Combination of different models distribution and volatility to estimate value at risk of the stock index

• Main Authors: Chih-Hao Huang, 黃志豪
• Other Authors: Chang-Cheng Changchen
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/94t72q

碩士 === 長榮大學 === 經營管理研究所 === 98 === Financial assets mostly have the leptokurtosis and the fat-tailed. Therefore, this study considered the characteristics of financial data, using skewness and leptokurtosis distribution, and traditional symmetric distribution was relatively. Using Simply moving average (SMA) method, and financial time series models Exponentially weighted moving average (EWMA(λ=0.94)), GARCH(1,1), GJR-GARCH(1,1), can be more accurate to capture the fat-tailed volatility clustering.This papper using four developed countries sotck index and four Asian tigers to estimate value at risk. The period January 1, 1992 until July 31, 2009, using 99.5%, 99%, 95%, 90% confidence interval, forecast the next day value at risk. Empirical results show that the use of skewed distribution models, and consider the time series models to estimate the value at risk was better than symmetric distribution and SMA method.