Developing Non-linear Dynamic Model to Estimate Value at Risk, Considering the Effects of Asymmetric News: Evidence from Tehran Stock Exchange

Empirical studies show that there is stronger dependency between large losses than large profit in financial market, which undermine the performance of using symmetric distribution for modeling these asymmetric. That is why the assuming normal joint distribution of returns is not suitable because of considering the linier dependence, and can be lead to inappropriate estimate of VaR. Copula theory is basic tool for multivariate modeling, which is defined by using marginal and dependencies between variables joint distribution function. In addition, Copulas are able to explain and describe of complex multiple dependencies structures such as non-linear dependence. Therefore, in this study, by combining symmetric and asymmetric GARCH model for modeling the marginal distribution of variables and Copula functions for modeling financial data and also use of DCC model to determine the dynamic correlation structure between assets, try to estimate the Value at Risk of investment portfolio consists of five active index In Tehran Stock Exchange. The results demonstrate excellence of GJR-GARCH(1,1) with the distribution of t-student for marginal distribution. t-Copula model, estimates the Value at Risk model less than the Gaussian Copula in all cases.