Investigating the Long-Term Memory of Total Price Index of the Tehran Stock Exchange (A Case Study: Pharmaceutical Industry)

This study meets two first، investigating the presence of long-term memory of total price index in the pharmaceutical industry of the Tehran Stock Exchange; second، evaluating the accuracy of predicting models which include the long-term memory of the total price index of this industry. In this research، the methods of Maximum Likelihood (MLE)، Whittle، GPH، and Sperio have been used in order to evaluate the fractional integration parameter (market memory). In the beginning، among the four above-mentioned methods، two methods of MLE and Whittle could process the best pattern of ARFIMA to the data. Then، by using statistics and choosing criteria for the best one، this pattern was selected as the best for evaluating by MLE method. The total price index in the pharmaceutical industry of the Tehran Stock Exchange has a long-term memory، and the best pattern for predicting the stock total price index of pharmaceutical industry is ARFIMA (1، 0. 13، 1). It can be claimed with high certainty that the time-series of total price index in the pharmaceutical industry of the Tehran Stock Exchange has a long-term memory. The existence of this feature is a good reason for rejecting the weak shape of the market efficiency assumption. According to this assumption، the price of assets should not be predictable by historical data. The existence of long-term memory of total price index in the pharmaceutical industry indicates self-correlation among evidence with high intervals. Thus، it is possible to use historical data in order to predict the future indices which will be provided by applying a profitable strategy.