UC Riverside

We develop sentiment indexes to study the relationship between sentiment and stock returns. For the first index, we use only market return data; for the second, we use traditional indicators including the closed-end fund discount, NYSE share turnover, and the equity-share of new issues. In sample we find that rising sentiment leads to rising returns and lower risk. We also show that sentiment cycles lead bear markets. In a forecasting exercise we develop two-stage model averaging (2SMA). 2SMA is a flexible framework that allows researchers to incorporate prior economic information into their forecasts. Through 2SMA we develop two-stage Bayesian model averaging (2SBMA) and two-stage equal-weighted averaging (2SEWA). With these techniques we forecast stock returns using investor sentiment. We find that the 2SBMA and 2SEWA forecasts beat their traditional model-averaging counterparts and beat the benchmark random-walk plus drift in a statistically significant manner.