January 2019
An unobserved components model with stochastic volatility is used to decompose aggregate Euro area HICP inflation into a trend, seasonal and irregular components. Estimates of the components based only on aggregate data are imprecise: the width of 68% error bands for the seasonally adjusted value of aggregate inflation is 1.0 percentage points in the final quarter of the sample. Estimates are more precise using a multivariate model for a 13-sector decomposition of aggregate inflation, which yields a corresponding error band that is roughly 40% narrower. Trend inflation exhibited substantial variability during the 2001-2018 period and this variability closely mirrored variation in real activity