This paper aims at verifying the nature of the relationship between financial development and economic growth in Brazil. From the theoretical point-of-view, although the prevailing view through which financial development generate direct impacts on economic growth, there are distinct positions. Some authors argue that, in fact, the two variables are jointly determined, while others simply insert this relation in the theoretical framework of demand for financial assets. In face of this, the causality test proposed by Granger was adopted for the empirical verification. Thus, using the data from 1947-2000, 1963-2000 and 1970-2000, the obtained results empirically support the existence of direct and unidirectional impacts of financial development on economic growth, when the following ratios are used as financial development indicators: banking credit to private sector/GDP, financial system credits to private sector/GDP and public resources in the financial system/M2.