Studying Long Memory of Tehran Stock Exchange

Abstract

According to the efficient market hypothesis, prices in stock market follow the random walk theory. In such a market, data is published rapidly and affect the stock price. Thus a large proportion of financial theory is extended based on random walk models for assets prices and returns. Long memory is one of violation of efficient market that tells time series of capital market index does not follow the random walk theory. This study examined long memory as one of the properties of time series for Tehran dividend stock price index and industrial index to test the assumptions, Autoregressive Fractionally Integrated Moving Average (ARFIMA) model is used. The results indicated that the concept of long memory existed for two index in Tehran stock exchange. The result of this research shows that there is a conception of long memory in Capital Market of Iran. Not only capital market efficiency does not obey the random walk as Dickey Fuller test, but also it has long memory. Also its time series is not unstable, in other words it is steady.