Os determinantes da inflação brasileira recente: Estimações utilizando redes neurais

Abstract

In this paper we assess the determinants of recent Brazilian inflation, specifically since 1999 when the Inflation Targeting System was implemented. Using a Neural Network model, we evaluate: (1) if the demand pressures, measured by the output gap and unemployment gap, have a clear and systematic impact on inflation, (2) if “imported inflation”, including tradable goods inflation in dollars and the nominal exchange rate changes, exerts a significant influence on inflation, and (3) if the channel of interest rates cost is empirically verified in the Brazilian case. We conclude that recent Brazilian inflation depends mainly on inflation inertia, “imported inflation” and changes in the nominal interest rate.