Abnormal Returns in the Ibovespa Using Models for High-Frequency Data

Abstract

This article aims to identify profitable trading strategies based on the effects of leads and lags between the spot and futures equity markets in Brazil, using high frequency data. To achieve this objective and based on historical data of the Bovespa and the Bovespa Future indexes, four forecasting models have been built: ARIMA, ARFIMA, VAR, and VECM. The trading strategies tested were: net trading strategy, buy and hold strategy, and filter strategy – better than average predicted return. The period of analysis of this paper extends from August 1, 2006 to October 16, 2009. In this work, it was possible to obtain abnormal returns using trading strategies with the VAR model on the effects of leads and lags between the Bovespa index and Bovespa Future index.

Keywords

• trading strategies
• high-frequency data
• IBOVESPA