Momentum Trading and Performance with Wrong Return Expectations

Abstract

Many long-term institutional investors use strategic portfolio allocation to maximize net asset value assuming that long-run expected returns are constant. This approach is complemented by tactical asset allocation when an asset class is perceived to offer higher expected returns in the short run. When investment performance is evaluated in the interim, it can be difficult to assess the merit of private beliefs about changing expected returns. The authors examine the annual expected return and dynamic trading under wrong assumptions about time-varying expected returns and show that a misspecified portfolio policy can outperform the correct policy. They also find that the particular assumptions about how expected returns vary through time can be identified from observed rebalancing of a portfolio. The assumption that expected returns are changing implies that momentum trading is optimal under certain conditions. In contrast, contrarian trading is always optimal when prices are assumed to follow a random walk around a constant long-term trend. These results can inform the annual performance evaluation of institutions, such as endowments, pension plans, mutual funds, and hedge funds that have long-term investment objectives.

TOPICS: VAR and use of alternative risk measures of trading risk, statistical methods, portfolio construction