erc/metu
INTERNATIONAL CONFERENCE IN ECONOMICS  IV
September 13-16, 2000, Ankara

 

Testing the Random Walk Hypothesis for Brazilian Stock Market

Benjamin M. Tabak (Central Bank of Brazil, Unýversity of Brazil and Catholic University of Brasilia)

Abstract

In this paper the random walk hypothesis is tested for a set of daily/weekly Brazilian stock data given by the Săo Paulo Stock Exchange Index (IBOVESPA) in the period of 1989-1998. Both the homoscedastic and heteroscedastic versions of Lo-ackinlay's variance ratio test were used. Our results suggest that prices do not follow a random walk for daily/weekly data for short horizon periods, which would suggest that the Brazilian stock market contains some predictable components. Therefore significant returns to active management could be observed. To test robustness of this result a rolling variance ratio test for different investment horizons was conducted, and it is concluded that prior to 1994 the random walk hypothesis is rejected but after that it cannot be rejected. Institutionally maturing markets, increasing liquidity and the openness of Brazilian markets for international capital can explain the increase of efficiency of the Brazilian stock market.

Economic Research Center
Middle East Technical University
06531 Ankara Turkey
Phone: + 90 312 210 3044, 210 2003
Fax: +90 312 210 1244
e-mail: metuerc@metu.edu.tr