| Sumario: | Researchers James Tobin and Harry Markowitz and many other writers have made valuable contributions to the problem of optimal risk decisions by emphasizing analyses of means and variances. Recently, researchers Karl Borch and Martin Feldstein have re-emphasized the lack of generality of mean-variance analysis and evoked a reply from Tobin, as of October 1970. But nobody has replied to the author's paper which suggested that most of the interesting propositions of risk theory can be proved for the general case with no approximations being involved. This same paper pointed out all the realms of applicability of mean-variance analysis and also its realms of non-applicability. In this article, the author reemphasizes an aspect of the mean-variance model that seems not to have received sufficient attention, namely the usefulness of mean and variance in situations involving less and less risk. The Tobin-Markowitz analysis of risk-taking in terms of mean and variance alone is rigorously applicable only in the restrictive cases where the statistical distributions are normally Gaussian or where the utility-function to be maximized is quadratic.
|
|---|