This study investigates the dynamic relationship between three macroeconomic variables in Brazil: Interest rate, exchange rate and Ibovespa index. The first variable analyzed was the interest rate, using the prices of futures contracts CDI traded on the BMeF Bovespa maturing in 360 days and raised through the Bloomberg platform. The second variable analyzed was the exchange rate real/dolar real historic provided by the Central Bank of Brazil. The third variable analyzed was the Ibovespa index raised by Economática program. For this study were collected daily data for the three variables between January 4,1999 to September 4, 2015 in order to study the dynamic relationship between the variables and the existence of long-term memory and analysis of shocks volatility by applying econometric models. The VAR model was applied with 13 lags for the analysis of the dynamics related to study the explanatory power between the three variables. The FIPARCH model was applied to test the existence of long memory and analyze the impacts of volatility shocks in the variables. The results were significant and showed a greater explanatory power of the exchange rate on the remaining variables, the existence of long memory and that the conditional volatility is more affected by positive shocks to the Ibovespa index and more affected by negative shocks to the exchange rate and the interest rate. This study is important for professionals in business and government to plan their short and long term actions to control and planning of the economy and to contribute for others studies of this topic.