VAR estimation using various extreme value theory methods

Portfolio risk management is a complicated process, which requires an attentive data analysis and a strict evaluation of losses. Value at Risk (VaR) is a financial tool, which is usually used for risk management. Unfortunately, estimating VaR in a common way, we can miss some possible losses. If we apply extreme value theory for estimating VaR, we can get more reliable results of financial losses. Extreme value theory is widely used in finance and insurance to forecast rare and very extreme events. This thesis focuses on methods of extreme value theory,such as POT and Hill estimator, and applying them to to estimate daily VaR of "Microsoft" stock prices for a 4 year-period. This thesis shows the reliability and utility of using extreme value theory to estimate VaR, and compare these results with with those obtained using normal distribution