Using ARFIMA and FIGARCH methods in Markowitz mean variance portfolio optimization: An application on ISE-30 index stocks

In finance literature, there are some problems about Markowitz mean variance portfolio optimization model. One of these problems is how to determine the expected return of stocks which are used in calculations of portfolio optimization. In this study, whether enhanced optimized portfolios may be obtained via using fractional integrated models that ensure return forecast is examined. Return forecast data is obtained via ARFIMA model, and variance forecast data is obtained via FIGARCH model and then, dynamic portfolio optimizations for 42 months is formed by using obtained data. Performances of these portfolios are compared with equivalent dynamic optimized portfolios which use classical Markowitz expected returns. According to the results, the hypothesis investigated is not supported on ISE-30 Index stocks for forecast period including “Mortgage Crises”, which is originated from USA.