Portfolio Optimization Using Markowitz’s Mean-Semi Variance Method on Tehran Stock Exchange

Subject Areas : Futurology

F. Heibati ¹ , R. Haddadzadeh ²

1 - نویسنده مسئول
2 - ندارد

Received: 2006-09-18 Accepted : 2007-04-16 Published : 2008-03-20

Keywords: Modern Portfolio Theory, Post-Modern Portfolio Theory, Downside Risk, Mean-Semivariance, Efficient Frontier,

Abstract :

This paper presents an alternative to the standard mean-variance efficient frontier model underpinningmodern portfolio theory applications. We present a more practical alternative - A Mean-Semivarianceefficient frontier - that takes into consideration advances in “Post Modern Portfolio Theory” as it pertains toasset allocation. The implications for advisors and planners are profound:An investor’s minimal acceptable return is a critical determinant of their optimal portfolio;Efficient frontiers need not be continuous, reflecting the realities of market behavior and demonstratedvolatility;This new method for calculating frontiers is a preferred device for developing strategic asset allocations.What is so efficient about the “efficient frontier?” The Markowitz (1952) model is employed to determinethe optimal mix of risky securities. This methodology involves determining what the minimum riskcombination of securities or asset classes is for a given return. Risk, in this context, is defined as the standarddeviation of returns of a composite portfolio. By plotting these risk-return combinations, an “efficientfrontier” is generated (with the “efficient” part being the upper boundary). Is this really the preferred way tolook at risk and use it in portfolio selection? If it is not, then is this frontier really “efficient?”In this paper we have modified the traditional Markowitz paradigm by redefining risk. The definition of riskwe employ in this paper is “Semi-Standard Deviation” instead of “Standard Deviation”. Then we haveconstructed efficient frontier for top fifty securities of Tehran stock exchange using Downside Risk approachor “Mean-Semivariance” method. In this research we achieved more efficient frontier using this method thanthe traditional one.