What core inflation indicators measure?

Whether excluding food and energy components from overall price indices produces a useful indicator for monetary policy purposes is widely debated. The proposals of model based measures of underlying inflation are scarce and the evidence on their performance is limited. In the paper the multidimensional performance of exclusion and model based core inflation indicators is compared in the period of persistently low inflation and interest rates. Providing new measures of underlying inflation we look for specific features of such indices as: tracking trend, appropriate smoothing, unbiasedness with respect to the cost-of-living index, good approximation of the demand pressure, and good short- to medium-term forecasting abilities. To this end, we extract permanent and transitory components of headline HICP and core inflation in the sample of 26 European Union countries for the period 2002-2016 using bivariate unobserved correlated components model and maximum likelihood estimator. We construct an aggregate performance measure, named Core Inflation Score, to capture different dimensions of underlying inflation indicators which could be of interest in monetary policy analysis.