Abstract
This study evaluates the inflation forecasts produced by Phillips curve models with and without ARMA modeling of their errors, considering a sample that contains developed and developing countries. The aim of this study is to provide empirical evidence that this simple reformulation of the Phillips curve can serve as a benchmark for studies that propose econometric or time series models more elaborated to predict the rate of inflation. The results show that the use of ARMA components in the Phillips curve decrease considerably its mean square error of forecast for all countries in the sample.
Key Words:
inflation; forecasting; Phillips Curve.