Seed Rasekhi; Amir Khanalipour
Volume 1, Issue 1 , January 2011, , Pages 101-132
Abstract
This paper has examined the long memory of oil market volatility. For this purpose, the paper has employed different types of long run ARCH models including FIGARCH-BBM, FIGARCH-chung, FIEGARCH, FIAPARCH-BBM and FIAPARCH-chung and short run ones including GARCH, EGARCH, GJR AND APARCH with three different ...
Read More
This paper has examined the long memory of oil market volatility. For this purpose, the paper has employed different types of long run ARCH models including FIGARCH-BBM, FIGARCH-chung, FIEGARCH, FIAPARCH-BBM and FIAPARCH-chung and short run ones including GARCH, EGARCH, GJR AND APARCH with three different assumptions of normal, t-student and generalized error distributions. Results obtained from all long run models indicate the volatility persistence, i.e. the long memory of oil market volatility. Furthermore, with regard to Akaike’s information criterion, FIAPARCH-chung with assumption of t-student distribution has the best performance. Also, according to Schwarz Criterion, FIGARCH-chung model with assumption t-student distribution is the best model in modeling volatility of oil market. Based on the results, long run models considering long memory property of volatility indicate a better performance than the short run ones. Finally, based on obtained results, asymmetric distributions including t-student and GED are found to be more suitable than normal distribution.
