Measuring core inflation

In this paper we argue that measured (RPI) inflation is conceptually mismatched with core inflation: the difference is more than just "measurement error". We propose a technique for measuring core inflation, based on an explicit long-run economic hypothesis. Core inflation is defined as that component of measured inflation that has no (medium-to) long-run impact on real output - a notion that is consistent with the vertical long-run Phillips curve interpretation of the co-movement in inflation and output. We construct a measure of core inflation by placing dynamic restrictions on a vector autoregression (VAR) system.