| Summary: | 碩士 === 東吳大學 === 企業管理學系 === 91 === Financial asset returns are known to be nonnormal and tend to be leptokurtic with fat tails — as is commonly found in academic research. The Standard EWMA estimator with the normality assumption (used in JP Morgan''s RiskMetrics® model) will be inefficient and lead to understate the true Value at Risk (VaR) if the asset returns are fat-tailed. We employ a range of EWMA family estimators such as Power EWMA, Standard EWMA, and Robust EWMA proposed by Guermat & Harris (2002) to forecast the VaR based on the daily closing indices of TSEC Taiwan 50, TAIEX, FTSE 100, and DJIA. The results demonstrate that the members of the family of EWMA estimators with the power exponential distribution (also known as the generalized error distribution) have improved the inability to capture the non-normality returns in estimating VaR over the Standard EWMA estimator, and show that Power EWMA performs an excellent accuracy in VaR estimation.
|
|---|
