The analysis of market efficiency is presented in three versions: the weak form, the half-fort and the fort. If a market be efficient doesn't have information available for the agents, that assist them to forecast behaviors future and thus the market follows a random walk. If a market doesn't be efficient, it is possible to assimilate information, through past returns, that can allow that an agent has a better performance of what another one. The objective of this work is to empirically test the Hypothesis of Market Efficiency, in its weak form, to verify if historical series of index IBOVESPA and rate of exchange Dollar/Real have some power to forecast returns future. Two methodologies for analysis of variable return will be applied. The first consists of collecting data with a daily frequency and transform the positive or negative variations into a binary sequence, to analyze the vectors divided in four different sizes and to conclude, through conditional probability, if have or not available information in the market that assist the agents to get some competitive advantage. The second will deal with the autocorrelation of the absolute values of returns by the DFA (Detrended Fluctuation Analysis).
News published in Agência FAPESP Newsletter about the scholarship: