Abstract
Our article contains three critical elements. First, by way of motivation, we consider active optimal mean-variance portfolios within the confines of a three-risky-asset model (based on common equity, corporate bonds, and long-dated Treasury securities) and a short-term, riskless security. Within that context, our article models and numerically demonstrates the impact of the magnitude and sign of the interest-rate term premium in optimal asset allocation. Then, based on interest-rate survey data, we demonstrate the occurrence of a negative term premium in the Treasury market, and proceed to permit such an occurrence within a model whereby asset returns covary with a market portfolio. Finally, turning to empirics, we document the relationship between the sign of the term premium and the sign of the covariance between Treasury and equity returns. Recognizing the deficiency in using historical data as inputs to active optimal mean-variance portfolios, we seek to obtain forward-looking inputs to render our model prospective (rather than merely retrospective): we design and implement tests for obtaining forward-looking measures of covariation.