During the pandemic, annual CPI inflation in the US rose sharply to a peak of 9% in June 2022. This column argues that the rise of inflation can be explained with a framework in which inflation depends on three factors: long-term expectations, the tightness of the labour market as measured by the vacancy-to-unemployment ratio, and large changes in relative prices in particular industries, such as energy and autos. Furthermore, it shows that the same framework can explain the retreat in inflation since the peak. In 2025, core inflation remains above the Federal Reserve’s target and would require further cooling of the labour market.
After a quarter century of quiescence, US inflation rose sharply during the COVID-19 pandemic. The 12-month CPI inflation rate rose from an average of 2.1% over 2017-2019 to 9.0% in June 2022, alarming economists and the public. Then, inflation started declining and in March 2025 it stood at 2.4%, most of the way back to its pre-pandemic level. A large body of research has studied that experience (see, for example, Benigno and Eggertsson 2023, Bernanke and Blanchard 2024, Bolhuis et al. 2025, Dao et al. 2024, and Giannone and Primiceri 2024).
One contribution to this literature is our analysis in Ball et al. (2022). That paper analysed data through September 2022, when inflation was near its peak, and sought to explain why inflation had risen sharply. We offered an explanation based on three factors: the tight labour market over 2021-2022, as captured by an extremely high ratio of job vacancies to unemployment; sharp rises in relative prices in certain industries, such as energy and autos; and a modest rise in long-term inflation expectations.
In a recent update of that earlier work (Ball et al. 2025), we ask whether the factors that explain the rise in inflation can also explain the fall since late 2022. We conclude that they can. All three drivers of the inflation increase have reversed: the tightness of the labour market has diminished, prices in sectors such as energy and autos have fallen, and long-term inflation expectations have returned to pre-pandemic levels.
Headline and core inflation
The starting point of our analysis is a decomposition of headline inflation into two components: core inflation and headline shocks. We measure core inflation with the weighted median inflation rate from the Federal Reserve Bank of Cleveland, and headline shocks with deviations of headline inflation from the weighted median. We seek to explain core inflation with expected inflation and labour market tightness — the variables in the Phillips curve — and pass-through from headline shocks. The headline shocks themselves arise from large relative-price changes in particular industries.
Figure 1 plots the path of CPI inflation since 2020 and our decomposition. The vertical line indicates September 2022, when we completed our earlier work. We see that the inflation surge over 2021-2022 reflected both rising core inflation and positive headline shocks, and both declining core and negative shocks have contributed to the subsequent disinflation.
Figure 1 CPI inflation: Headline, core, and headline inflation shocks, 2020–2025 (percent)
Note: Core inflation is the weighted median CPI inflation rate from the Federal Reserve Bank of Cleveland. Vertical line indicates September 2022.
Three drivers of the inflation cycle
In conventional macroeconomics, one determinant of the inflation rate is expected inflation. Like a number of other researchers (for example, Hazell et al. 2022), we measure expected inflation with the ten-year-ahead forecast of CPI inflation from the quarterly Survey of Professional Forecasters. Expected inflation rose over 2021 and 2022, reaching 2.95% in 2022Q4, prompting some to warn that it might be the start of a de-anchoring of expectations from the Federal Reserve’s target. Then, in 2023, the rise was sharply reversed, and by 2025Q1 expectations stood at 2.3%, close to their pre-pandemic level. Evidently the Federal Reserve’s tightening and communication over 2022-2023, as well as progress in reducing actual inflation, reassured forecasters that inflation would remain anchored.
Another tenet of macroeconomics is that inflation is influenced by tightness or slack in the labour market. Following recent studies such as Furman and Powell (2021) and Barnichon et al. (2021), and as in our 2022 paper, we measure the tightness of the labour market by the ratio of job vacancies to unemployment (V/U). This ratio rose rapidly over 2021 and peaked at 2.0 in March 2022, far above the average level of 0.6 over 1985-2019, and the highest at any point since 1951 — the first year for which data are available. After the peak, V/U fell steadily until mid-2024 and stood at 1.1 in February 2025 — close to the December 2019 level, if still somewhat higher than the historical average.
Underlying the behaviour of the V/U ratio are the well-known shifts that have occurred in the Beveridge curve relating job vacancies and unemployment. As Figure 2 shows, there was a stable Beveridge curve from 2009 through March 2020, and then a dramatic outward shift at the start of the pandemic. Starting in late 2022, falling vacancies moved the economy downward in the Beveridge space, and by the end of 2024 the economy was back on the pre-pandemic Beveridge curve. This downward movement meant that the V/U ratio fell, reducing inflation, even though the unemployment rate never rose above 4.2%.
Figure 2 The behaviour of the Beveridge curve, 2009–2025 (percent of the labour force)
Note: July 2009 to March 2020 covers the pre-COVID-19 expansion, based on National Bureau of Economic Research (NBER) business cycle dates, and the first month of the COVID-19 era. The figure reports log-linear curves fitted to each period.
Both expected inflation and labour-market tightness influence core inflation. The final factor influencing inflation in our framework is headline inflation shocks — deviations of headline from core arising from sharp changes in sectoral prices. These shocks influence headline inflation directly and can also pass through into future core inflation because they influence costs of production and wage demands throughout the economy.
Our 2022 paper identified three variables that explain most of the positive headline shocks that occurred as inflation rose: changes in the relative prices of energy and auto-related items (each measured as the monthly inflation rate in the sector minus median inflation), and a measure of backlogs of work that captures supply chain disruptions. In our recent update, we find that movements in the same three variables — with opposite signs — also explain most of the negative headline shocks that have contributed to falling inflation since late 2022.
Decomposing the path of inflation
Our central findings are summarised in Figure 3, which decomposes headline inflation into core inflation and headline shocks, with core inflation further decomposed into contributions from inflation expectations, the vacancy-to-unemployment ratio (V/U), and pass-through from headline shocks.
Figure 3 Accounting for the rise and retreat of headline inflation (percentage points)
Note: ‘Pass-through’ denotes pass-through of headline shocks to core inflation. ‘Expectations’ denotes contribution of long-term (SPF) inflation expectations to core inflation. ‘V/U’ denotes contribution of ratio of vacancies to unemployed. Vertical line indicates September 2022.
We see that the rise in 12-month inflation from January 2020 to the peak in June 2022 is explained primarily by positive headline shocks (grey bars), which raised inflation both directly and by passing through into core inflation (gold bars). By early 2023, however, the effects of these shocks had faded away and inflation remained high because of extremely high levels of V/U (red bars).
The decline in inflation since then has reflected falling V/U and negative headline shocks. Movements in energy and auto-related prices have been the most important sources of headline shocks. Based on March 2025 data, core inflation remains above target levels mainly because V/U remains above its historical norm.
These findings come from a framework developed in our 2022 paper to explain the rise in inflation over 2021-2022. In our view, the good fit of the framework in explaining the disinflation since 2022 strengthens the credibility of our approach. In our 2025 paper, we show how the fitted values of inflation from our estimated equations follow actual inflation closely over the entire period from 2020 through 2025 — even when the equation is estimated with a sample ending in September 2022, in which case the fitted values after then are out-of-sample conditional forecasts.
Where is inflation heading?
After analysing recent inflation, it is natural to speculate about the future. In March 2025, 12-month core (median) inflation stood at 3.5%, which exceeds the level consistent with the Federal Reserve’s target. This gap reflects the fact that the vacancy-to-unemployment ratio, at 1.1, is still higher than the historical norm, implying upward pressure on underlying inflation. Reducing inflation to the Fed’s target would — other things equal — require a further cooling of the labour market that pushes V/U closer to its average of 0.6 over 1985-2019.
Since 2023, the inward shift of the Beveridge curve has allowed vacancies to fall, reducing V/U and inflation, without a substantial rise in unemployment. It is possible this benign trend will continue, but that would require the Beveridge curve to shift inward relative to its pre-pandemic position; the reversal of the outward shift over 2020-2022, which has already occurred, is not enough. If the Beveridge curve stabilises at its current position or shifts outward again as a result of new labour market disruptions, a normalisation of V/U will require a movement along the curve, with a significant rise in unemployment as well as a fall in vacancies.
The future could also bring either positive or negative headline shocks. We might, for example, see disinflationary shocks resulting from the recent decline in international energy prices, or inflationary shocks from tariffs and their effects on global supply chains. Unfortunately, our findings suggest that such shocks would have asymmetric effects, because inflationary shocks pass through into core inflation and disinflationary shocks do not. Overall, how soon inflation returns to target will depend on which shocks occur, the evolution of V/U, and whether long-term inflation expectations remain at target-consistent levels.
Source : VOXeu
