A Note on Errors in Variables and Estimates of Systematic Risk

Abstract

Australian capital markets, being relatively “thin”, present the researcher with a potentially large “errors in the variables” problem. The standard formulation of the problem is inappropriate for estimating securities' systematic risks. The unusual feature in this context is that the regressor (the rate of return on the “market index”) is defined as the sum of the feasible regressands (the rates of return on individual securities). As a result, errors in the regressor are not independent of errors in the regressand. An alternative formulation is suggested, which is simple and similar to the conventional formulation. While the alternative is more realistic, it is not sufficiently so: the real problem of errors in this context is much more complex.