There is an inherent problem with comparing and ranking competing Value at Risk (VaR) and Expected shortfall (ES) models since we are measuring only a single realization of the underlying data generation process. The question is whether there is any significant statistical difference in the...