Inefficiency of the Brazilian Stock Market: the IBOVESPA Future Contracts

TL;DR

This paper investigates the inefficiency of the Brazilian stock market by analyzing cross-correlations with foreign markets, return distributions, and neural network forecasts, revealing significant deviations from market efficiency.

Contribution

It combines correlation analysis, return distribution study, and neural network forecasting to demonstrate market inefficiency in Brazil, which is a novel multi-method approach.

Findings

Strong dependence on foreign indices like S&P 500 and CAC 40

Discretely fat-tailed distribution of daily log-returns

Neural network forecasts achieve profitable accuracy

Abstract

We present some indications of inefficiency of the Brazilian stock market based on the existence of strong long-time cross-correlations with foreign markets and indices. Our results show a strong dependence on foreign markets indices as the S\&P 500 and CAC 40, but not to the Shanghai SSE 180, indicating an intricate interdependence. We also show that the distribution of log-returns of the Brazilian BOVESPA index has a discrete fat tail in the time scale of a day, which is also a deviation of what is expected of an efficient equilibrated market. As a final argument of the inefficiency of the Brazilian stock market, we use a neural network approach to forecast the direction of movement of the value of the IBOVESPA future contracts, with an accuracy allowing financial returns over passive strategies.