Ranking portfolio analysis

Abstract

Purpose: the authors apply the econometric model with a discrete dependent variable to build a portfolio of securities. Discussion: the authors note that Sharpe used a linear econometric model for a diagonal model of portfolio investment building. The authors propose to use a nonlinear regression model with a discrete dependent variable instead of a linear model to build a portfolio. This allows you to reflect the binary nature of a financial asset yield with the use of probability obtaining the positive result from investments in the asset. As a result, probabilistic estimates become a criterion on the basis of which it is necessary to form a portfolio. This requires a new approach that uses principles different from those of the optimization approach. Results: the authors proposed to form a portfolio of securities on the basis of a probability preference matrix. It is shown that the eigenvector of this matrix determines the structure of the portfolio, which is called rank because of its properties. The properties of the ranking portfolio and the method of its construction significantly expand the capabilities of the portfolio analysis apparatus.