Forecasting Irish Inflation after the crisis: Evaluating Multiple Bayesian Approaches

This paper presents a follow up to the work of Bermingham et al (2012), and to the long literature on forecasting Irish ination. Using a suite of approaches, we nd that Bayesian techniques compare favourably with existing... [ view full abstract ]

This paper presents a follow up to the work of Bermingham et al (2012), and to the long literature on forecasting Irish ination. Using a suite of approaches, we nd that Bayesian techniques compare favourably with existing forecasting methods used by the Central Bank of Ireland.

The contribution of our paper is twofold. First, using Bayesian Model Averaging we conduct an exercise to identify which real activity measures are important when forecasting Irish inflation. We attempt uncover the to uncover the main drivers of inflation and how these have changed post crisis. We then use the results of this "battery" approach to inform our choice of variables in a bayesian VAR framework and contrast its performance with a univariate unobserved components model with stochastic volatility.

Our findings confirm those of Bermingham et al (2012) that there is a role for omestic measures of economic slack, but that it remains imperative to condition on

external variables when forecasting ination. A signicant contribution of this paper

is in pinpointing the relative importance of domestic and external variables over time

while a further contribution is in adding to the suite of models available for forecasting

ination in Ireland.