Uncertainty in seasonally adjusted services inflation: the role of Easter and travel

Prepared by Martin Eiglsperger, Mario Porqueddu and Elisabeth Wieland

Published as part of the ECB Economic Bulletin, Issue 5/2025.

When assessing inflation developments, the ECB complements the monitoring of annual inflation rates by analysing measures of “momentum” to gauge shorter-term developments. Momentum is defined as the three-month-on-three-month percentage change in the seasonally and calendar-adjusted price index (annualised for comparability).^[1] Annualised shorter-term rates need to be interpreted with caution, not least on account of exceptional effects that may considerably increase uncertainties pertaining to seasonal and calendar adjustment, as illustrated in this box.

HICP services momentum increased sharply from March to April 2025 and remained elevated in May and June, in particular reflecting changes in the seasonally and calendar-adjusted month-on-month rate of change owing to travel-related items (Chart A, panel a). While part of the increase in the momentum of services inflation in early 2025 reflected some annual repricing effects, the exceptional surge in travel-related services in April points to a stronger than usual impact from this year’s Easter holiday period (Chart A, panel b).^[2] Travel-related services usually associated with holiday effects – including Easter – encompass air transport, package holidays and accommodation services. Since the seasonally and calendar-adjusted rate of change in HICP services items only extracts average Easter effects, exceptional effects from Easter-related travel still remain in the adjusted series.^[3] This explains the sharp swing in the seasonally and calendar-adjusted month-on-month rate of change from 0.3% in March to 0.7% in April, the subsequent drop to ‑0.1% in May and the increase to 0.5% in June.

Chart A

Seasonally and calendar-adjusted services dynamics and travel-related services

a) Services momentum and month-on-month rates of change	b) Price dynamics of travel-related services
(annualised three-month-on-three-month changes and month-on-month percentage changes)	(month-on-month percentage changes)

Seasonal and calendar adjustment uses statistical estimations to remove these effects; however, the trend as well as cyclical and irregular movements remain present after adjustment. Seasonal adjustment is a statistical estimation technique, which extracts patterns from a time series that appear every year in the same month, affect the series in a similar manner and can be expected to reappear in the years to come.^[4] The ECB conducts separate seasonal adjustments of the main components of the euro area HICP. The HICP for services is also calendar adjusted for Easter, which particularly affects prices for travel services. Removing the seasonal and calendar effects, the remaining adjusted series is composed of trend-cyclical and erratic movements, including potential outliers and one-off effects.^[5]

The recent volatility in the seasonally and calendar-adjusted HICP for services reflects a strong impact of Easter in April, which was only partially captured by the ECB’s calendar adjustment. The dominant role played by travel-related services in the uptick in HICP services inflation in April (Chart A, panel b) points to a strong impact of the late timing of Easter this year (April in 2025 vs March in 2024), likely exacerbated by the Catholic and Orthodox Easter celebrations falling on the same date. The ECB’s calendar adjustment removes average Easter effects, thereby increasing the month-on-month rate of change in HICP for services by about 0.1 percentage points in March 2025 and decreasing this rate by 0.1 percentage points in April 2025. Calendar adjustment identifies the average historical impact of moving calendar dates such as Easter. It is more uncertain than estimations of seasonal effects as statistical evidence is more limited and country-specific school holiday schedules differ. As a result, exceptional Easter-related effects that deviate from historical average effects persist even after seasonal and calendar adjustment and consequently pass through to the momentum indicator.^[6]

An analysis of the ECB’s seasonally and calendar-adjusted HICP for services reveals that estimation uncertainty is generally greater during the second quarter of the year. Revisions of seasonally adjusted backdata usually occur whenever seasonal profiles are re-estimated, as observed in the real-time vintages of the series. Chart B, panel a) presents the range of revised values, i.e. the area between the largest and the smallest reported values for a certain reporting month, over the entire period. In some months, replacing the lowest reported value with the highest one would shift the seasonally and calendar-adjusted month-on-month rate of change by up to 0.3 percentage points. Zooming in on calendar months, seasonally and calendar-adjusted services prices tend to show larger ranges of revised values during the second quarter of the year, suggesting that changes in the timing of school holidays across years as well as evolving seasonal patterns contribute to more uncertain estimates (Chart B, panel b).^[7]

Chart B

Historical range of seasonally adjusted HICP for services

a) Historical range of revised values for seasonally adjusted HICP for services	b) Average range of revised values by calendar month
(month-on-month percentage changes)	(month-on-month percentage changes)

Moreover, seasonally and calendar-adjusted results depend on model choices and parameter settings. While Eurostat provides general recommendations on seasonal adjustment, it does not produce seasonally and calendar-adjusted HICP series. As a result, data users may adopt different approaches to seasonal adjustment. Chart C illustrates how different settings – such as those affecting outliers, trend estimation and seasonal patterns – can significantly alter results compared with the ECB’s preferred approach. According to these alternative methods, the range of differences in seasonally and calendar-adjusted rates of change for the HICP for services is, on average, higher than the range of differences arising from revised values. These alternative approaches may have partially removed exceptional travel-related price increases in April 2025 by attributing them to seasonality to some extent, resulting in lower rates of change for that month. By contrast, the ECB’s preferred approach keeps the above-average effect separate, while only average Easter effects are extracted.^[8]

Chart C

Range of seasonally adjusted HICP for services using alternative seasonal adjustment methods

(month-on-month percentage changes)

Overall, the volatility in seasonally and calendar-adjusted services dynamics should be interpreted cautiously. April 2025 was marked by sharp price increases in travel-related services. Calendar adjustment procedures rely on average calendar effects based on current and past data. In the ECB’s seasonal and calendar adjustment, the higher than average part remains present in the adjusted April 2025 value. The effect is treated as an outlier, not as a seasonal movement; it is higher than a normal Easter effect, while the average Easter effect is extracted by the ECB’s calendar adjustment. Part of the increase could be explained by shifts in demand preferences for tourism services or a redirection of tourism demand from the summer to the spring (reflecting climate change), potentially altering seasonal dynamics on a permanent basis.

References

Bodnár, K., Fabbri, A., Rubene, I. and Zekaite, Z. (2025), “Where do we stand with inflationary pressures arising from price resetting?”, Economic Bulletin, Issue 4, ECB.

Eiglsperger, M. (2019), “A new method for the package holiday price index in Germany and its impact on HICP inflation rates”, Economic Bulletin, Issue 2, ECB.

European Central Bank (2016), “Harmonised Index of Consumer Prices – Easter effects and improved seasonal adjustment”, Economic Bulletin, Issue 3.

Eurostat (2024), “ESS Guidelines on seasonal adjustment”, Publications Office of the European Union, Luxembourg.

Lane, P. (2023), “Underlying inflation”, lecture at Trinity College, Dublin, 6 March.

Lis, E. and Porqueddu, M. (2018), “The role of seasonality and outliers in HICP inflation excluding food and energy”, Economic Bulletin, Issue 2, ECB.

Mehrhoff, J., Eiglsperger, M. and Haine, W. (2011), “Seasonal adjustment and reliability of euro area GDP - Increased uncertainty in times of unusual developments?”, Verein für Socialpolitik – annual conference.

Schnorrenberger, R., Schwind, P. and Wieland, E. (2024), “Forecasting HICP package holidays with forward-looking booking data”, Technical Paper, No 04, Deutsche Bundesbank.