Determining the drivers of inflation in real time is a central challenge for central banks. This column introduces a new financial markets-based model to identify the types of shocks driving changes in inflation expectations in near real-time. After the initial fall in demand in 2020, supply and policy factors became key drivers of inflation expectations. As the economy reopened in 2021, supply bottlenecks kept inflation elevated, until stronger demand became the main driver from mid-2022 onward. Furthermore, it shows that pass-through to inflation varies by shock type: global value chain disruptions are especially persistent, widespread, and inflationary.
Post-pandemic inflation posed a core challenge for central banks: determining the drivers of inflation in real time to adjust policy accordingly. Early on, the discussion centred on overheating demand and tight labour markets (Summers 2021, Blanchard 2021, Reifschneider and Wilcox 2022). The debate then evolved to incorporate ‘supply chain disruptions’ – sudden disruptions that interrupt the flow of goods and materials in the global supply chain – as a key driver of both the scale and persistence of inflation. Recent studies support this view: Comin et al. (2024) explain theoretically how these disruptions push up prices, while Ascari et al. (2024) and Bai et al. (2024) quantify their impact relative to energy and other supply shocks. Their findings point to strong and lasting price pressures from supply-chain breakdowns. Consistently, using data up to 2024, Bernanke and Blanchard (2025) and Dao et al. (2024) estimate that supply shocks played a major role in the early phase of the inflation surge.
However, the origin of the shocks (and hence a quantification of their persistency) was not available to policy makers in real time, when it was most needed. Initially, several prominent policymakers misdiagnosed the inflation surge as transitory. For instance, it took more than six months — and five Federal Open Market Committee (FOMC) statements — for the term ‘transitory’ to outlast its own definition. Consequently, the Federal Reserve began raising interest rates only when inflation had reached 7.9%, launching one of the steepest and fastest tightening cycles in its history, increasing policy rates by 5.25 percentage points over just 16 months. Flash estimates of key macroeconomic aggregates were not helpful either because, as shown by Giannone and Primiceri (2025), they understated the strength of the post-pandemic recovery, presenting a more optimistic outlook for inflation than subsequent revised data revealed.
But macroeconomic data are not the only clues to inflation. Financial markets also reveal real-time expectations through instruments like inflation-linked swaps (ILS rates) — contracts whose value reflects where investors think inflation is headed. These market-based measures track actual inflation closely (see Figure 1) and offer an early warning signal. The challenge is using this timely information effectively to extract meaningful signals about emerging inflationary pressures and their drivers.
Figure 1 Correlation of inflation expectations with realised inflation
Figure 1 Correlation of inflation expectations with realised inflation
Notes: The figure reports the estimated coefficient and the R-squared obtained regressing realized inflation on three measures of inflation expectation: market-based inflation expectations (1-year 1-year inflation linked swaps, upper row) the one-year ahead Michigan inflation survey (middle row) or the Federal Reserve Bank of Cleveland's expected inflation rate (bottom row). The dependent variable is realized US CPI inflation between periods t-1 and t+k. The regression is estimated at monthly frequency. The vertical lines show 95% and 68% confidence intervals.
In a recent paper (Cassinis et al. 2025), we tackle this question by building a mixed-frequency model for the US that combines weekly financial data – specifically, stock prices, ten-year yields, energy prices, and one-year-forward one-year inflation-linked swap rates (1Y1Y ILS) – with monthly macro indicators (industrial production and the supply chain index by Benigno et al. 2022). This setup helps identify the types of shocks driving changes in inflation expectations in near real time. We distinguish five main sources of price movements: demand shocks, monetary policy shocks, energy shocks, domestic supply shocks, and global value-chain (GVC) shocks, which are linked to supply disruptions (see Figure 2). Importantly, we use the low-frequency variables to inform the identification of higher frequency shocks by rejecting draws of the model that are inconsistent with standard theory – like an expansionary demand shock that lowers industrial production. These cross-frequency restrictions help to filter noise from financial data and to better link them to macro aggregates.
Drivers of post-Covid inflation
The model helps explain the post-Covid-19 inflation (see Figure 2). In early 2020, inflation-linked swap rates plunged as demand collapsed amid lockdowns and uncertainty. This drop quickly reversed as supply and policy factors took over. Global value-chain (GVC) disruptions (red area) – from lockdowns and factory shutdowns to shortages of key inputs like semiconductors – pushed inflation expectations higher even as output remained weak. At the same time, the Federal Reserve launched a rapid tightening cycle, contributing as much to rising inflation expectations as GVC shocks between 2021 and 2022. Energy markets added to the pressure: OPEC+ production cuts in spring 2020 drove a strong energy-related rebound in expected inflation. Finally, as the global economy reopened in 2021, supply bottlenecks (from port congestion and soaring shipping costs) kept inflation elevated, until stronger demand became the main driver from mid-2022 onward.
Figure 2 Drivers of market-based inflation expectations during the post-pandemic period
Figure 2 Drivers of market-based inflation expectations during the post-pandemic period
Notes: The figure reports the median historical decomposition for the period between January 2020 and March 2024. The black line reports the cumulated percentage changes of each variable, standardised to zero at the first observation. The model is a Bayesian mixed-frequency vector autoregression (VAR) (Eraker et al. 2014, Ghysels 2016, and Schorfheide and Song 2015) designed to identify the structural shocks driving US inflation expectations at a weekly frequency. Weekly financial variables (yields, stock and commodity prices, ILS rates) are combined with lower-frequency macroeconomic indicators within a unified framework. Specifically, two monthly variables are used: US industrial production and the New York Fed’s Global Supply Chain Pressure Index (Benigno et al. 2022). The model is identified combining sign and magnitude restrictions, leveraging on the Supply Chain Index’s exogeneity from demand and energy factors.
Consistent with findings from Giannone and Primiceri (2024) and Bernanke and Blanchard (2025), our results show that US inflation has shifted from being driven mainly by supply shocks to being increasingly shaped by demand forces. Importantly, this insight can be obtained in real time, highlighting the model’s ability to track the drivers of inflation expectations as they evolve. Since it relies on financial market data, the analysis can be updated continuously, is not affected by data revisions and could guide policy decisions in (quasi) real time.
Link to inflation
But for financial market signals to guide policy, they must contain real information about price developments. We test this by running local projections that map the identified weekly shocks onto a range of US price indicators, down to the industry level, to trace how they spread through the economy. The results show that inflation pass-through varies greatly by shock type. Energy shocks cause an immediate but short-lived spike in headline inflation, which fades as weaker demand brings energy prices back down. Domestic supply shocks, by contrast, have a slower but stronger effect on core inflation, offset partly by lower output and energy prices. The most persistent and widespread inflation pressures stem from global value-chain (GVC) shocks: by disrupting access to key inputs and capital goods, they push up production costs across sectors, sustaining high producer and consumer prices. And because these disruptions originate abroad, their inflationary impact is not offset by weaker domestic demand or lower energy prices. This evidence is in line with existing research, like Ascari et al. (2024) and Bai et al. (2024), but results from a financial market-based model.
Figure 3 Reaction of prices to high-frequency supply shocks
Figure 3 Reaction of prices to high-frequency supply shocks
Notes: The figure reports the impulse responses to energy supply, global value chain (GVC), and domestic supply shocks. The shaded areas denote the 68% and 95% confidence intervals. Errors are bootstrapped. Impulse responses are computed by local projections, controlling for industrial production, energy prices and the US dollar exchange rate.
Analysing industry-level data confirms that energy shocks primarily affect energy-related and energy-intensive sectors, while global value chain (GVC) shocks reverberate across nearly all manufacturing and service industries.
Conclusion
This column shows how to leverage insights from financial markets to inform policy decisions in quasi-real time. The results show that not all supply shocks are equal; global value chain disruptions are especially persistent, widespread, and inflationary. Monitoring these shocks in near-real time is essential for timely and well-calibrated policy actions. Crucially, not all supply shocks can be ‘looked through’ – some, such as GVC and domestic shocks, may call for a more proactive policy response.
References
Ascari, G, D Bonam and A Smadu (2024), "Global supply chain pressures, inflation, and implications for monetary policy", Journal of International Money and Finance 142: 103029.
Benigno, G, J Di Giovanni, J J Groen and A I Noble (2022), "The GSCPI: A new barometer of global supply chain pressures", FRB of New York Staff Report 1017.
Bernanke, B and O Blanchard (2025), "What caused the US pandemic-era inflation?", American Economic Journal: Macroeconomics 17(3): 1–35.
Blanchard, O (2021), "In Defense of Concerns Over the 1ドル.9 Trillion Relief Plan", Peterson Institute for International Economics, 18 February.
Cassinis, M G, M Ferrari Minesso and I Van Robays (2025), "Supply Shocks and Inflation: Timely Insights from Financial Markets", ECB Working Paper 3096.
Comin, D A, R C Johnson and C J Jones (2024), "Supply Chain Constraints and Inflation", NBER Working Paper 31179.
Dao, M C, P-O Gourinchas, D Leigh and P Mishra (2024), "Understanding the international rise and fall of inflation since 2020", Journal of Monetary Economics 148: 103658.
Eraker, B, C W Chiu, A T Foerster, T B Kim and H D Seoane (2014), "Bayesian mixed frequency vars", Journal of Financial Econometrics 13(3): 698–721.
Giannone, D and G Primiceri (2024), "The drivers of post-pandemic inflation", NBER Working Paper 32859.
Ghysels, E (2016), "Macroeconomics and the reality of mixed frequency data", Journal of Econometrics 193(2): 294–314.
Reifschneider, D and D Wilcox (2022), "The case for a cautiously optimistic outlook for US inflation", Peterson Institute for International Economics Policy Brief PB22-3, March.
Schorfheide, F and D Song (2015), "Real-time forecasting with a mixed-frequency var", Journal of Business & Economic Statistics 33(3): 366–380.
Summers, L (2021), "The Biden Stimulus Is Admirably Ambitious. But It Brings Some Big Risks, Too", Washington Post, 4 February.
