Statistical Inference in Two-Parameter Portfolio Theory with Multiple Regression Software

Abstract

The purpose of this paper is to demonstrate how multiple regression software may be used for computing estimates of efficient set parameters and for performing tests of mean-standard deviation efficiency. Regression software also is shown to be useful for selecting, from a set of assets, a subset that maximizes performance and for comparing the performance of the set to the subset. The underlying multiple regression model fitted by the software has no relation to the analysis; the regression software is employed simply as a computing device. Since the multiple regression procedure is familiar to most finance researchers and since regression software is commonly available, the techniques presented here should be of wide interest.