The recent inflation episode has revived economists’ interest in understanding how retailers adjust their prices in response to large aggregate shocks. In a low-inflation environment, prices are sticky and the frequency of price changes is broadly constant (in line with the predictions of standard time-dependent models of price rigidity, like the Calvo model). In particular, the frequency of price changes does not vary with the level of inflation or aggregate shocks, which implies similar delays in the inflation response to both small and large shocks. To explain the sharp inflation surge in 2021-2022, Cavallo et al. (2024) argue that “large shocks travel fast”: facing a large aggregate shock, retailers adjust their prices more quickly than in normal times, and the frequency of price changes increases with the aggregate shock. This implies that state-dependent pricing becomes more prevalent (Gagliardone et al. 2025, Bunn et al. 2026), inflation is less inertial, the Phillips curve is non-linear (Ascari et al. 2025, Boar et al. 2025) and monetary policy trade-offs differ (Karadi et al., 2025).

These findings call for systematic evidence at scale. In a new paper (Gautier et al. 2026), we document new stylised facts on price rigidity in the euro area during the 2021-2024 inflation episode. We build on around 190 million individual price quotes collected across nine euro area countries (Austria, Estonia, France, Germany, Greece, Italy, Latvia, Lithuania and Spain, together covering 83% of the euro area HICP) to compute price adjustment statistics for a common sample of 166 detailed product categories spanning food, non-energy industrial goods and services (together covering about 60% of the HICP basket of products). The 2021-2024 inflation episode, a large, broad-based shock occurring after a decade of low and stable inflation, provides the before-and-after design that makes these statistics analytically powerful. Three main findings stand out.

Fact #1: The frequency of price changes almost doubled, then normalised unevenly

Before 2020, the monthly frequency of price changes was stable at around 8.2% across the nine countries. It then rose sharply in 2022, averaging about 12% that year and peaking at 15.7% in January 2023, nearly double the pre-pandemic average (Figure 1). This rise was widespread: more than two thirds of individual product categories showed a statistically significant increase in frequency in 2022. However, the magnitude of the increase differed across sectors: food experienced the largest increase (7 percentage points above its 2019 average in 2022) while non-energy industrial goods and services each rose by around 3 to 3.5 percentage points.

Figure 1 Monthly frequency of price changes, increases and decreases in the euro area (excluding sales)

Note: the graph plots the average monthly frequencies of price changes (black lines), increases (red lines) and decreases (blue lines) in the euro area, the solid lines plots the +/- 6-period-window moving averages of the frequencies (source Gautier et al. 2026). Euro area HICP inflation excluding energy is represented by the grey histogram (source Eurostat)

The frequency of price changes subsequently fell, returning close to its pre-pandemic level in 2024 for food and non-energy industrial goods, but it declined more slowly for services, where the frequency remained about 2 percentage points above its pre-pandemic average in 2024.

Repricing shifts are a systematic feature of the inflation episode in all euro area countries. The increase in the frequency of price changes in 2022 was more pronounced in Baltic states which experienced a larger inflation shock – contrasting with the absence of country heterogeneity in price stickiness during the low inflation period.

A similar mechanism is visible in cross-sectional regressions at product level. We show that products with a larger share of imported energy inputs in their cost structure had their prices adjusted significantly more often in 2022 and 2023, but not in 2020 or 2021 (Figure 2). This selective response to cost exposure is the fingerprint of state-dependent pricing: firms respond more when the gap between their actual price and their optimal price widens more.

Figure 2 Cross-sectoral frequency of price changes in 2020-2024 and the share of imported energy and raw material input 

Note: The figure shows the estimates of an OLS regression relating year fixed effects estimated at the product level (capturing how the frequency of price changes (or increases or decreases) differs in year t from that in year 2019) to the share of energy inputs in the product-level cost structure constructed from EA input-output matrices.
Source: Gautier et al. (2026)

The difference between how prices of goods and services respond to the inflation episode is itself informative: the slower normalisation in services reflects both a lower structural degree of state dependence – with more stable probabilities of price changes (flatter hazard rates) and more pronounced seasonal January peaks – and wage pressures that continued into 2023-2024. Empirical estimates of monetary policy transmission show that transmission lags to services inflation are substantially longer than for headline inflation, with the peak impact taking more than two years to materialise (Zlobins 2025). Our sectoral results provide some evidence on micro-level mechanisms behind this heterogeneity.

Fact #2: The distribution of price changes shifted in composition, not in scale

During the surge, the average size of (non-zero) price changes rose from 1.5% before the pandemic to 5.5% in 2022 (Figure 3). This number, taken at face value, suggests firms made larger individual adjustments. We show that this interpretation is wrong. The rise was driven almost entirely by a compositional shift: the surge in price adjustment reflects a marked increase in the frequency of price rises and a slight decrease in the frequency of price reductions (Figure 1). Before the pandemic, around two-thirds of all price changes were increases, by 2022 that share had risen to 82%.

Figure 3 Monthly average of (non-zero) price changes, increases and decreases in the euro area (excluding sales)

Note: The graph plots the average sizes of (non-zero) price changes (black lines), increases (red lines) and decreases (blue lines) in the euro area, the solid lines plots the +/- 6-period-window moving averages of the sizes (source Gautier et al. 2026). Euro area HICP inflation excluding energy is represented by the grey histogram (source Eurostat)

Both the average size of price increases and average size of price decreases, taken separately, barely moved. The distribution of price changes shifted toward increases and away from decreases, while the shapes of the two sub-distributions remained essentially stable. The higher moments of the price change distribution, including kurtosis, were barely affected during the surge.

This has a precise theoretical reading. In a state-dependent model, a large positive aggregate cost shock shifts the distribution of price gaps leftward: most firms find their prices too low relative to the optimum. Firms accumulating large positive gaps respond by increasing prices. The result is exactly what the data show: more frequent price increases, fewer price decreases, and stable sizes of price increases and decreases taken separately.

Fact #3: State-dependent pricing became more prevalent

The debate among policymakers during the inflationary period centred on whether firms’ pricing behaviour became more responsive to shocks in high-inflation states, changing the speed of transmission and the persistence of inflation.

We find evidence that retailers are indeed more likely to adjust prices when mispricing is larger. To show this, we construct empirical hazard rates defined as the probability of adjustment as a function of the estimated gap between actual and optimal benchmark prices. Figure 4 shows that the probability of a price change increases with the absolute price gap, meaning that when the price gap is larger, retailers are more likely to adjust their prices. We document that hazard rates are V-shaped in all nine countries and across all major sectors. The slope varies across sectors, being steeper for food and non-energy industrial goods and flatter for services, consistent with differences between goods and services observed during the normalisation.

Figure 4 Adjustment hazard rates as a function of the price gap

Note: This graph represents the probability of price changes (black line), the probability of price increases (red line) and the probability of price decreases (blue line) as a function of the price gap defined as the log difference between the observed price (in a country and product category) and a benchmark desired reset price (proxied by the average of updated prices in the product category).
Source: Gautier et al. (2026).

We also decompose product-level inflation into two components: the extensive margin (how often prices change) and the intensive margin (how large those non-zero changes are). In the low-inflation period, the intensive margin dominated: most of the variation in product-level inflation came from variation in the size of price changes rather than their frequency. During the surge, the extensive margin (the frequency of price changes) became significantly more correlated with product-level inflation than in ‘normal’ times.

Local projection regressions linking counterfactual inflation rates to oil supply news shocks, as identified by Känzig (2021), confirm that the extensive margin responded more strongly to those shocks during the high-inflation period than before, while the intensive margin’s contribution remained broadly similar. This shift from intensive to extensive margin dominance under large shocks is a clean prediction of state-dependent models.

A simple macro model counterfactual quantifies the aggregate consequence: had frequency remained at its pre-pandemic level, as a standard Calvo model would imply, peak inflation would have been approximately one percentage point lower.

What these statistics offer to research and policy

Overall, the euro area inflation surge did not simply reflect ‘higher inflation’ but also a different repricing behaviour. The micro evidence shows that state dependence, a key behavioural ingredient generating nonlinear inflation dynamics, is present in euro area consumer price data in economically meaningful magnitude. Price rigidity statistics we provide at product level across nine countries and fourteen years are the natural empirical input to several active research programmes.

Alvarez et al. (2022) establish that kurtosis and frequency together constitute a sufficient statistic for the real effects of monetary shocks across a wide class of models, making these moments key inputs for the empirical assessment of monetary non-neutrality. Karadi et al. (2025) show that the degree of state dependence in repricing shapes the sacrifice ratio following large shocks: when the frequency of price changes is high, prices are more flexible and the output cost of fighting inflation is lower, allowing the central bank to lean more forcefully against inflation. Related work combining state-dependent pricing with production networks shows that supply-side shocks can generate rapid cascading repricing waves that demand-side frameworks miss entirely (Ghassibe and Nakov, 2026). The moments of the price change distribution provided by in our paper, during the recent inflation episode at a granular sectoral level across the euro area, provide the empirical foundation for all these frameworks and for the broader question of how pricing mechanisms shape monetary transmission in a large, heterogeneous currency area.

Authors’ note: The views expressed in this column are those of the authors and do not necessarily reflect those of the National Central Banks nor the Eurosystem.

References

Alvarez F, F Lippi, A Oskolkov (2022), “The Macroeconomics of Sticky Prices with Generalized Hazard Functions”, The Quarterly Journal of Economics 137(2): 989–1038.

Ascari, G, A Carrier, E Gasteiger, A Grimaud and G Vermandel (2025), “Monetary Policy in the Euro Area, when Phillips Curves … are Curves”, CEPR Discussion Paper No. 20489.

Boar, C, A Blanco, C Jones and V Midrigan (2025), “The Inflation Accelerator”, NBER Working Paper No. 32531

Bunn, P, N Bloom, C Menzies, P Mizen, G Thwaites, and I Yotzov (2026), “Why inflation may respond faster to big shocks: The rise of state-dependent pricing”, VoxEU.org, 6 February.

Cavallo, A, F Lippi, and K Miyahara (2024), “Large Shocks Travel Fast.” American Economic Review: Insights 6(4): 558–74.

Gagliardone, L, M Gertler, S Lenzu, and J Tielens (2025), “Micro and macro cost-price dynamics in normal times and during inflation surges”, VoxEU.org, 6 June.

Gautier, E, C Conflitti, D Enderle, L Fadejeva, A Grimaud, E Gutiérrez, V Jouvanceau, J-O Menz, A Paulus, P Petroulas, P Roldan-Blanco and E Wieland (2026), “Consumer Price Stickiness in the Euro Area During an Inflation Surge”, ECB Working Paper 3181 and CEPR Discussion Paper 21149.

Ghassibe, M and A Nakov (2026), “Pricing cascades: inflation in a networked economy”, VoxEU.org, 4 February.

Karadi, P, A Nakov, G Nuño, E Pasten and D Thaler (2025), “Why monetary policy should crack down harder during high inflation”, VoxEU.org, 4 May.

Känzig, D (2021), “The macroeconomic effects of oil supply news: evidence from OPEC announcements”, American Economic Review 111(4): 1092–1125.

Zlobins, A (2025), “Monetary policy transmission in the euro area: Why this time it’s different”, VoxEU.org, 26 February.