Volatility Modelling and Parametric Value-At-Risk Forecast Accuracy: Evidence from Metal Products

Abstract

In this paper, we investigate the one-day-ahead VaR and ES accuracy of four metal daily return series including Aluminium, Copper, Nickel and Zinc. Since, all sample presents volatility clustering, volatility asymmetry, and volatility persistence, we have assessed five GARCH-type models including three fractionary integrated models assuming three alternative distributions (normal, Student-t and skewed Student-t distributions). Estimates results reveal the performance of AR (1) - FIAPARCH model under a skewed Student-t distribution. We have computed one-day ahead VaR and (ES) for both short and long trading positions. Backtesting results show very clearly that the skewed Student-t FIAPARCH model provides the best results for both short and long VaR estimations. These results present several potential implications for metal markets risk quantifications and hedging strategies.