Uç Değerler Teorisi ve Riske Maruz Değer

Abstract

Value-At-Risk (VaR) is a method, measures the maximum loss of the investment for a given confidence interval and holding period. When the classical VaR models are analyzed, it is shown that these models are based on the normal distribution assumption. Because of the fact that financial return series have fatter tail than the normal distribution, VaR forecasts, calculated under the normal distribution assumption, don t reflect the real market conditions. As a result of literature review related to topic, Extreme Value Theory (EVT) is used to model fat-tail structure in financial returns series. At thesis study, firstly classical VaR models and then EVT with theoretical structure are analyzed comprehensively. In application part, banks in BİST-30 index, using their daily return series, VaR forecast of static VaR models are obtained; results are interpreted statistically and financially.