Non-Linear Dynamics in Futures' Returns

This research work studies non-linear dynamics and deterministic behavior of Russian stock index futures prices, considers the behavior of stock index futures basis and also finds forecasting opportunities of index futures returns. The analysis of presence of non-linear dynamics is conducted through the study of correlation dimensions of futures price series and the set of BDS independence test for testing non-linearity. The presence of deterministic relationship and chaos in price series is checked with Lyapunov exponents test. The behavior of basis is analyzed using VAR model. Opportunities for returns forecasting is checked with ARMA, AR-GARCH-GED and neural network models. Besides, the possible reasons for index futures non-linear dynamics are discussed. As a result, index futures prices are subject to non-linear dynamics, and prices are not identically and independently distributed. Neural network model shows high forecasting performance which makes evidence of index futures market inefficiency. RSTI futures has volatility clustering effect, but MICEX futures are subject to deterministic chaos under certain conditions. The results of this work will make a contribution to efficiency of algorithmic trading.