基于双曲线记忆HYGARCH模型的动态风险VaR测度能力研究

发布时间：2018-01-29 18:20

本文关键词： 金融市场典型事实 HYGARCH 动态风险测度　出处：《中国管理科学》2011年06期 　论文类型：期刊论文

【摘要】：本文针对金融市场的典型事实特征,运用自回归分数移动平均(Fractional Integrated Autoregressive Moving Average,ARFIMA)模型与双曲线记忆广义自回归条件异方差模型(Hyperbolic Memory Generalized Autoregressive Conditional Heteroscedasticity,HYGARCH)模型、分数协整非对称自回归条件异方差(Fractional Integrated Asymmetric Power Autoregressive Conditional Heteroscedasticity,FIAPARCH)模型和分数协整指数广义自回归条件异方差(Fractional Integrated Exponential Generalized Autoregressive Conditional Heteroscedasticity,FIEGARCH)模型结合,并运用有偏学生t分布(Skew Student t Distribution,SKST)来捕获金融收益分布形态,以此开展动态风险测度研究,进而运用返回测试(Back-Testing)中的似然比率测试(Likelihood Ratio Test,LRT)和动态分位数回归(Dynamic Quantile Regression,DQR)方法对风险模型的准确性与精度进行联合检验。通过实证研究,得到了一些非常有价值的实证结论:ARFIMA(1,d,1)-FIAPARCH(1,d,1)-SKST模型与ARFIMA(1,d,1)-HYGARCH(1,d,1)-SKST模型均表现出卓越的风险测度能力,但没有绝对优劣之分;ARFIMA(1,d,1)-FIEGARCH(1,d,1)-SKST模型在成熟市场的表现能力差强人意;本文引入的所有风险模型在中国大陆沪、深股市表现优越且没有实质性差异。
[Abstract]:The typical facts according to the characteristics of the financial market, using the fractional autoregressive moving average (Fractional Integrated Autoregressive Moving Average, ARFIMA) model with hyperbolic memory generalized autoregressive conditional heteroskedasticity model (Hyperbolic Memory Generalized Autoregressive Conditional Heteroscedasticity, HYGARCH) model, fractional cointegration asymmetric autoregressive conditional heteroskedasticity (Fractional Integrated Asymmetric Power Autoregressive Conditional Heteroscedasticity, FIAPARCH) model and fractional cointegration exponential generalized autoregressive conditional heteroskedasticity (Fractional Integrated Exponential Generalized Autoregressive Conditional Heteroscedasticity, FIEGARCH) model combination, and the use of biased Student t distribution (Skew Student t Distribution, SKST) to capture the financial income distribution, in order to carry out the dynamic risk measure research, and then use the Return test (Back-Testing) in the likelihood ratio test (Likelihood Ratio, Test, LRT) and dynamic quantile regression (Dynamic Quantile, Regression, DQR) joint inspection accuracy and precision of the risk model. Through empirical research, obtained some valuable empirical conclusions: ARFIMA (1, D, 1) -FIAPARCH (1, D, 1) -SKST model and ARFIMA (1, D, 1) -HYGARCH (1, D, 1) -SKST model showed a risk measure of extraordinary ability, but there is no absolute good or bad; ARFIMA (1, D, 1) -FIEGARCH (1, D, 1) -SKST just passable performance ability the model in mature markets; this paper introduces all the risk model in Chinese, Shanghai and Shenzhen stock market performance and no substantive differences.

【作者单位】：成都理工大学商学院;
【基金】：国家自然科学基金资助项目(71171025,71071131,70771097) 教育部“新世纪优秀人才支持计划”(NCET-08-0826),教育部人文社会科学研究青年基金(10YJCZH086) 成都理工大学高层次人才科研启动基金(HJ0038),成都理工大学优秀创新团队培育计划项目(2010TD01)
【分类号】：F224;F830.9
【正文快照】： 1引言近年来,全球金融市场又发生了一系列重大风险事件(Risk Events)或称危机事件(Crisis E-vents),如:1997年的亚洲金融危机、2007年的美国次贷危机转化为全球性“经济危机”以及最近发生的迪拜财务危机事件,等等。所有这些事实,又一次证明了:维护金融经济安全,推动经济的发

【参考文献】