Time-varying ARFIMA-GARCH model with symmetric thresholds: applications to inflation

This study suggests a structural modification of the basic ARFIMA-GARCH model by allowing for time-varying baseline mean and, especially, symmetric threshold GARCH. By applying it to the inflation of G7 countries, we find that past excessive positive or negative shocks have positive impacts on future volatility and GARCH persistence. Compared with the ARFIMA-GARCH model, the model in this study has superior performances in identifying and characterizing structural changes and excessive shocks.