Estimating of value at risk and expected shortfall by using conditional extreme value approach in Tehran Securities Exchange

This paper investigates the relative performance of Value-at-Risk (VaR) and expected shortfall (ES) models using daily overall index data from TSE for a period of 8 years from 2008 to 2016. The main emphasis of the study has been given to Conditional Extreme Value Theory (CEVT) and to evaluate how well Conditional EVT model performs in modeling tails of distributions and in estimating and forecasting VaR and ES measures. We also compare them with parametric approaches. We have compared the accuracy of Conditional EVT approach to VaR and ES estimation with other competing models. We use Bernoulli coverage and Independence of violation tests for backtesting the VaR models and McNeil & Frey’s Backtest and Model Confidence Set to assess the performance of the ES models. The best performing VaR and ES models is found to be the Conditional EVT. MCS function result for ES also shows that the Conditional EV with student's t standardized residuals, Conditional EV with normal standardized residuals and GARCH with student's t residuals models are respectively ranked first to third.