Rational explanation for rule-of-thumb practices in asset allocation

Abstract

Naive asset allocation and other ad-hoc techniques are commonly practiced by fund managers in the industry. Such strategies, however, are deemed mean-variance (MV) sub-optimal according to modern portfolio theory. Nonetheless, taking estimation risk into considerations, such practices are consistent with rational theory. In practice, the potential advantage from MV optimization is weighed against the severity of estimation risk. This paper proposes a set of decision rules to determine the optimal fund under estimation risk. A mixed strategy that deploys the proposed decision rules implies a convex improvement in terms of out-of-sample Sharpe-ratio.