Non-linear dynamics in financial asset returns: the predictive power of the CBOE volatility index

In this paper we attempt to predict the direction of change of the S&P500 index over the period 8 April 1998 to 5 February 2002 by means of a recurrent neural network (RNN). We demonstrate that the incorporation in the trading rule of the Chicago Board Options Exchange (CBOE) volatility index changes strongly enhances its profitability during 'bear' market periods. This improvement is measured in comparison with a RNN including changes of estimated conditional volatility measures, a linear autoregressive model as well as to a buy-and-hold strategy. We suggest a number of theories that are consistent with our findings.