Some CPI Component Movements and Their Implications

Dean Baker has an interesting article conjecturing about future declines in the CPI, being driven by gasoline, cars, and food. The argument seems plausible, depending on what happens in the supply chains and the oil markets.  I want to consider what might happen, depending on other components.

First, consider what has driven overall inflation is in large part the surge in durable goods prices.

Figure 1: CPI-all urban (bold black), CPI-durables (light blue), CPI-nondurables (light green), and CPI-services (red), all in logs, 2020M02=0. NBER defined recession dates peak-to-trough shaded gray. Source: BLS (November 2021 release) via FRED, NBER, and author’s calculations. 

Should public health concerns diminish, then as consumption switches back to services, durable goods prices might stabilize and even come down. Remember, supply chain problems involving getting goods from where they’re produced to where they’re demanded are driven in large part by the surge in demand for goods — as opposed to diminished supply capacity at ports, etc. This is a variation on Baker’s argument, and others.

Another area of concern revolves around housing costs, which are unambiguously rising:

Figure 2: CPI-all urban (bold black), CPI-owner occupied equivalent rent (blue), and CPI-rent of primary residence (tan), all in logs, 2020M02=0. NBER defined recession dates peak-to-trough shaded gray. Source: BLS (November 2021 release) via FRED, NBER, and author’s calculations. 

Rent of primary residence  accounts for 7.6% of total weight in October 2021, and owner equivalent rent of residences, 23.5% (see Table 6 in the release). Growth in the former accounted for 1.3 percentage points of the 9.8% m/m CPI inflation rate in November, while growth in rent accounted for 0.4 percentage point.

A reader notes that the accelerated rise in owner occupied equivalent rent is likely to persist.

Average lag time from house price index to cpi owners equivalent rent is 18 months.

YoY inflation is forecast to remain, and likely to even increase, in 2022

The CPI measure of rents is also measured in a way that lags commonly observed rent measures.

Here is a picture of owner occupied equivalent rent (and rent) versus house prices.

Figure 3: CPI-owner occupied equivalent rent (blue, left log scale), and CPI-rent of primary residence (tan, left log scale), Case-Shiller 20-city house price index (green, right log scale). NBER defined recession dates peak-to-trough shaded gray. Source: BLS (November 2021 release) via FRED, S&P via FRED, NBER, and author’s calculations. 

A visual inspection seems to uncover a relationship, particularly in y/y changes (but not in long-run cointegration). At 18 months (using non-overlapping changes), the relationship is 0.08 – that is a one percentage point increase in the y/y growth rate leads to a 0.08 percentage point increase in the OER component growth rate. Taken literally, the September reading of 17.5% y/y change (log terms) implies that in March 2023, OER inflation will be 1.4 percentage points higher as a consequence. Since I have little faith in the regression (given some of the diagnostics), I’m not going to bet on this number.