Monetary Policy and Exchange Rate: Effects on Disaggregated Prices in a FAVAR Model for Brazil

Abstract

This article investigates what type of nominal price rigidity there is in Brazil by testing if there are different price responses to sector-specific, monetary policy, and exchange rate shocks. We estimate the effects of these shocks on disaggregated prices of the Brazilian consumer price index (IPCA), from 1999 to 2011, using a factor-augmented vector autoregressive model (FAVAR). There is innovation in the Bayesian technique used to estimate the model, which allows for the consistent imposition of priors directly on the parameters of the structural form of the model and the use of sign restrictions in the impulse responses to identify the model. We combine the Gibbs sampling procedure developed by Bernanke, Boivin, and Eliasz (2004) to estimate a FAVAR with the Gibbs sampling procedure developed by Waggoner and Zha (2002 and 2007) to estimate a Bayesian structural vector autoregressive model (BSVAR). The main results are: (i) series-specific shocks are the main determinants of variances in highly disaggregated price series, but macroeconomic shocks have more influence over aggregated series; (ii) macroeconomic shocks are more persistent than series-specific shocks; (iii) almost all the signs of the price responses to monetary and exchange rate shocks are the ones predicted by the theory; and (iv) the magnitude of the responses to monetary and exchange rate shocks are heterogeneous across sectors.