基于EGARCH-GED模型下的股市风险测度研究.PDF

第 34卷 第 2期 () http: / /www. ku stjourna.l com / V ol 34 N o12 2009年 4 月 Journal ofK unm ingU n iversity of Science and Technology ( Science and Technology) Ap r 2009 do:i 10 3969 / j issn 1007- 855x 2009 02 02 3 EGARCH - GED 王喜报, 刘文奇 ( 昆明理工大学 理学院, 云南 昆明 650093) : 根据我国股市市场收益的基本特征, 首先运用AR - EGARCH 模型来捕获上海证综合指数 收益序 的自相关性、波动集聚性和杠杆效应; 然后利用广义误差分布估计其厚尾分布, 建立了 能准确度量时变风险价值的AR - EGARCH - GED模型, 并与基于正态分布和学生 t分布的AR - EGARCH 模型所计算的风险价值效果进行比较 最后, 通过实证分析, 并利用后验测试, 表明 基于AR - EGAR CH - GED模型的风险价值能更好地刻画我国股市的市场风险 : VAR; AR - EGARCH 模型; 广义误差分布; 后验测试; 股市风险; 风险度量 : F830 91 : A : 1007- 855X( 2009) 02 - 0 104 - 04 Stud on R iskM easur em ent in Ch inese Stock M arket Based on EGARCH - GED M odel W ANG X2i bao, LIU W en2q i ( F acu lty of Scien ce, Kunm in g Un iversity of Scien ce and T echnology, Kunm ing 650093, Ch in a) Abstr act: F irstly, b ased on the b asic characterist ics of return series in Ch in ese Stock M arket, AR - EGAR CH m ode l is u sed to ob ta in its e ffect of autocorre la tion, c lu stering fluctu ation an d especia lly the / leverage effect0 of flu ctuat ion Gen eralized error distribution is then adop ted to estmi ate th e fat tail of return sequ en ces Th e AR2 EGAR CH - GED m odel for ca lcu lat ing value at risk is estab lish ed and then comp ared w ith th e AR - EGAR CH m ode ls based on n orm a l distribution an d t distribu tion F in ally, an emp irica l an alysis in comb in ation w ith p oste2 rior testing show s th at th e AR 2EGARCH - GED m ode l is a b etter m ethod to describe th e m ark et risk in Ch inese Stock M ark et