Forecasting inflation time series using score‐driven dynamic models and combination methods: The case of Brazil

Abstract

AbstractThis paper evaluates the inflation forecasting models of a small open economy. We compare forecasts of Brazilian inflation (consumer price index—IPCA) based on the generalized autoregressive score (GAS) approach with the forecasts reported in the Focus Bulletin (the Brazilian Central Bank's weekly consensus survey report), as well as with other benchmark models. We selected a restricted number of the variables used most in the literature to compose the forecasting models. Furthermore, based on a confidence set, we applied a wide range of combination methods to the forecast components aiming at improving the forecasting power. Point and density forecast tests were performed to verify the performance of the models. The results showed that the GAS models had good predictive performance for the IPCA, and it was possible to improve their accuracy by combining them with other models.