The Economic Value of Exploiting Time-Varying Return Moments

Abstract

This paper studies the economic value of exploiting time variation in risk premia and in the volatility of stock returns for a real-time investor in a dynamic setting. I find that ignoring time variation in these return moments leads to economically and statistically significant utility costs. Time-varying risk premia play a more important role than time-varying volatility in forming portfolio weights and in improving portfolio performance for the real-time investor who can rebalance portfolios quarterly. In addition, dynamic policies are more susceptible to parameter uncertainty than myopic policies. This effect further reduces the utility costs of myopic behavior. This study also indicates that conducting an out-of-sample evaluation is necessary to assess the value of a dynamic portfolio policy because in-sample analysis can be misleading.