Modelos econométricos para a previsão da taxa de inflação nos países emergentes

Abstract

Inflation is a process by which there is an increase in prices of goods and services, causing the devaluation. This process is present most strongly in underdeveloped countries, due to the grade of instability of their economies, generating various negative effects, for example, a level below the optimal growth and increasing poverty in a country. For these reasons, inflation is characterized by economic instability be monitored by economic agents and countered by central banks, which require forecasts of this important variable to form expectations about the future behavior of the price level of the economy. Aiming to contribute with empirical evidence on this issue, this study estimated and generated predictions for the inflation rate from ARMA, VAR and Phillips Curve (CP) using samples from emerging countries. One of the main results, it was found that in underdeveloped countries the use of ARMA models predictions generates more efficient than models without Phillips Curve modeling residues. But when it includes the transfer function in the Phillips curve, this generates efficient forecasts only in half of the countries analyzed. The analysis of variance of inflation reveals that a possible explanation for this difference in performance prediction of transfer functions in development can be attributed to the higher variance of this variable