VaR techniques are becoming increasingly popular, but how well do they perform in practice? This thesis deals with the concepts of VaR, and then describe in detail the six methods for computing it: the historical simulation, the variance-covariance methods using moving average and exponential average, Hull and White model, and Monte-Carlo simulation using moving average and exponential average. It shows that different methods of computing VaR generate widely varying results, suggesting the choice of VaR method is very important. This thesis examined it using daily data on different exchange rates and stock indices and portfolio. And it used two methods to verify the accuracy of the model: estimates based on the binomial distribution and the mean absolute percentage error measure.
The purpose of the study is to provide the risk managers of financial institutions with the meaning of VaR value and the appropriate VaR methods.
