عنوان مقاله [English]
Value at Risk (VaR) measures risk exposure at a given probability level and is very important for risk management. In this paper, mainly EVT models are compared to other well-known models such as GARCH, Historical Simulation and Filtered Historical Simulation. Then evaluation their models with different back testing such as Kupiec test, Christoffersen test and Lopez Loss function.
Our results indicate that using conditional methods and Extreme Value Theory to forecast Value at Risk, is better than other models. And we should examine different methods for forecast Value at Risk, then select the best method for any tails of distributions.