.3 percent, instead of 9.4 percent. Our adjusted peak of core CPI inflation in 1951 is 5 percent, compared to an official peak of 7.2 percent. Constant expenditure shares make the current elevated rate of inflation look more serious relative to past peaks than the official numbers would suggest. Our dataset contains 32 components that cover around 90 percent of the overall CPI since 1946, as listed in Appendix Table 1 of the paper. Figure A.1 in the Appendix plot the inflation rate of components over time.
Methodology
We use public data from the BLS over time to explore the change in the nature of inflation during the post-war period. Our dataset contains 32 components that cover around 90 percent of the overall CPI since 1946, as listed in Appendix Table 1 of the paper. Figure A.1 in the Appendix plot the inflation rate of components over time.
Using our data on inflation rates and weights of the 32 components, we construct two new measures of inflation. First, we replicate the official headline and core CPI inflation rate. To ensure our bottom-up estimate of official headline CPI equals the published series, we add a residual CPI component. This residual component mainly covers recreation and information. We then adjust CPI inflation pre-1983 by estimating OER using the CPI rent series. We backcast what we think that measures of OER would have been pre-1983 had the post-1983 method been used. We do this by regressing OER on rental inflation post-1983. Our dataset also includes series that use constant weights from other time periods to adjust for changing consumption habits.
Citation of the data
Publications and research reports based on this database must cite it appropriately using
Bolhuis, M. A., Cramer, J. N., & Summers, L. H. (2022). Comparing Past and Present Inflation (No. w30116). National Bureau of Economic Research.
For questions on the dataset and paper, email Judd Cramer at judd.cramer@gmail.com or Marijn Bolhuis at mabolhuis@gmail.com.