Evaluating Market Risk of Coal
Torng-Her Lee, Haimin Chen, Chin-Ping Tsai

Abstract
We use daily price of spot, futures and swap contracts to study the market risk of coal. The Value at Risk (VaR), GARCH (1, 1), diagonal-VECH, diagonal-BEKK and CCC models are applied in this paper. The empirical results demonstrate that both variance-covariance method and GARCH model pass Basle’s back-test at 99% confidence level, which outperforms the historical simulation method. Meanwhile, although variance-covariance and GARCH models also have good performance at 95% confidence levels, most contracts pass back-test except RFM, which falls into yellow zone in variance-covariance method and RSQ, which falls into yellow zone in GARCH method. Historical simulation method is the poorest of three methods to calculate VaR of coal return series. The empirical result of MGARCH shows that cross volatility of RS/RFQ is higher than other combinations in diagonal VECH, CCC and diagonal BEKK model. However, the degree of closeness of the three MGARCH models in explaining all combinations shows no high cross volatility.

Full Text: PDF