Anger in predicting the index futures returns

This paper aims to investigate how different emotions affect the subsequent index futures returns. We test the forecasting regressions which predict the S&P 500 index futures returns with lagged text-based emotion (anger, joy, fear, optimism, and gloom) indices and find asymmetric forecasting power exists between pessimism and optimism emotion indices. We show that only the text-based anger index could reliably perform at predicting index futures return in-sample and outperform the prevailing unconditional mean out-of-sample. Notably, the predictive power of the text-based anger index persists after controlling for other emotion indices, investor sentiment indices, and fundamental variables known to predict the futures market. And the asset allocation conditioning on text-based anger index can generate substantial economic benefits. Furthermore, the anger index influences the index futures return through both the discount rate and cash flow channels.