A Comparison of Portfolio Selection Models via Application on ISE 100 Index Data

ALTUN E., Tatlidil H.

11th International Conference of Numerical Analysis and Applied Mathematics (ICNAAM), Greece, 21 - 27 September 2013, vol.1558, pp.1438-1441 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 1558
  • Doi Number: 10.1063/1.4825788
  • Country: Greece
  • Page Numbers: pp.1438-1441
  • Hacettepe University Affiliated: Yes

Abstract

Markowitz Model, a classical approach to portfolio optimization problem, relies on two important assumptions: the expected return is multivariate normally distributed and the investor is risk averter. But this model has not been extensively used in finance. Empirical results show that it is very hard to solve large scale portfolio optimization problems with Mean-Variance (M-V) model. Alternative model, Mean Absolute Deviation (MAD) model which is proposed by Konno and Yamazaki [7] has been used to remove most of difficulties of Markowitz Mean-Variance model. MAD model don't need to assume that the probability of the rates of return is normally distributed and based on Linear Programming. Another alternative portfolio model is Mean-Lower Semi Absolute Deviation (M-LSAD), which is proposed by Speranza [3]. We will compare these models to determine which model gives more appropriate solution to investors.