Abstract:

This paper explores the predictive power of the S&P 500 Implied Correlation Indexfor the S&P 500 Index returns. We propose three models in which future S&P 500Index multi-period returns are regressed on the current information set of the implied correlation index changes. The result shows that changes in this implied correlation index, i.e. current weekly change and changes in the past, are strongly linked to the S&P 500 Index future multi-period returns and our models consistently outperform the random walk model using the Superior Predictive Ability test.