Long Memory usage in Portfolio Optimization using the Copula Functions: Empirical evidence of Iran and Turkey Stock Markets

The main objective of this paper is to optimize and manage the portfolio by using copula functions. Copula function has been using as a powerful and flexible tool for the determination of dependency structure. Research data include the Iran stock market index and the Turkey stock market index. The present study seeks to find the effect of long memory on the structure of dependence between returns and optimal portfolio structure. In the first step, we compare the dependence structure between the net returns and the filter generated from the ARFIMA-GARCH process returns to investigate the impact of long memory on them. In the second step, the influence of the dependence structure between net returns and filtered returns on portfolio optimization has been investigated. The results indicated that the model can be fitted to the return of time series and the best pattern is the frank pattern. The results also indicated the existence of long memory in the mean and variance of stock return on the Iran stock market and the existence of long memory in the variance of the Turkey stock market. All models allocate more percentage of capital to Iran stock market and lower percent to Turkey stock market.