Supply, Demand and the Post-Lockdown Inflation Surge

Economists track inflation in many different ways, such as looking at “core” versus “headline” or monthly versus annual rates. This blog post discusses the advantages of tracking inflation from a different perspective: analyzing individual categories of goods and services via the classic market supply and demand framework.

To start, consider a diagram of an intersecting upward-sloping supply curve with a downward-sloping demand curve. (See the first figure.)

Supply-Demand Diagram for an Equilibrium in a Market for a Good or Service

An outward shift in the demand curve, holding fixed the supply curve, drives up both the price and quantity in a market for a good or service. This increase in both price and quantity will still occur even when the supply curve shifts at the same time, as long as the shift is not too large. On the other hand, an inward shift in the supply curve, holding fixed the demand curve, drives up the market price while decreasing quantity.

With this in mind, it’s natural to label the observed price growth for a good as “positive demand-driven” if the quantity of that good increases. Similarly, when both quantity and price for a good fall, one could label the observed disinflation for that good as “negative demand-driven”; when quantity falls, but price increases, one would label the change “negative supply-driven.” “Positive supply-driven” price-quantity movements then occur when the quantity sold for that good rises, but its price falls.

Supply- and demand-driven price changes might occur for a variety of reasons. For example, bad weather might unexpectedly reduce crop production, or a dockworker strike might prevent some goods from reaching stores.

What Does This Mean for Consumer Inflation Economywide?

We can apply this thinking to classify inflation for various goods and services, such as new autos and financial services, as driven by supply and demand factors. We can then combine supply- and demand-driven inflation across these expenditure categories to infer the portions of overall inflation due, loosely speaking, to supply and demand factors. For example, the classification might imply that in January 2025, a 0.4% increase in the price of new autos would be positive demand-driven inflation if new-auto sales rose. Similarly, a 0.2% increase in the price of financial services would be negative supply-driven inflation if financial services fell in quantity during the same month.

Taking it a step further, if new autos and financial services were the only things consumed and they were consumed in about equal proportions over time, then average economywide consumer inflation would be about 0.3% in that month. Two-thirds of overall inflation would be demand-driven, and one-third would be supply-driven.

Consumption Inflation Decomposition: The Method

Only recently have economists started tracking category-level consumer inflation using their associated movements in quantities. Adam Shapiro, an economist and vice president at the Federal Reserve Bank of San Francisco, used the supply-demand framework described to classify inflation at the consumption-category level into supply- and demand-driven components.Shapiro’s method is slightly more complicated than what we’ve described thus far. He assigns the direction of price and quantity movements based on whether the unanticipated (or unforecastable) change in price and quantity move up or down, instead of whether the actual values move in either direction. See his 2024 article, “Decomposing Supply- and Demand-Driven Inflation,” in the Journal of Money, Credit and Banking.

In our recent working paper, we implemented an extension of Shapiro’s method, which distinguishes between the trend component of inflation and inflation attributable to supply and demand shocks.See Bill Dupor and Marie Hogan’s working paper, “Was the Post-Lockdown Inflation Surge Mainly Supply Driven?” Federal Reserve Bank of St. Louis Working Paper No. 2025-007A, March 2025. We generalize his approach in a few ways.The generalization was developed by Maximiliano Dvorkin, an economic policy advisor at the Federal Reserve Bank of St. Louis, in his work on U.S. labor markets. Most importantly, we distinguish between the current and past effects of those shocks. This allows us to parse the part of inflation that’s expected in the absence of supply- and demand-side “shocks” (the trend), as well as the parts of inflation explained by the ongoing expected effects of shocks in previous periods (past) versus shocks happening right now (current).

Applying the Method to the 2021-22 Inflation Surge

We applied the method to study the inflation surge in 2021 and 2022. We found that the main source of inflation was driven by demand—i.e., categories in which both price and quantities grew faster than expected—rather than by supply.

The second figure plots year-over-year inflation decomposed into five components: trend (light gray), past supply-driven (light blue), current supply-driven (dark blue), past demand-driven (light red), and current demand-driven (dark red).More specifically, we report results for core personal consumption expenditures inflation in the figures. Overlaying the colored bars, the black line shows total inflation.

Prior to the pandemic, annual inflation was slightly below trend (1990-23), which is estimated to be 2%. After inflation dropped sharply during the pandemic’s initial lockdown phase, it began rising sharply in late 2020. As shown in the figure, most of this price growth was demand-driven. Initially, this demand-driven inflation was roughly split between current and past shocks; as inflation remained elevated into late 2021 and 2022, past demand shocks as well as past supply shocks took on larger roles.

Decomposition of Core PCE Inflation by Type of Price Change

SOURCE: Dupor and Hogan (2025).

NOTE: PCE=personal consumption expenditures.

The third figure plots the same information, except as a cumulative change in the price level rather than as a sequence of inflation rates; also, we zoom in on 2021-22 and transform the data by subtracting trend price growth to highlight the contribution of supply and demand shocks. The black line indicates the rise in the adjusted price level based on actual data, and the stacked columns separate the price change into its supply and demand components (both current and past). The figure shows that two-thirds of the price growth was due to demand. If we restrict attention to goods and services with market-based prices (excluding both imputed and nonmarket-based prices), three-quarters of the price growth was due to demand.

Decomposition of Adjusted Core PCE Price Level Path

SOURCE: Dupor and Hogan (2025).

NOTES: Height of stacked column does not equal the height of black line for each month because of higher-order terms included in the black line. The path is adjusted to exclude the trend component of the price level change. PCE=personal consumption expenditures.

One advantage of studying consumption categories is that we can look to see where demand- and supply-driven inflation are coming from. The fourth figure plots the cumulative price growth (net of trend) attributable to supply- and demand-driven inflation, in colors blue and red, respectively, for categories considered in our decomposition between December 2020 and December 2022. We limit the figure to the 10 largest categories in terms of total (supply and demand) cumulative contributions. The black line under each bar shows total price growth net of trend. We weight category-level inflation by share of the total consumption basket; thus, a value of 0.2 means that nontrend price growth in that category contributed to 0.2 percentage points of additional price growth in overall inflation over those two years.

Top 10 Consumption Categories Sorted by Size of Nontrend Contribution to Core PCE Price Level Growth: December 2020 to December 2022

SOURCE: Dupor and Hogan (2025).

NOTE: Data are not annualized. Housing=Imputed rental of owner-occupied nonfarm house.

For a subset of categories, the oft-mentioned supply story for inflation is accurate: High inflation in new and used cars, for example, was largely supply-driven. We also observe strong supply-driven price growth in financial services.  However, these majority supply-driven categories do not make up most of the cumulative price growth in these two years. Rather, several of the categories contributing most to inflation, like household durable goods and meals at restaurants, are entirely or primarily demand-driven. We also observe offsetting supply and demand effects in health care categories like hospital services and doctor visits (not pictured). On net, these health care categories exhibited relatively little price growth above trend, masking a significant negative supply component (prices rising and quantities falling faster than expected) and a positive demand component (prices and quantities rising faster than expected).

Limitations of the Method

The method described here is a powerful tool for tracking price and quantity comovement, as well as recording whether these changes are due to new “shocks” or the lingering expected impact of shocks that occurred in the past.

Still, the insights we learn from the supply-demand decomposition have limitations. First, the method does not tell us what specific fundamental events or shocks, such as technological innovation or changing consumer preferences, drive movements in prices and quantities; instead, it provides a classification of category-level inflation based on the observed comovement. Second, it is not informative about the slopes of market supply or market demand curves. Third, the method does not disentangle real from nominal price changes, which is essential for fully understanding macro fluctuations and informing macro policy.

Notes

1. Shapiro’s method is slightly more complicated than what we’ve described thus far. He assigns the direction of price and quantity movements based on whether the unanticipated (or unforecastable) change in price and quantity move up or down, instead of whether the actual values move in either direction. See his 2024 article, “Decomposing Supply- and Demand-Driven Inflation,” in the Journal of Money, Credit and Banking.
2. See Bill Dupor and Marie Hogan’s working paper, “Was the Post-Lockdown Inflation Surge Mainly Supply Driven?” Federal Reserve Bank of St. Louis Working Paper No. 2025-007A, March 2025.
3. The generalization was developed by Maximiliano Dvorkin, an economic policy advisor at the Federal Reserve Bank of St. Louis, in his work on U.S. labor markets.
4. More specifically, we report results for core personal consumption expenditures inflation in the figures.