FEDS Note: The Effect of Oil Prices on Core Inflation: Recent Evidence from the Advanced Economies

FEDS Notes

November 09, 2026

The Effect of Oil Prices on Core Inflation: Recent Evidence from the Advanced Economies

Andrea De Michelis, Musa Orak, Avani Pradhan, and Albert Queralto1

Elevated oil prices and heightened geopolitical uncertainty have renewed interest in a longstanding policy question: do supply shocks like oil-price increases pose a transitory or a persistent threat to price stability? This note sheds light on this question by providing new evidence on whether the inflation process has changed. In an earlier note, De Michelis et al. (2024) documented that core inflation had become more inertial after the pandemic, implying that supply shocks might be expected to have more persistent effects if this greater inertia continued. This note updates those results using a similar methodology with data extended through the middle of this year and documents a second change in the inflation process: an increase in the pass-through from oil prices to core inflation.

Because the post-pandemic period remains short, we estimate the model using a panel of five major advanced economies rather than data from a single country, which allows us to detect common shifts in persistence and pass-through more robustly. We find that the pass-through from oil prices to core inflation, which was statistically indistinguishable from zero before the pandemic, rose to an economically and statistically significant level afterwards. A horse race among nested specifications shows that allowing for stronger post-pandemic pass-through, on top of stronger persistence, delivers a marked improvement in our model's ability to track core inflation dynamics in recent years. Our estimated model implies that the recent rise in oil prices could significantly raise core inflation in the advanced economies over the next few years were the greater persistence and higher pass-through to continue. Finally, we use a structural global macroeconomic model to assess the broader implications of the changes in the inflation process, focusing on how these changes bear on the appropriate monetary policy response to supply shocks.

Admittedly, our empirical results are derived using a reduced-form approach that cannot distinguish a structural change in the underlying inflation process from the possibility that global supply shocks have become more frequent, more persistent, or both. The results presented here depend on our specification and set of controls, and different specifications might yield different results. Relatedly, while our panel approach expands the information set by pooling together data across the major advanced economies, it does not allow us to control for some specific factors that have seemingly contributed to persistently elevated inflation in the U.S. because of the lack of comparable data in other countries. In addition, even if our results stem from a structural change reflecting, for instance, a shift in behavior of price setters, we cannot say whether this change will endure or reverse. These caveats notwithstanding, we view our results as suggestive of an upside risk to inflation. We also stress that the results shown here do not represent any official forecasts of inflation in the U.S. or foreign economies.

Empirical framework: an augmented Phillips curve for core inflation

Our empirical strategy follows De Michelis et al. (2024) and relies on a backward-looking augmented Phillips curve for core inflation. In this reduced-form framework, inflation expectations depend on past inflation outcomes. More specifically, core inflation follows:

$$$$ \pi_{i,t} = \alpha_{i} + \gamma_{1}\pi_{i,t-1} + \gamma_{2}\pi_{i,t-2} + \lambda x_{i,t} + \beta s_{t} + \sum_{j=0}^{4} \delta_{j} o_{i,t-j} + \varepsilon_{i,t} $$$$

where $$\pi_{i,t}$$ is core inflation in economy $$i$$ in quarter $$t$$, $$x_{i,t}$$ is the output gap, $$s_{t}$$ is a measure of supply-chain disruptions, and $$o_{i,t}$$ is relative oil-price inflation.2 The specification includes four lags of oil-price inflation because energy price changes take time to pass through to core prices, and it controls for supply-chain disruptions through the Global Supply Chain Pressure Index (GSCPI).3 The sum $$\gamma_{1} + \gamma_{2}$$ measures the degree of inertia, or persistence, in the inflation process, while the sum of the $$\delta_{j}$$ coefficients measures the cumulative pass-through from oil prices to core inflation.

To let the inflation process change after the pandemic, we augment this equation with two sets of interaction terms, one on the lags of core inflation and one on the lags of oil-price inflation, each interacted with a post-pandemic dummy $$D_{t}^{\text{post}}$$ that equals one from 2021:Q1 onward and zero before:

$$$$ \pi_{i,t} = \alpha_{i} + \gamma_{1}\pi_{i,t-1} + \gamma_{2}\pi_{i,t-2} + \eta_{1}\pi_{i,t-1}D_{t}^{\text{post}} + \eta_{2}\pi_{i,t-2}D_{t}^{\text{post}} \\ + \lambda x_{i,t} + \beta s_{t} + \sum_{j=0}^{4} \delta_{j} o_{i,t-j} + \sum_{j=0}^{4} \zeta_{j} o_{i,t-j}D_{t}^{\text{post}} + \varepsilon_{i,t} $$$$

Here the persistence interaction terms $$\eta_{1}$$ and $$\eta_{2}$$ allow inflation inertia to shift after the pandemic, so that post-pandemic persistence is given by $$\gamma_{1} + \gamma_{2} + \eta_{1} + \eta_{2}$$. Symmetrically, the pass-through interaction terms $$\zeta_{j}$$ allow the response of core inflation to oil prices to change, so that post-pandemic pass-through is given by $$\sum_{j} {\delta }_{j} + \sum_{j} {\zeta }_{j}$$. A positive and significant sum of the $$\zeta_{j}$$ would indicate that a given oil-price shock transmits more strongly to core inflation than it did before the pandemic. This single specification therefore nests the two distinct questions we ask: has the core inflation process become more persistent, and has oil-price pass-through become stronger?

To answer these two questions, we estimate the equation on panel data for five major advanced economies: Canada, the euro area, Japan, the United Kingdom, and the United States, which improves our ability to more precisely identify common post-pandemic changes over a relatively short sample.4 The panel regression includes country fixed effects to absorb time-invariant differences across countries. The sample period spans from 2000:Q1 through 2026:Q2, with the post-pandemic period starting in 2021:Q1. Core inflation is measured using core PCE prices for the U.S. and core CPI prices for the other economies; oil inflation is measured from quarterly Brent prices in local currency and enters the model relative to core inflation.5 Supply-chain pressures are measured by the Global Supply Chain Pressure Index of the Federal Reserve Bank of New York. The estimates of the output gap—measured as percent of potential GDP—come from the Bank of Canada, the Federal Reserve Bank of New York (euro area), the Cabinet Office of Japan, Haver Analytics (U.K.), and the Congressional Budget Office (U.S.).

How well do the models track post-pandemic core inflation?

Before examining the estimated coefficients, it is useful to assess how much each of the additional post-pandemic effects improves the model's ability to track the data. Figure 1 plots four-quarter core inflation for the panel of advanced economies against the in-sample predictions of three progressively richer specifications, each of which is a restriction of the equation above:

- No post-pandemic dummies (dashed black line). The most restrictive case, with constant coefficients and no dummies for the post-pandemic period ($$\eta = \zeta = 0 $$).

- Stronger persistence (red line). Interacts the post-pandemic dummies with persistence coefficients but not with the pass-through coefficients ($$\zeta = 0$$).

- Stronger persistence and pass-through (blue line). The full specification, with interactions with both persistence and pass-through coefficients.

Of course, the specifications are nearly indistinguishable before 2021, when the post-pandemic dummies and interaction terms are inactive by construction. The horse race is therefore decided over the post-pandemic period. There, the differences are stark. The specification without post-pandemic interactions (the black dashed line) captures the broad contour of the inflation surge but noticeably underpredicts its peak and subsequent stickiness. Allowing persistence to rise (the red line) closes part of the gap, especially during the decline from the peak, when the model predicts inflation remaining elevated for longer. Adding stronger pass-through as well (the blue line) tracks both the climb and the peak of the surge most closely.

Table 1 makes this comparison precise by reporting the root-mean squared errors (RMSEs) of each specification relative to actual core inflation, both over the full charted sample and over the post-pandemic period. The results are consistent across the two windows: each additional layer of post-pandemic flexibility lowers the RMSE, and the full specification performs best by a wide margin. Over the post-pandemic period, moving from constant coefficients to stronger persistence and pass-through lowers the RMSE from 0.37 to 0.17, with the single largest improvement in fit coming from the pass-through interaction.

Table 1. Root-mean squared errors of core-inflation estimates by specification

Note: RMSE measures forecast accuracy against realized four-quarter advanced-economy core inflation, calculated as the simple average across five advanced economies: Canada, euro area, Japan, the U.K., and the U.S.

Source: Authors' calculations.

Rising persistence and stronger oil pass-through

Table 2 reports the panel estimates from the full specification. The persistence results confirm and extend De Michelis et al. (2024), which uses a similar framework. Pre-pandemic persistence—the sum of the coefficients on the first two lags of core inflation—is estimated at 0.35 and is highly significant. The post-pandemic increase in persistence is 0.30, also highly significant, implying that persistence roughly doubles to about 0.65 from 2021:Q1 onward. The result that core inflation has become substantially more inertial is therefore robust to extending the sample through 2026:Q2 and to allowing oil-price pass-through to change at the same time.

Table 2. Changes in persistence and pass-through coefficients

Note: Persistence is the sum of the coefficients on the first two lags of advanced-economy core inflation estimated from the panel regression. Pass-through is the sum of the coefficients on current and lagged relative oil-price inflation. The post-pandemic increases are the sums of the corresponding interaction coefficients. The lower bound and upper bound report 95 percent confidence intervals. *** p < 0.01, ** p < 0.05.

Source: Haver Analytics; Authors' calculations.

The pass-through results are the more novel contribution. Before the pandemic, the response of core inflation to oil prices was very small and statistically insignificant: the point estimates in Table 2, which are for one-unit changes, imply that a 10-percentage point increase in relative oil-price inflation raised core inflation by only about a statistically insignificant 0.01 percentage point. After the pandemic, the pass-through rises to a statistically significant 0.14 percentage point. Although still modest in absolute terms, this estimate represents a meaningful qualitative change: a channel that was effectively switched off before the pandemic became active afterwards. Together, the two panels of Table 2 imply that oil-price shocks now transmit to core inflation both more strongly and more durably than before.6

Figure 2 translates these coefficient changes into a cumulative response, tracing the effect on the core price level of a permanent 10 percent increase in oil prices under three specifications. Under pre-pandemic dynamics (the black line), the effect is very small, raising the core price level by only 0.05 percent after three years. Allowing persistence to rise (the red line) lifts the cumulative effect to just under 0.1 percent, because the same modest pass-through now decays more slowly. Allowing pass-through to strengthen as well (the blue line) raises the response much further, to 0.35 percent after 12 quarters, with most of the effect building over roughly eight quarters.7 The bulk of this increase comes from the stronger pass-through, amplified by higher intrinsic persistence of the core inflation process.

What does the model imply for the 2026 oil price shock?

We next use the estimated model to gauge how the recent rise in oil prices might affect core inflation over the next few years. We feed into the model three key inputs: core inflation data through 2026:Q2; the Brent futures curve (as of early August 2026) for the path of oil prices from 2026:Q3 onward; and the reading of the GSCPI through June 2026, at 1.63 standard deviations above its historical mean and projected forward at an estimated quarterly decay rate of 0.8.8 Figure 3 reports the resulting projections under our three specifications.

The three specifications imply materially different paths for core inflation. With no post-pandemic dynamics (the black line), core inflation edges up only modestly, peaking near 2.2 percent. Allowing for stronger persistence (the red line) raises the path to a plateau of about 2.5 percent. Adding stronger pass-through from oil prices as well (the blue line) produces the highest and most front-loaded response: core inflation climbs to nearly 2.8 percent around mid-2027, roughly 0.6 percentage point above the path implied by pre-pandemic dynamics, before receding toward about 2.2 percent by the end of 2028. The difference between the red and black lines—about 0.3 percentage point at the peak—reflects the stronger persistence; the additional difference between the blue and red lines—a further 0.3 to 0.35 percentage point—reflects the higher pass-through. The timing differs across channels: the higher pass-through dominates the near-term peak but fades as the oil impulse in the futures curve normalizes, so the blue path eventually falls back below the red one; the stronger persistence, by contrast, produces a more durable elevation that keeps core inflation above the no-persistence path well into 2028. The post-pandemic inflation process can thus turn even a temporary relative-price shock into a more lasting increase in underlying inflation.

These forecasts assume that the higher persistence of core inflation and stronger oil-price pass-through estimated for the post-pandemic period remain in place at least through 2028. We therefore view these paths for core inflation as an upper bound on the likely impact rather than a central prediction. There are several reasons why persistence and pass-through may prove lower than assumed. The post-pandemic surge in inflation was extraordinary and the result of a confluence of factors, including reopening demand, fiscal stimulus, labor-market tightness, supply-chain disruptions, and high energy prices (as argued, for instance, by Bernanke and Blanchard, 2024). As such, the post-pandemic interaction terms likely capture more than a pure structural oil-to-core channel and, in line with Kilian (2009), may partly reflect the co-movement of oil prices with broader global demand pressures. In a more normal environment, both persistence and pass-through could therefore be weaker than our post-pandemic estimates imply. To that extent, the estimated amplification may overstate the effect of an isolated oil shock and the extent of lasting structural change in the inflation process remains uncertain.

Consistent with this interpretation, second-round effects have not yet materialized; however, risks are mounting. Indeed, some advanced-economy central banks, including the European Central Bank and the Bank of England, have noted that measures of underlying inflation remain broadly stable but also highlighted that inflation risks have increased. The risk highlighted by our estimates is conditional: if oil prices remain elevated for longer than implied by futures pricing, stronger post-pandemic transmission could reassert itself and push core inflation closer to the path implied by our model.

Macroeconomic and monetary policy implications through a structural model

We conclude by assessing the macroeconomic and monetary policy implications of the estimated changes in the inflation process that we have documented above. We examine this issue using the GEMUS model, a structural open-economy macroeconomic framework developed by Federal Reserve Board staff. The model comprises the U.S. and a foreign block, which includes a small region of oil-producing countries. We generate an increase in oil prices through a reduction in that region's supply of oil to global markets. The model also incorporates trade in final consumption goods, intermediate inputs, and capital goods, as well as a broad set of features found to be important in accounting for the data, including nominal and real rigidities and financial frictions.9

The black lines in Figure 4 depict the effects on advanced economies of a 10 percent increase in oil prices under a model calibration corresponding to the pre-pandemic inflation process—with low intrinsic inflation persistence and low pass-through from oil prices to core inflation. Under these conditions, the effects of the oil price shock are manageable: activity declines a little over 0.1 percent, headline inflation rises noticeably but briefly, and core inflation increases only modestly. Monetary policy responds modestly, tightening slightly in response to higher core inflation.10 The red lines show the effects of the same shock under the higher inflation persistence indicated by our post-pandemic estimates. With greater persistence, the rise in core inflation is larger and longer lasting, and the contraction in activity is somewhat more severe, although the differences are not large. Monetary policy continues to respond only modestly.

The picture differs, however, when higher persistence is accompanied by stronger pass-through, shown by the blue lines. Here we assume that firms' marginal costs respond more strongly to an increase in oil prices—engineered through a higher cost share of oil in production—and calibrate this increase to match the rise in pass-through estimated in our regressions.11 The increases in both headline and core inflation are now much larger and more persistent, and the GDP hit is about twice as large. Monetary policy is put in the uncomfortable position of having to tighten materially to contain inflation even as the economy weakens. The decline in activity reflects both supply-side forces, as firms face higher costs, and demand-side forces, as higher oil prices and tighter monetary policy restrain spending. The model exercises illustrate the structural mechanisms that might be embedded in the empirical estimates.

Conclusion

Our analysis suggests that oil-price shocks may now pose greater challenges to monetary policy. Oil-price shocks were relatively manageable before the pandemic, when intrinsic inflation persistence was lower and there was more limited oil prices-to-core prices pass-through. However, the post-pandemic data point to a shift in the inflation process that, if it continues, would amplify the response of core inflation to oil-price shocks: the stronger pass-through raises the initial core-inflation response, while higher persistence makes that response more durable. The evidence should be interpreted cautiously, because the post-pandemic interactions are estimated from a short and unusual period and so it is uncertain how long they may continue. Nevertheless, they illustrate an upside risk to inflation should oil prices remain elevated or supply-chain disruptions persist for an extended period.

References

Alp, Harun, Matthew Klepacz, and Akhil Saxena (2023). "Second-Round Effects of Oil Prices on Inflation in the Advanced Foreign Economies," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, December 15, 2023.

Benigno, Pierpaolo, and Gauti B. Eggertsson (2024). "Revisiting the Phillips and Beveridge Curves: Insights from the 2020s Inflation Surge," NBER Working Paper 33095.

Bernanke, Ben, and Olivier J. Blanchard (2024). "An Analysis of Pandemic-Era Inflation in 11 Economies," PIIE Working Paper 24-11. Washington: Peterson Institute for International Economics, May.

Cuba-Borda, Pablo, Albert Queralto, Ricardo Reyes-Heroles, and Mikaël Scaramucci (2025). "Trade Costs and Inflation Dynamics," International Finance Discussion Papers 1411. Washington: Board of Governors of the Federal Reserve System.

De Michelis, Andrea, Grace Lofstrom, Mike McHenry, Musa Orak, Albert Queralto, and Mikaël Scaramucci (2024). "Has the Inflation Process Become More Persistent? Evidence from the Major Advanced Economies," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, July 19, 2024.

Erceg, Christopher J., Marcin Kolasa, Jesper Lindé, and Andrea Pescatori (2024). "Can Energy Subsidies Help Slay Inflation?" IMF Working Paper No. 2024/081.

Federal Reserve Bank of New York. "Global Supply Chain Pressure Index."

Kilian, Lutz (2009). "Not All Oil Price Shocks Are Alike: Disentangling Demand and Supply Shocks in the Crude Oil Market," American Economic Review, 99 (3): 1053–69.

Peneva, Ekaterina, Jeremy Rudd, and Daniel Villar (2025). "Retrospective on the Federal Reserve Board Staff's Inflation Forecast Errors since 2019," Finance and Economics Discussion Series 2025-069. Washington: Board of Governors of the Federal Reserve System.

Taylor, John B. (1999). A historical analysis of monetary policy rules. In J. B. Taylor (Ed.), Monetary Policy Rules (pp. 319–348). University of Chicago Press.

1. The note conveys the authors' own views, not those of anyone else associated with the Federal Reserve Board or the Federal Reserve System. We thank Cian Nicholas for excellent research assistance and Shaghil Ahmed for helpful comments. Return to text

2. This specification uses actual oil-price changes rather than attempting to isolate oil supply shocks using structural decomposition methods (e.g., Kilian, 2009). This choice reflects our focus on the issue that is relevant for policy: conditional on observed oil-price movements and cyclical conditions, how much do core prices respond? While oil prices rise for various reasons, our inclusion of the output gap controls for the state of economic activity, helping to isolate the cost-push component. Return to text

3. We also augmented the specification by controlling for relative core import price inflation but found no evidence of additional explanatory power for core inflation in the panel regression. While Peneva et al. (2025) showed that for the U.S. it is important to control for import prices excluding oil and other energy products, we lack a consistent measure for it across major advanced economies. Relatedly, there are no tariff-inclusive measures of import prices for the foreign advanced economies. We therefore did not include import prices in our baseline specification. Return to text

4. Country-specific OLS estimates deliver largely similar results, with the rise in core inflation persistence post-pandemic broad-based and significant across all five economies. The increase in oil pass-through shows more cross-country variation: it rose significantly in Canada, Japan, and the U.S., but was weaker and not statistically significant in the euro area and the U.K., possibly reflecting Europe's natural gas price shock following Russia's invasion of Ukraine. Return to text

5. Using energy CPI in place of oil prices delivers similar, indeed somewhat larger, estimates of the post-pandemic increase in pass-through. We also test our specifications without contemporaneous oil prices, which delivers largely similar persistence and pass-through results. Return to text

6. As a robustness check, we replace the output gap with the vacancy-to-unemployment ratio (V/U), motivated by Benigno and Eggertsson (2024), who emphasize labor-market tightness and its nonlinear relationship with inflation in explaining the post-pandemic U.S. inflation surge. Canada is excluded from this analysis because its vacancy series only starts in 2015. The estimated post-pandemic increases in persistence and pass-through are 0.292 and 0.010, respectively, close to the baseline estimates of 0.298 and 0.013, and remain statistically significant. Given that results align for both measures, we retained the output gap in the baseline to preserve the five-economy sample. Return to text

7. These estimates can be compared to those of Alp et al. (2023), who found statistically significant but modest second-round effects of oil prices on inflation using data through 2022 from a large panel of advanced economies: a permanent 10 percent oil-price increase raises the core CPI level by about 0.1 percent over roughly eight quarters. That magnitude is close to what we estimate before allowing pass-through to strengthen, but our full specification implies a larger effect because it captures full oil-to-core pass-through and allows both pass-through and persistence to shift after the pandemic, rather than estimating a single average relationship over 2000–2022 with one COVID dummy. A further difference is country coverage: their estimates focus principally on the euro area, the United Kingdom, and Canada, whereas our panel also includes the United States and Japan. Return to text

8. The decay rate is obtained by regressing the GSCPI on one quarterly lag of itself over 2020:Q1–2024:Q4, which yields a slope of approximately 0.8; the GSCPI is measured in standard deviations from its historical mean, so this implies a return toward the mean with a half-life of roughly three quarters. Return to text

9. GEMUS stands for Global Economy Model for U.S. Policy Analysis. The version used here is an extension of the framework developed in Cuba-Borda et al. (2025). Return to text

10. Monetary policy in the model follows an inertial rule similar to that in Taylor (1999), with the policy rate responding to core inflation and the output gap. Return to text

11. The assumption of a higher cost share of oil in production proxies for the underlying reasons why oil pass-through could be stronger, like larger indirect effects on wages or on the prices of other inputs. For example, Erceg et al. (2024) find empirically that an exogenous oil price shock raises real wages in the euro area but lowers them in the U.S., while also generating a larger increase in core inflation in the euro area than in the U.S. We also experimented with generating higher pass-through by raising the frequency of price adjustment—which produces a larger increase in prices for a given rise in marginal cost—and found similar implications for the near-term effects on activity and the policy rate. Return to text

De Michelis, Andrea, Musa Orak, Avani Pradhan, and Albert Queralto (2026). "The Effect of Oil Prices on Core Inflation: Recent Evidence from the Advanced Economies," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, November 9, 2026, https://doi.org/10.17016/2380-7172.4190.

Disclaimer: FEDS Notes are articles in which Board staff offer their own views and present analysis on a range of topics in economics and finance. These articles are shorter and less technically oriented than FEDS Working Papers and IFDP papers.