Marijan A. Bolhuis, Jude NL Kramer, Lawrence H. Summers 22 June 2022
The monthly Consumer Price Index (CPI) has been closely scrutinized as concerns over inflation have risen in the United States. The CPI rose 8.6% in the twelve months ending May, the highest annual growth since 1981. This may not be the highest because traders now expect this rate to be around 9% in the next four months.1 An alarming number, it will still be well below the official March 1980 high of 14.8%. To reflect on the dynamics and monetary policy of today’s tight inflation, economists are re-examining the inflation of the Vulcar era and previous inflation (Reis 2021, DeLong 2022, Ha et al. 2022). In a recent study (Bolhuis et al. 2022b), we point out important differences in measurement and warn against rapid comparisons. The way homeownership costs were measured in the headline CPI before January 1983 shows the highest CPI inflation in the past, especially in the Vulcar-era, artificially high. Using today’s methods, CPI estimates prior to 1983 show that current inflation, especially core inflation, is much closer to the previous peak than it appears at first glance. Joining the method also reveals that the previous measure of shelter inflation was mechanically responsive to the Federal Reserve’s interest rate policy. Therefore, the plans to reduce inflation cannot be based on the CPI’s pre-1983 period.
Although historical CPI inflation bears more resemblance to the following adjustment today, we emphasize some important differences with the Vulcar-era. Today’s problems are more manageable for a number of reasons. Most importantly, long-term inflation expectations reached close to double digits in the late 1970s (Hazell et al. 2020). Current expectations are well anchored (Krugman 2022), which suggests that the faster the inflation problem is solved, the easier it can be. In addition, the main problem we highlight is that it does not affect the Federal Reserve’s preferred measure of inflation, the Personal Expenditure Expenditure (PCE) Price Index (Wilcox 2022). The lower part of the housing in the PCE should allow for a quick return to the trend (Ballhuis et al. 2022a).
Measuring Inflation: The Housing Introduction
Housing is both a cost well and an investment. Between 1953 and 1983, the Bureau of Labor Statistics (BLS) assessed the cost of home ownership for CPI without distinguishing between the two. It has created a measure that broadly captures homeowners costs, changes in home prices, mortgage interest rates, property taxes, and maintenance costs as insurance and input. Much higher than rent, this estimate was responsive to changes in interest rates and home price fluctuations. This method created a series of unstable shelters prior to 1983, as shown in the top panel of Figure 1. This inaccurate measurement in December 1982 found a complete 26.1% by weight of the headline CPI. The costs of these shelters were also directly affected by Fed policy, due to the impact of the federal funds rate on mortgage rates. During the tight cycles of 1967-1969, 1972-1974 and 1977-1981, shelter inflation rose sharply, falling sharply only when the speed of hardening decreased.
Figure 1 Shelter CPI inflation and federal fund rate, current since 1954
Formula: Bureau of Labor Statistics.
Comments: Percentage change from 12 months ago. Changes to the OER began in January 1983.
In 1983, after ten years of internal controversy, BLS exchanged ownership costs for homeowners’ equivalent rent (OER). By estimating what a homeowner will get for their home in the rental market, BLS has removed the aspect of housing investment to discourage owner-occupants from using residential services. Since this 1983 shift, the shelter CPI has been much less volatile and much more correlated with the rental CPI (bottom panel of Figure 1). Using publicly available BLS data for the post-war period, we generate new estimates of CPI headlines and core inflation that can be compared over time.
Our dataset contains 32 components that cover about 90% of the overall CPI since 1946. We make three measurements of inflation, presented in Figure 2. First, we replicate the official headline and the original CPI inflation.2 We then adjusted the CPI inflation before 1983 by estimating the OER using the CPI rental series.3 Finally, to assess the importance of differentiating between the volatility of CPI components for overall inflation, we created a second version of the approximate CPI series that uses 2022 weights over the entire period and aligns pre-1983 data with approximate OER.
Figure 2 Official and approximate title CPI
Formula: Bureau of Labor Statistics, author’s calculations
Comments: Percentage change from 12 months ago. Left: Home Ownership Costs Approximate OER pre-1983 replaced. Right: The cost of home ownership is replaced by the estimated OER pre-1983 and the weight of the quantity is fixed at the 2022 level.
Comparison of the inflation cycle
Our estimates suggest that the current rate of inflation is closer to the top of the cycle than other official CPI data. In Table 1, we use our estimates to compare today’s inflation dynamics with the three inflation cycles where the annual growth rate of the official headline CPI has risen above the current reading of 8.6%. The Volker-era inflation peak (March 1980) is currently estimated at 14.8%, when adjusted for conversion from home ownership costs to OER only 11.6%. When measured using the OER method, the core CPI growth in the same month dropped from 12.5% to 6.7%. These large differences reflect both the substantial weight of the OER on the index, especially the core CPI, and the lower peaks of the estimated OER compared to home ownership costs.
Table 1 Past inflation cycle and today’s
Formula: Bureau of Labor Statistics, author’s calculations
Comments: We define ‘start’ and ‘end’ as the local minimum for the official annual headline CPI growth at the beginning and end of each cycle. We define ‘peak’ as the official annual headline CPI for each cycle and the local maximum of core CPI growth. For the official Core CPI, we use the CPI Low Food Series for the period prior to 1958. Today’s basis: Home ownership costs are replaced with the estimated OER prior to 1983. Today’s basis and weight: Home ownership costs are replaced by the approximate OER pre-1983 and the amount of weight is fixed at the 2022 level. A start: July 1949. Top: February 1951 for the title and June 1951 for the original. End: October 1954. b Start: June 1972. Top: December 1974 for the title and February 1975 for the original. End: December 1976. Beginning: April 1978. Top: March 1980 for the title and June 1980 for the original. End: July 1983. d Start: May 2020. Top: March 2022 for both title and original.
Our estimates further indicate that the past inflationary cycle looks more volatile than today due to the larger weight of transitory commodity components in past measurements. In the early 1950s, for example, food and clothing headlines accounted for about 50% of the CPI index. Headline inflation has dropped to 9% to 2% in one year, after the growth rate of these elements went above 10% due to the Korean War-induced deficit. Today, though, food and clothing only gets 17% of the headline CPI weight. The index bears witness to this shift from transit commodity components to less volatile services, increasing the weight of the ‘sticky’ industry in CPI across the board (Bryan and Meyer 2010). The current approach makes housing inflation particularly sticky (Bolhuis et al. 2022a). Such changes in the measurement method complicate the comparison between CPI rates.
We note again that the Federal Reserve’s preferred measure of inflation since 2000 is the PCE Price Index, which has consistently measured rental inflation and does not suffer from the housing problem we examine. Despite its shortcomings, the BLS recently reported that more than 2 million workers were under collective bargaining agreements that linked their wages to the CPI. The CPI index also affects the incomes of about 80 million people due to statutory measures: 47.8 million Social Security beneficiaries, about 4.1 million military and federal civil service retirees and survivors, and about 22.4 million food stamp recipients. The CPI is also used as an input for countless other contracts in the United States that will touch almost every American household. Thus, slower than expected decline in CPI may become self-strengthening in CPI and PCE as it takes more time to get out of the inflation system. In addition, we note that the public and journalists first look to the CPI for the current state of inflation. We leave out the role of measured CPI and PCE inflation in shaping long-term inflation expectations for future exploration, but caution against completely ignoring the CPI.
Some ways to debate inflation today
Previous inflation cycles appear to be more volatile and responsive to Fed policy due to differences in how housing inflation was measured before and after 1983. We are highlighting three effects. First, our observations indicate that the current inflation regime is close enough to that of the late 1970s, if the inflation rate is compared directly. Our revised estimates indicate that core CPI inflation is now at the 1979 experience level. Second, we conclude that it was mistakenly shifted to collect the effects of the financial titan from the CPI disinflation published by Paul Volker in the early 1980s. Volcanic rate of core CPI deflation is significantly slower when measuring using today’s method. At the pace achieved in the 1980s, using these revised estimates, it would take about three years for today’s headline CPI to return to 2%. Finally, many commentators say that monetary policy works through housing – as it was in the past. Once the rate stabilizes, shelter inflation is mechanically submerged. Today, in view of the rearranged structure of the sheltered inflation discussed elsewhere (Bolhuis et al. 2022a), residential services can be a significant barrier to headline and especially core inflation. This could be a headache for the Fed even after housing prices stop doubling.
Author’s Note: Monthly estimates are publicly available here. Please specify using NBER working paper. The opinions expressed herein do not necessarily reflect the views of the authors and the IMF, its executive board, or the IMF management.
Bolhuis, MA, JN Cramer and LH Summers (2022a), “Upcoming Rise in Residential Inflation”, NBER Working Paper No. w29795.
Bolhuis, MA, Cramer, JN, & Summers, LH (2022b), “Comparing Past and Present Inflation”, NBER Working Paper No. w30116.
Bryan, M and B Meyer (2010), “Are some prices higher in CPI than others? We think so “, Economic Comments (2010-02).
DeLong, BJ (2022), “America’s Macroeconomic Outlook”, Project Syndicate, 25 March.
Ha, J, MA Kose, H Matsuoka, U Panizza and D Vorisek (2022), “Anchoring Inflation Expectations in Emerging and Developing Economies”, VoxEU.org, 14 July.
Hazell, J, J Herreno, E Nakamura and J Steinsson (2020), “The Slop of the Phillips Curve: Evidence from US States”, NBER Working Paper No. w28005.
Krugman, P (2022), “Wonking Out: How Low Must Must Inflation Go?”, New York TimesJune 3.
Reis, R (2021), “Loss of Inflation Anchor”, Brookings Papers on Economic Activity.
Wilcox, D (2022), “US INSIGHT: Summers Is Right on CPI. Happily, PCE is the Fed’s focus “, Bloomberg Intelligence, June 8.
1 https://www.marketwatch.com/story/us-inflation-expected-to-keep-running-hot-traders-see-4-straight-months-of-roughly-9-or-higher-cpi-readings -11655399963
2 To ensure that our bottom-up estimates of the official CPI are equal to the published series, we add a residual CPI component.
3 We ‘backcast’ what we think is a measure of OER before 1983, if the post-1983 method was used. We do this by regressing OER on rental inflation since 1983. From 1979 to 1983, we tested our estimates against the previous series of BLS that consistently measured CPI inflation using current methods.