Essays on Momentum and Risk

The efficient market hypothesis stipulates that investors are unable to consistently gain risk adjusted returns with the information known to them at the time of the investment. The expected return, conditional on the information set known to investors, is determined from an assumed expected return theory (asset pricing model). However, it has previously been shown that past winners outperform past losers. A trading strategy taking a long position in previous winner stocks and a short in previous loser stocks earn positive statistically and economically significant risk-adjusted returns. These results are confirmed in international markets, but also in different asset classes. A number of alternative asset pricing models explaining momentum returns imply that momentum should be stronger among high uncertainty assets. Many of these alternative asset pricing models build on investor psychology. This premium, with higher momentum returns among high risk stocks has also been empirically documented. This dissertation evaluates some behavioral explanations to momentum returns by their implications. Behavioral explanations often imply that the momentum anomaly is stronger when uncertainty about information is high. Essay one confirms that there is an overreaction to information causing momentum to be high when uncertainty is high. However, when uncertainty is low momentum still exists, now caused by a slow incorporation of new information into asset prices. Contrary to what many behavioral models imply, the results in essay three suggest that uncertainty about information in the portfolio formation period does not cause a stronger momentum anomaly. Stock prices imply a larger under-reaction to positive and relatively reliable information, than to more uncertain information. Based on previous literature, the results in this study suggest that the driver of the return premium in high volatility assets is the general risk level of single stocks, and not uncertainty about portfolio formation period information. Using equity index futures data, essay two links time series momentum profits to volatility states. Time series momentum portfolio returns are driven by assets in a low volatility state. These results support the general finding that momentum is not caused by uncertainty about portfolio formation period information