THE EFFECT OF ERRORS IN MEANS, VARIANCES, AND COVARIANCES ON OPTIMAL PORTFOLIO CHOICE

How should an analyst allocate resources to estimate means, variances, and covariances in mean/variance portfolio selection problems? Evidence indicates that errors in means are over ten times as damaging as errors in variances, and over twenty times as damaging as errors in covariances. Moreover, the relative impact of errors in means is even greater at higher risk tolerances. Hence, the emphasis should be on obtaining the best mean estimates, followed by good variance estimates, followed by decent covariance estimates.