Resumen
The goal of this paper is to evaluate Brazilian stock market efficiency using daily data for the Sao Paulo Stock Exchange Index from January 1995 to December 2012. We employ a variance ratio statistic with wild bootstrap, developed to test linear dependency, to test for the Random Walk Hypothesis. We also use the generalized spectral test for the nonlinear case. We employ moving subsamples with fixed size, checking the existence of random walk behavior. We test whether market efficiency depends on market conditions (Adaptative Markets Hypothesis - AMH). We cannot reject both the RWH and AMH.