Published as part of the ECB Economic Bulletin, Issue 8/2024.
Monetary policy affects consumer prices through a number of channels, while its impact, in terms of both speed and magnitude, varies across consumption categories. The post-pandemic inflation surge was a result of an unprecedented combination of shocks, including supply chain disruptions, energy shocks and the pent-up demand from the re-opening of the economy. The ECB responded forcefully by unwinding the accommodative monetary policy stance that had supported the economy through the pandemic and moving it into restrictive territory. The overall disinflation process that followed reflected the fading of supply shocks and the effectiveness of the steep and decisive interest rate hiking policy. At the same time, the disinflation process was accompanied by persistent dynamics in core inflation – defined as the Harmonised Index of Consumer Prices excluding energy and food (HICPX). This box analyses the heterogeneous pass-through of monetary policy shocks to euro area inflation, with a focus on the distinct behaviour of the individual prices of the goods and services included in the HICPX. Focusing on this index provides considerable insight into developments in an inflation component that is considered to typically capture more persistent dynamics.
An assessment of the monetary policy pass-through to disaggregated prices can complement standard analyses of aggregate inflation. This box presents an estimate of the impact of monetary policy shocks on the prices of each of the 72 COICOP-4 items in the HICPX basket.[1] Following the estimation of item-specific Bayesian vector autoregressive models (BVARs),[2] the individual items across goods and services are classified, according to their responsiveness to monetary policy shocks over a three-year horizon, into three categories of sensitivity: (i) highly sensitive, (ii) moderately sensitive, and (iii) non-sensitive.[3],[4] In this way, it is possible to assess which items from the core inflation basket respond strongly to monetary policy shocks, as well as which items respond swiftly or only with long lags. Such information provides valuable insights into the pass-through of monetary policy to aggregate inflation in the euro area.
The non-energy industrial goods (NEIG) category accounts for more items classified as sensitive to monetary policy than the services category. Items either highly or moderately sensitive to monetary policy make up 33% of the euro area HICPX basket and constitute a larger share of NEIG (44%) than of services (26%).[5] In general, the sensitive category (combining highly and moderately sensitive items) consists of a mix of durable, semi-durable and non-durable goods, while sensitive services are primarily related to recreation and transportation. Chart A illustrates the peak impact of monetary policy shocks on a selection of individual items identified as highly sensitive. Overall, there is considerable heterogeneity in the strength of the monetary policy pass-through to individual items within this category. On average, across the highly sensitive items shown in Chart A, the peak impact of monetary policy on prices is somewhat greater for services than for NEIG. Among services items, the peak impact is greatest for “Passenger transport by air”, followed by “Combined passenger transport” and “Package holidays”. Among NEIG items, it is greatest for “Recording media”, followed by “Motor cars” and “Clothing materials”. The stronger impact of monetary policy shocks on some highly sensitive services than on highly sensitive NEIG items might be explained by the discretionary, leisure-related nature of these services.[6]
Chart A
Peak impact of monetary policy on highly sensitive items
(x-axis: maximum cumulative percentage change; bubble size: weight of item in HICPX)
Sources: Eurostat and ECB calculations.
Notes: The bubbles depict the peak impact on the items most responsive to monetary policy shocks over the three-year horizon. The results are based on the median of the posterior distribution of impulse responses normalised to a 25 basis-point increase in the one-year German Bund. The size of the bubbles is related to the weight of a particular item in the HICPX and based on 2024 consumption weights.
Monetary policy has a similar peak impact on goods and services items classified as sensitive. Chart B compares the impulse responses of sensitive and non-sensitive items to a 25-basis point monetary policy shock.[7] Despite the overlap of the credibility bands for both groupings, the impulse responses of sensitive items are more clearly concentrated in negative values and are different from zero based on the 68% credibility bands. After about 20 months, a 25-basis point tightening shock reduces the cumulative price change of sensitive services and sensitive durable goods by around 1.5 percentage points. Moreover, durable goods exhibit a more forceful response relative to semi- and non-durable goods items, in line with prior evidence.
Chart B
Responses of sensitive and non-sensitive NEIG and services aggregates to monetary policy shocks
(x-axis: years; y-axis: cumulative percentage changes)
Sources: Eurostat and ECB calculations.
Notes: The lines show the median posterior distribution of the impulse responses, while the shaded areas denote the 68% credibility bands. The impulse responses are normalised to a 25-basis point increase in the one-year German Bund.
The inflation rates of items classified as sensitive to monetary policy has declined more than the inflation rates of non-sensitive items since the peak of core inflation. HICPX inflation peaked at 5.7% in March 2023, with both sensitive and non-sensitive items contributing significantly to the overall figure (sensitive items accounted for around 2.6 percentage points – Chart C). Since then, the impact of restrictive monetary policy, together with the fading of the extraordinary shocks, has gradually fed through to prices, particularly for sensitive items. Recent data shows a marked decline in the contribution of sensitive items, which accounted for only 0.8 percentage points of the 2.7% HICPX inflation in October 2024. This left non-sensitive items, in particular non-sensitive services such as rent, medical-related services and some insurance items, as the main driver of core inflation.[8] At the peak, non-sensitive services contributed 2.1 percentage points to the 5.7% HICPX inflation, while the latest figures show a contribution of 1.7 percentage points, which accounts for nearly two-thirds of the recent developments in HICPX inflation.
Chart C
HICPX inflation over time – decomposition into items sensitive and not sensitive to monetary policy shocks
(annual percentage changes; percentage point contributions)
Sources: Eurostat and ECB staff calculations.
Note: The latest observations are for the third quarter of 2024.
While the granular analysis confirms the role of sticky services inflation as the main driver of aggregate inflation recently, it also highlights heterogeneity within the services category. Varying sensitivity is documented not only across the two core inflation subcomponents (NEIG and services), but also within each category. The disaggregated analysis also suggests that, despite most services items exhibiting a rather sluggish response to the latest tightening cycle, monetary policy has been successful in dampening price increases across a range of services items primarily related to recreation and transport services. Taken together, this evidence emphasises how identifying items with exceptionally strong responses in a granular way can help to assess the breadth of transmission to aggregate inflation and to monitor it in a timely manner.
The Classification of Individual Consumption by Purpose (COICOP) standardises the consumption basket items across countries. The four-digit classification in the euro area includes 93 categories of prices. For further details see the Eurostat website.
Based on Allayioti, A., Górnicka, L., Holton, S. and Martínez Hernández, C., “Monetary policy pass-through to consumer prices: evidence from granular price data”, Working Paper Series, No 3003, ECB, Frankfurt am Main, 2024. The estimation uses item-specific Bayesian vector autoregressive models (BVARs) with a range of macro-financial controls. The sample varies across items, covering the period between the early 2000s and September 2023. Monetary policy shocks are as in Jarocinski, M. and Karadi, P., “Deconstructing Monetary Policy Surprises – The Role of Information Shocks”, American Economic Journal: Macroeconomics, Vol. 12(2), 2020, pp. 1-43. The shocks were updated using the database of surprises by Altavilla, C., Brugnolini, L., Gürkaynak, R.S., Motto, R. and Ragusa, G., “Measuring euro area monetary policy”, Journal of Monetary Economics, Vol. 108, 2019, pp. 162-179.
Within 36 months of the shock, items with at least three consecutive months of negative and statistically significant price responses are categorised as sensitive to monetary policy shocks. The remaining items are classified as non-sensitive. The sensitive items are further split into “highly” and “moderately” sensitive, depending on whether their maximum negative response is above (moderately sensitive) or below (highly sensitive) the median response across all sensitive items.
A similar classification for consumption, prices and earnings in the United States is conducted by Andreolli, M., Rickard, N. and Surico, P., “Non-Essential Business-Cycles”, NBER Working Paper, 2024.
Taken together, the items classified as sensitive to monetary policy shocks contribute about one-third of the dynamics of the HICPX and are evenly split between NEIG (50.1%) and services (49.9%).
The literature documents a high sensitivity of consumer energy prices to monetary policy shocks. See, for example, Ampudia, M., Ehrmann, M. and Strasser, G., “The effect of monetary policy on inflation heterogeneity along the income distribution”, BIS Working Paper, No 1124, September 2023.
Goods are further split into durables and others, following several studies that document how spending on durables tends to be more cyclical and more responsive to monetary policy changes than spending on non-durables or services. See for example Dedola, L. and Lippi, F., “The monetary transmission mechanism: Evidence from the industries of five OECD countries”, European Economic Review, Vol. 49(6), 2005, pp.1543-1569.
20 out of 28 items classified as late movers overlap with our classification of items not sensitive to monetary policy. See “The heterogeneous developments of the components of euro area core inflation”, Economic Bulletin, No. 4, Banca d’Italia, October 2023. Examples of such items include rent, medical and dental-related services, and insurance linked to health and transport.