Modeling and Forecasting Tehran Stock Exchange Return using ARFIMA and FIGARCH

During the last decades long-memory processes have evolved as an important part of time series analysis and long memory parameter in asset returns has important evidences for many paradigms in modern finance theory. If asset returns display long memory, the series realizations are not independent over time, realizations from the remote past can help forecast future returns. Therefore the presence of long memory in asset returns contradicts the weak form of the market efficiency hypothesis. Also this characteristic has fundamental effect on time series prediction methods. This research examines the presence of long memory in series of return and volatility of Tehran Stock Exchange. Significance evidence of long memory is found in first and second moments of Tehran Stock Exchange return series. Also predictive accuracy of AMRA, GARCH, ARFIMA and FIGARCH models compared in variety of forecast horizons with recursive and rolling estimation schemes. The results of this research show that ARMA model perform better in 1-step ahead forecast, while for greater forecast horizons, including weekly, monthly, seasonal and yearly predictions, FIGARCH model outperform other alternatives.