The dynamics of precious metal markets VaR: A GARCHEVT approach

The data analysis of the metal markets has recently attracted a lot of attention, mainly because the prices of precious metal are relatively more volatile than its historical trend. A robust estimate of extreme loss is vital, especially for mining companies to mitigate risk and uncertainty in metal price fluctuations. This paper examines the Value-at-Risk and statistical properties in daily price return of precious metals, which include gold, silver, platinum, and palladium, from January 11, 2000 to September 9, 2016. An advanced two stage approach which combines GARCH-type models with extreme value theory is implemented. In the first stage, the conditional variance is modeled by different rolling univariate GARCH-type models (GARCH, EGARCH and TGARCH) under the GED error assumption in the returns of precious metal markets and compare the same with other well-known models. In the second stage, extreme value approach is applied to capture the tail behavior of distribution for the extracted standardized residuals. In comparison with the dynamic VaRs of these precious metals, we find that gold has the most steady and the highest VaRs, followed by platinum and silver; on the other hand our results show that palladium has the most volatile VaRs. The backtesting result confirms that our approach is an adequate method in improving risk management assessments and hedging strategies in the high volatile metal markets.