Curva de Philips para a previsão da taxa de inflação nos países do G7

Abstract

This paper compares the forecasts of quarterly inflation rate of countries that compose the G7, which is composed by the seven largest economies in the world, made from time series models such as ARMA and VAR models and the Phillips curve. Another approach is the use of the Phillips curve models with ARMA modeling of waste. There is no consensus about the superiority of forecasts generated by models of the Phillips curve on the time series models. Thus, based on the predictions generated, it is possible to evaluate the efficiency of predictive Phillips curve in comparison to the different models under study and thereby classifying the model predictive results in more efficient. The results showed that the ARMA model is more efficient compared to other models, with the exception of Germany. However, with the insertion of the ARMA process in modeling the Phillips curve, the situation was reversed in favor of these models that generate predictions became more efficient in almost all developed countries, France being the only exception. VAR models have not obtained good results for any of the countries analyzed