Forecasting Daily Value-at-Risk of the Tehran Stock Exchange Index using Realized GARCH Approach

As a measurement of risk, Value-at-Risk has always been attractive to researchers in the field of risk management and capital markets. This importance has always led them to develop models to increase the accuracy of VaR estimation. GARCH models are used to estimate conditional variances in parametric approaches of calculating VaR. In the latest developments, Hansen presented the Realized GARCH model using intra-day data. Regarding the recent high fluctuations of the Tehran Stock Exchange Index, the use of Realized GARCH models to compute VaR can increase the accuracy of forecasting. In this paper the three different distributions, Normal, T, and GED, are combined with conventional GARCH models, and the new Realized GARCH model. Next, the VaR of the Tehran Stock Exchange index is forecasted using the rolling window sampling method. And finally, the accuracy of predicted VaR has been evaluated and compared using a two-step Backtest method. The outcome of this study indicates that using the new Realized GARCH model in forecasting VaR will tend to result in more accurate estimates, whether at the 5% or 1% level.