Testing for Lead-Lag Effects Between the American and the Brazilian Stock Markets

Abstract

This article examines the existence of lead-lag effects between the U.S. stock market, represented by NYSE and the Brazilian stock market, represented by Bovespa, i.e., whether upward and downward price movements in the NYSE are followed, on average, by similar movements in Bovespa, which would enable predicting stock prices in the Brazilian market, thus providing arbitrage opportunities. The existence of this effect would indicate a relative segmentation between these two markets, which would violate the efficient market hypothesis, whereby stock prices are unpredictable. Cointegration between the two markets was identified as well as the existence of bi-directional causality (Granger test). The results obtained from VECM, TSLS and GARCH regressions showed that the two markets are segmented and that returns of the Bovespa Index (Ibovespa) are to a large extent explained by the stock price movements in the Dow Jones Index some minutes beforehand. However, the results also show that the practice of arbitrage based on the lead-lag effects is not economically feasible due to transaction costs.