Modelos de ciclos reais de negócios aplicados à economia brasileira

Abstract

This article tests if standard business cycle models can reliably reproduce aggregate Brazilian time series. The article first describes some features of the Brazilian data that the models are meant to reproduce. Then it describes a version of the standard business cycle model with indivisible labor, along with the cash-in-advance extension of that model, both of which assume that fluctuations in economic activity are caused only by shocks to technology. Some key parameters are estimated using the Generalized Method of Moments but others, such as taxes, by calibration, The fit of the standard models to the data is disappointing, but adding tax distortions to the cash in advance model is shown to significantly improve its ability to mimic some of the observations. However, a number of key stylized facts, such as the excess volatility of consumption, were not reproduced by any of the models studied.