Time Series Properties of four Latin American Equity Markets: Argentina, Brazil, Chile and Mexico

Abstract

Variance ratio tests indicate that the equity markets of Argentina, Brazil and Mexico follow random walks, but not those of Chile. The low correlations among the four markets suggest that investments in these countries can contribute to reduce portfolio risk. The research on the random walk hypothesis has been heavily concentrated on the large equity markets of the United States, Canada, Japan and Europe (summers 1986; Fama and French 1986a, 1986b; Lo and MacKinlay 1988, and Poterba and Summers 1988). Even though some studies have been conducted for stock markets of developing countries (Errunza 1983 and Errunza and Losq 1985), little research has been done in Latin American capital markets (Errunza and Losq 1987). This paper employs the variance-ratio test to investigate the random walk hypothesis for the following four Latin American equity markets: Argentina, Brazil, Chile and Mexico. Two versions of the variance-ratio tests are implemented : first, the variance-ratio under the maintained hypothesis of homocedasticity and, second, the heteroscedasticity-robust variance-ratio. The empirical results reported in the paper indicate that the random walk hypothesis is rejected for Chile but it is generally confirmed for Argentina, Brazil and Mexico. Therefore, American investors might not be able to develop investment strategies that can be generate abnormal returns in these three countries. However, the low correlation among these markets suggests that American investors can reduce the risk of their portfolios by diversifying in international stocks of these countries.