Impactos dos choques na política monetária e no câmbio no Brasil: um modelo de autorregressão vetorial estrutural aumentada por fatores dinâmicos

Abstract

This dissertation aims to estimate the impact of exogenous shocks in monetary policy and exchange rates on economic variables in Brazil, using the information contained in a large number of these variables. To obtain a parsimonious model estimated a model FAVAR (Factor-Augmented Vector Autoregression) where the information contained in a broad subset of selected variables is condensed into a small number of dynamic factors. These factors were included as endogenous variables, variables with which the information was condensed into factors in a SVAR model (Structural Vector Autoregression). All coefficients and unobservable components of FAVAR were estimated jointly employing the Gibbs sampler. The identification of exogenous shocks was achieved through restrictions on signs, the impulse-response functions of variables which the information was condensed using a less dynamic version of the Mundell - Fleming stochastic. Besides having been obtained the impacts of the shocks in monetary policy and exchange rates on a large number of economic variables, we obtained the following additional results: shocks in monetary policy, considering the model FAVAR, have a much smaller effect on the overall level price and level of production than in a VAR model, and the exchange rate shocks do not change significantly impacts when comparing the results of FAVAR and VAR.