An Evaluation of Alternative Methods for Estimating Systematic Risk

The available evidence on the properties of return distributions does not convincingly support a particular method for estimating the parameters of the market model. The underlying stimulus for this paper is to resolve this issue empirically by evaluating the performance of three alternative estimational methods to OLS for deriving the beta or systematic risk of securities. These are the mean absolute deviation, a non-parametric method, and the generalised least square method. The empirical tests include estimates of beta using the four estimators, an examination of whether the beta estimates belong to the same underlying distribution, and finally, an evaluation of the performance of the estimators for predicting monthly rates of return.