Innovations in the ARIMA - GARCH Modeling in Forecasting Gold Price

• Main Authors: Siti Roslindar, Yaziz, Roslinazairimah, Zakaria, Noor Azlinna, Azizan, Maizah Hura, Ahmad, Agrawal, Manju, Boland, John
• Format: Conference or Workshop Item
• Language: English
• Published: 2014
• Subjects: Q Science (General)
• Online Access: http://umpir.ump.edu.my/id/eprint/8613/1/fist-2014-roslinazairimah-Innovations_in_the_ARIMA.pdf; http://umpir.ump.edu.my/id/eprint/8613/; https://drive.google.com/file/d/0B02jW7Y1R3ICX0s0dG9XN3I1dVU/view?usp=sharing

Gold has been the most popular commodity as a healthy return investment due to its unique properties as a safe haven asset. Therefore, it is crucial to develop a model that reflects the pattern of the gold price movement since it become very significant to investors. In developing a model, the innovations for the standardized error in diagnostic checking should be chosen appropriately to make the model fit and adequate to the data. Previous study showed that hybrid of ARIMA-GARCH is a promising approach in modeling and forecasting gold price. In this study, we employ different innovations to the ARIMAGARCH model to provide a better understanding in the modeling of gold price series. The innovations in this study are Gaussian, t, skewed t, generalized error distribution and skewed generalized error distribution. By applying the hybrid model to daily gold price data from year 2003 to 2014, empirical results indicate that the ARIMA-GARCH with t innovations was found to perform better and fits the data reasonably well due to the heavier tails characteristics in the data series.