Choosing an optimal Model for Explaining & Forecasting the Volatility of Iranian Gold Price Returns: a Comparison of GARCH, IGARCH & FIGARCH Models

Subject Areas : Financial engineering

Mahdi Shahrazi ¹ (Assistant Professor of Economics, Golestan University)

Received: 2016-10-27 Accepted : 2017-01-12 Published : 2017-06-22

Keywords: FIGARCH, Volatility Modeling, Gold returns, Long-memory, Forecasting Performance,

Abstract :

This paper compares three models of the GARCH family to investigate the volatility dynamics of gold Price returns. Nowadays, GARCH-type models have been extensively used in modeling the volatility process of various asset price returns. Gold plays a critical role as a hedge against adverse market conditions. An accurate understanding about the gold volatility is important for the financial assets pricing, risk management, portfolio selection hedging strategies and value-at-risk policies. In this study, we use Iranian gold returns data from March 25, 2003 to December 25, 2015 and employ the GARCH(1,1), IGARCH(1,1) and FIGARCH(1,d,1) specifications. The research findings show that the FIGARCH is the best model to capture dependence in the conditional variance of the gold returns. Moreover, we examine the long memory behavior in the volatility of gold returns. According to the estimation results, the long memory parameter is positive and statistically significant. Consequently, long memory is an important characteristic of the gold volatility returns and should be taken into consideration in investment decisions. Also, the out-of-sample evaluation criteria (MAE, RMSE and TIC) select the FIGARCH(1,d,1) as the best forecasting model of gold volatility.