Measuring Value at Risk (VaR) of Exchange Rate in Iran

This study attempts to estimate value at risk using GARCH model at 95% confidence level for exchange rate return in Iran, and then selects appropriate model, comparing the results obtained from models with actual behavior of exchange rate and performing statistical tests. In the analytical part of the study in order to identify the state of used variables, ARCH test were used to determine ARCH effect, and GARCH models were used to obtain value at risk. In this study, daily data of exchange rate during the period May 12, 2003 to March 20, 2012 have been used to test the performance of provided models in explaining behavior of the average and returns volatilities, and calculating the value at risk. Based on the results, all coefficients of the value at risk except simple GARCH with normal distribution at 5% level were significant, and value at risk for this indicator showed that normality assumption of shocks’ distribution indicates value at risk significantly lower. Moreover, the results showed that simple GARCH model with normal error distribution indicates value at risk at the least level, and EGARCH model with t-error distribution indicates value at risk at the highest level. Therefore, a further increase in value at risk for the considered indicator would lead to wider distribution tails of that indicator.