Inflation Dynamics during COVID-19.

• Author: Cavallo, Alberto
• Position: Research Summaries

For over a decade, my research has explored the use of high-frequency microdata to measure inflation and other economic statistics in real time in an effort to make academic macroeconomic research more timely and useful for policymakers. The COVID-19 pandemic has provided a unique opportunity to test this methodology, particularly around the topic of inflation. After the crisis started, the United States experienced a relatively small decline in inflation in 2020, followed by a sudden surge in prices in early 2021. Understanding these inflation dynamics has been the focus of my recent papers. In particular, I studied the impact of COVID-related consumer price index (CPI) measurement distortions and supply disruptions, both of which can be quantified with new sources of high-frequency microdata.

Measurement Distortions with CPI Basket Weights

Early in the pandemic, W. Erwin Diewert and Kevin J. Fox warned that the standard fixed-basket-of-goods approach used by CPIs, which relies on category weights updated infrequently with lagged expenditure data, could introduce significant measurement bias. (1) In the US, CPI weights were last updated in December 2019, shortly before the pandemic dramatically changed consumer spending patterns. To quantify the extent of this basket bias, I relied on the work of Raj Chetty, John N. Friedman, Nathaniel Hendren, Michael Stepner, and the Opportunity Insights Team, who publish real-time consumer spending patterns based on credit and debit card transactions. (2) Using their high-frequency data, I updated the official CPI basket weights on a monthly basis, and combined them with CPIs for various categories of goods to compute an alternative index that can approximate the inflation rate of a COVID-adjusted consumption basket. (3)

With this "COVID CPI," I found that the annual CPI inflation rate in the US was significantly underestimated in 2020, as shown in Figure 1. This was because there was too much weight given to Transportation, a category initially experiencing deflation, and too little weight on Food at Home, where prices were rising. The distortion extended to the core CPI because non-energy transportation categories were impacted, as well as to the PCE (Personal Consumption Expenditures) Price Index, which has a chaining methodology that only partially adjusted to the sudden changes in consumer expenditure patterns.

The CPI basket distortion is temporary, but it is still affecting the annual inflation rate in mid-2021. In fact, US annual inflation is now being overestimated by about 0.7 percentage points. There are two reasons for this. First, the CPI was still placing too much weight on Transportation during the first half of 2021, a time when CPI categories such as Used Cars and Trucks experienced particularly strong increases in...