This paper investigates the Value at Risk (VAR) for Tehran Stock exchange using, four GARCH- type models, including GARCH(1,1), GARCH, Risk Metrics, and GARCH with optimal number of lags, Because most return series show fat-tailed distribution, we also consider the models with t-distributed errors. The results show that these types of models are quite successful in modelling average and variance of and estimating VAR for the retune series data. Risk Metrics outperforms the other models in modelling the data and estimation of VAR, regardless of the distribution of the errors. We, finally, provide a ranking of the indexes in terms of their VARs.
', './data/jte/coversheet/cover_en.jpg'))