Rodrigo Gustavo de Souza

This dissertation analyzed theoretically and empirically the relationship between exchange and prices in Brazil between 1999 and 2009. The theoretical session highlights if the channels of transmission of exchange rates and the two approaches to explain the relationship between these two variables, namely the microeconomic and the macroeconomic approaches. As for the microeconomic factors, the pass-through is emphasized as dependent on the following factors: replaceability of domestic and imported goods, the relationship between domestic and foreign firms, market structure, convexity of demand curve, multinational corporations and non-tariff barriers. However, the focus of this work was to treat macroeconomic determinants of pass-through. Based on the literature, the following conclusions can be drawn about the determinants of pass-through under the macroeconomic approach: 1) The pass-through is not exogenously determined, as posed by the traditional literature focused on the microeconomic, also it is dependent on the monetary regime, 2) the variables – inflation rate, degree of openness, real exchange rate and product gap – affect the pass-through and 3) the conduct of monetary and fiscal policies are extremely relevant in determining the pass-through. Subsequently, this study sought to analyze the Brazilian economy through the implementation of the Real Plan, describing the dynamics between prices and exchange rates. In the methodological chapter, this dissertation sought to discuss the main concepts regarding the econometrics of time series. The empirical session estimated exchange rate pass-through coefficient in Brazil, considering the existence of a structural break in the early 2003. The pass-through calculation until 2002 used the methodology of Vector Error Correction (VEC), because the non-stationary series followed a cointegrating relationship. For the second period, the methodologies employed were those of Vector Autoregressive (VAR) and Structural Vector Autoregressive (SVAR), because the variables, although stationary, presented no cointegrating relationships. Results indicated that the coefficient of transfer (pass-through) from 1999 to 2002 is significantly higher than the coefficient of the period 2003 to 2009. Another important result is that when estimating the pass-through for the second period, the results obtained using VAR and SVAR are quite alike, thus conferring strength to the results.