The time-varying evolution of inflation risks

About this event

This paper develops Bayesian inference for time series quantile regressions with time-varying parameters (TVPs). We transform the TVP quantile regression into an equivalent high-dimensional quantile regression model and derive a highly-efficient Gibbs sampler. The new methodology is able to bridge the empirically established benefits of TVP regressions for forecasting inflation, with the flexibility of quantile regression models for modelling the whole distribution of inflation. An application of this methodology points to a very good forecasting performance of quantile regressions with time-varying parameters augmented with specific credit and money-based indicators for the prediction of the conditional distribution of inflation in the Euro-Area, both in the short and the medium run, especially for tail risks.