| dc.description.abstract | This article investigates the extent to which various investor sentiment measures explain the returns in U.S. futures markets where there exist no short-sale restrictions. Consistent with the Miller’s (1977) hypothesis, our empirical results show that the aggregate stock market investor sentiment, individual investor sentiment, and consumer confidence indices provide limited predictive ability. Surprisingly, only the Consensus bullish sentiment index (CBSI), an index compiled based on the opinion of institutional investors, provides comprehensive predictive power for all futures contracts. However, the relationship between changes in CBSI and subsequent futures returns is positive, suggesting that the CBSI may not be a sentiment index measuring noise traders’ misreaction. A further analysis based on a GARCH-in-mean model indicates that, among the sample of 32 futures contract, 8 contracts exhibit a positive risk-return relationship, suggesting that risk premium hypothesis that institutional investors are rational investors. Interestingly, about one third of the futures contracts exhibit a negative risk-return relationship, suggesting that a higher return predicted by an increase in CBSI is accompanied with a lower volatility, thus supporting the smart money hypothesis. | en_US |