本研究係使用辛普森數值方法求標準化偏態t分配在不同厚尾、偏態參數值下之下方百分位值,再使用GARCH-N、GARCH-T與GARCH-ST模型來分別對台灣、南韓、新加坡與馬來西亞等四個國家股票市場指數進行90%、95%、99%及99.5%四種信賴水準一天期風險值之估計與預測。實證結果顯示,四個國家股票市場指數日報酬率的分配確實具有高峰、厚尾之特性且南韓、新加坡等國家股價指數日報酬率的分配呈現左偏而台灣、馬來西亞等國家股價指數日報酬率的分配呈現右偏。在低信賴水準下,GARCH-N模型最為準確。在高信賴水準下,GARCH-N模型準確性最差而南韓、新加坡等國家股價指數之GARCH-ST模型準確性最好,另台灣、馬來西亞等國家股價指數之GARCH-T模型準確性最好。
This investigation proposes a composite Simpson's rule, a numerical integral method, for estimating quantiles on the skewed student t (ST) under different fart-tailed and skewed parameters. Daily spot prices of TSEC, KOSPI, STRAITS and KLSE stock indices are used as data to examine the one-day-ahead VaR forecasting performance of the GARCH-N, GARCH-T and GARCH-ST models under 90%, 95%, 99% and 99.5% confidence levels. Empirical results show that the distribution of all asset returns exhibit leptokurtic and fat-tailed features. Moreover, it exhibits left-skewed distribution for KOSPI and STRAITS stock indices whereas right-skewed distribution for TSEC and KLSE stock indices. Furthermore, the GARCH-N models provide the most accurate VaR forecasts for low confidence levels. For high confidence levels, the GARCH-ST models provide the most accurate VaR for KOSPI and STRAITS stock indices whereas the GARCH-T models provide the most accurate VaR forecasts for TSEC and KLSE stock indices.