摘要 : This study evaluates the downside tail risk of coal futurescontracts (coke, coking coal and thermal coal) traded inthe Chinese market between 2011 and 2021, measured byvalue at risk (VaR). We examine the one-day-aheadVaRforecastin... 展开 This study evaluates the downside tail risk of coal futurescontracts (coke, coking coal and thermal coal) traded inthe Chinese market between 2011 and 2021, measured byvalue at risk (VaR). We examine the one-day-aheadVaRforecasting performance with a hybrid econometric anddeep learning model (GARCH-LSTM),GARCH familymodels, extreme value theory models, quantile regressionmodels and two naïve models (historical simulationand exponentially weighted moving average). We use fourbacktesting techniques and the model confidence set toidentify the optimal models. The results suggest that themodels focusing on tail risk or utilising long short-termmemory generate more effective risk management. 收起