基于APARCH-st模型的ES计算  

The computing of ES based on APARCH-st model

作　　者：王泓娜[1] 张金凤[2] 

机构地区：[1]辽宁师范大学数学学院,辽宁大连116029 [2]天津工业大学理学院,天津300160

出　　处：《辽宁师范大学学报（自然科学版）》2010年第4期416-419,共4页Journal of Liaoning Normal University:Natural Science Edition

摘　　要：为了刻画金融时间序列的尖峰厚尾性及股市中存在的杠杆效应,Ding、Granger和Engle(1993)提出了APARCH模型,但在关于此模型的研究中,通常假设条件分布为正态分布或对称厚尾分布,如t分布、GED分布.在兼顾波动性和分布两个方面的基础上对收益率序列建立APARCH-st模型,给出其VaR和ES的计算方法,并与基于APARCH-t模型的ES值相比较,结果表明,考虑偏态分布得到的ES估计更为保守.In order to depict the fat-tailed and leptokurtic properties of financial time series and leverage effect in stock market,Ding,Granger and Engle proposed APARCH model in 1993.It is usually assumed that in this model,the distribution of residuals is normal distribution or symmetrical fat-tailed distribution such as t and GED distribution.We set up APARCH-ST model and provide Value-at-risk（VaR） and expected shortfall（ES） Computing Method based on the consideration of volatility and distribution of the returns ratios.The result shows that compared with the ES value based on APARCH-t model,the ES estimation considered asymmetrical distribution is more conservative.

关 键 词：APARCH模型 st分布 VAR ES 

分 类 号：O213[理学—概率论与数理统计] F830.9[理学—数学]