|Banks using advanced VaR models are expected to hold in a lower amount subject to market risk (ASMR) than banks using simple VaR models because of measuring their risk relatively more accurately. The purpose of this study is to test the hypothesis that advanced VaR models which measures risks better are resulting a lower ASMR. In this study historical volatility, historical simulation, EWMA, GARCH (1,1), GARCH (1,1)-Bootstrap and GARCH (1,1)-GED models were used for VaR calculations. By backtesting the VaR measures the model security factor h has been identified and so the ASMR has been simulated. After the results have been discussed for the real data sets the same process was repeated with randomly generated six different data sets to test the consistence of the results. According to the findings, the hypothesis that advanced VaR models like GARCH (1,1)-Bootstrap and GARCH (1,1)-GED provides a lower ASMR was rejected.