Core Inflation: Robust Common Trend Model Forecasting

Abstract

Monetary authorities need a measure of the future inflation trend to keep inflation on target. Many alternative core inflation measures appear in the recent literature intending to avoid the deficiencies of the usual headline inflation index as a predictor. This price index is defined as some weighted average of the individual price change for a list of goods and services. Its use as a future inflation indicator is criticized in the literature because the products are heterogeneous in respect to the variability and some of the prices involved have relevant seasonal movements. A multivariate model simultaneously including the seasonal effects of each component of the price index and a common trend - core inflation - will be developed in this paper. The model is phrased as a dynamic model and a robust sequential filter is introduced. The posterior and predictive distributions of the quantities of interest are evaluated via stochastic simulation techniques, MCMC - Markov chain Monte Carlo. Different models are compared using the minimum posterior predictive loss approach and many graphical illustrations are presented.