Making Second Order Stochastic Dominance inefficient Mean Variance Portfolio efficient: Application in Turkish BIST-30 Index

The classical Mean Variance (MV) portfolio optimization has some weaknesses which do not satisfy today’s financial needs when working with real data. At the core, among other shortcomings, the requirement of normal distributed returns renders the MV optimized portfolios Second Order Stochastic (SSD) inefficient. In this paper, a new gradual method is introduced which eliminates the SSD inefficient stocks in the first step and applies the MV optimization in the SSD efficient subset in the second step. As an empirical example, the suggested method is applied to the Turkish BIST-30 Index. Once the application is completed, the Sharpe Ratio maximizing MV optimized portfolio of the suggested method and one of the original BIST-30 are compared with each other. The results show that although SSD efficient portfolio has a lower Sharpe ratio, a metric widely used in the market, risk averse investors take higher moments into account while making investment decisions. So the gradual method introduced in this paper expands the perspective of MV optimization by applying the SSD efficiency criteria as a pre screen.