基于跳扩散过程的ETF基金动态市场风险测度研究

发布时间：2018-03-11 21:50

本文选题：基金　切入点：风险　出处：《管理评论》2017年03期 　论文类型：期刊论文

【摘要】：本文分别构建了基于双指数-跳扩散过程及双因素-跳扩散过程的ETF基金收益率模型,研究认为后者对于价格的预测较为准确。建立了时变EVT-POT-GPD方法来确定ETF基金收益率标准残差的分位数,据此提出了动态市场风险测度方法,并以中国、香港、美国ETF基金的样本进行实证研究。结果表明,采用双因素模型并以一年期历史数据作为窗口数据进行ETF基金价格预测时的效果较好,坏消息及好消息对于ETF基金收益率的冲击具有较显著的非对称性影响和杠杆效应,而TGARCH模型在度量ETF基金收益率正向非对称性冲击及条件异方差特性时的效果较好。采用四种GARCH模型得到的ETF基金动态CVaR值均大于VaR值,这说明本文所构建的动态风险测度方法在风险估计方面更为保守和有效。当置信水平从95%上升到99%时,ETF基金的CVaR的增长率要快于VaR的增长率,这也表明当置信水平较高时本文所提出的动态CVaR风险测度方法敏感程度要高于动态VaR的。利用Back-testing及卡方检验发现,随着置信水平的提高,本文提出的动态CVaR及动态VaR测度方法的成功率均上升。
[Abstract]:In this paper, the ETF fund return model based on double exponent-jump diffusion process and double-factor-jump diffusion process is constructed, respectively. In this paper, we establish a time-varying EVT-POT-GPD method to determine the quantiles of the standard residual of the return rate of ETF funds, and then propose a dynamic market risk measurement method. The results show that the two-factor model and one-year historical data are used as window data to predict the price of ETF funds. Bad news and good news have significant asymmetric effect and leverage effect on the impact of ETF fund yield. The TGARCH model is more effective in measuring the positive asymmetric impact of the return of ETF funds and the characteristics of conditional heteroscedasticity. The dynamic CVaR values of ETF funds obtained by using four GARCH models are all higher than VaR values. This shows that the dynamic risk measurement method constructed in this paper is more conservative and effective in risk estimation. When the confidence level rises from 95% to 99, the growth rate of CVaR of ETFs is faster than that of VaR. It also shows that the dynamic CVaR risk measurement method proposed in this paper is more sensitive than that of dynamic VaR when the confidence level is high. With the help of Back-testing and chi-square test, it is found that with the increase of confidence level, Both the dynamic CVaR and the dynamic VaR measure proposed in this paper have higher success rates.
【作者单位】：西安理工大学经济与管理学院;
【基金】：国家自然科学基金项目(71171155) 陕西省教育厅专项科研计划(16JK1527) 西安市社会科学规划基金项目(17J92) 西安理工大学科技创新计划项目(2016CX009)
【分类号】：F831.51

【相似文献】