基于QR-t-GARCH(1,1)模型沪深指数收益率风险度量的研究
|
摘要
以沪深综合指数收益率为研究对象,在t-GARCH(1,1)模型与st-GARCH(1,1)模型的基础上,引入分位数回归,分别建立了QR-t-GARCH(1,1)模型与QR-st-GARCH(1,1)模型。失败率检验结果表明,在5%、2.5%、1%的显著性水平下,加入分位数回归的GARCH(1,1)模型较GARCH(1,1)模型对指数收益率的风险度量效果更好,而且对异常值的稳健性也更强。该模型可对指数收益率的风险特征进行全面描述。
Taking the Shanghai and Shenzhen comprehensive index yields as the research object, based on the models of t-GARCH(1,1) and st-GARCH(1,1), the models of QR-t-GARCH(1,1) and QR-stGARCH(1,1) were established by introducing the quantile regression. The results of failure rate test show that the GARCH(1,1) model which introduced quantile regression is more effective than the GARCH(1,1)model for the risk measurement of exponential yield at the significance level of 5%, 2.5%, 1% respectively,and its robustness for outliers is also stronger. The model can describe the risk characteristics of the exponential rate of return comprehensively.
Taking the Shanghai and Shenzhen comprehensive index yields as the research object, based on the models of t-GARCH(1,1) and st-GARCH(1,1), the models of QR-t-GARCH(1,1) and QR-stGARCH(1,1) were established by introducing the quantile regression. The results of failure rate test show that the GARCH(1,1) model which introduced quantile regression is more effective than the GARCH(1,1)model for the risk measurement of exponential yield at the significance level of 5%, 2.5%, 1% respectively,and its robustness for outliers is also stronger. The model can describe the risk characteristics of the exponential rate of return comprehensively.
引文
[1]Engle R F.Autoregressive condition heteroscedasticity with estimates of the variance of the United Kingdom inflation[J].Econometrica,1982,50:987-1008.
[2]Bollerslev T.Generalized autoregressive conditional heteroscedasticity[J].Journal of Econometrics,1986,(31):307-327.
[3]Engle R,Manganelli S.CAViaR:Conditional autoregressive Value at Risk by regression quantiles[J].Journal of Finance,2004,(22):367-381.
[4]Chen Hua,Kang Yixue.An empirical analysis of quantile regression based risk measurement[J].Management Engineering&Technology of Statistics,2012:37-45.
[5]刘新华,黄大山.中国股市风险CAViaR方法的稳定性分析及其时变建模[J].系统工程理论与实践,2005,(3):1-6.
[6]陈耀辉,朱盼盼.基于分位数回归模型的人民币汇率风险测度方法研究[J].长江大学学报,2015,(22):22-29.
[7]苟红军,陈迅,花拥军.基于GARCH-EVT-COPULA模型的外汇投资组合风险度量研究[J].管理工程学报,2015,(1):183-193.
[8]解其昌.分位数回归方法及其在金融市场风险价值预测中的应用[M].北京:中国农业科学技术出版社,2014.
[9]Artzner P,Delbaen F,Eber J M.Health D.Coherent measures of risk,mathematical finance[J].Mathematical Finance Theory Modeling Implementation,1999,9(3):203-228.
[10]Taylor J W.A quantile regression approach to estimating the distribution of multi-period returns[J].Journal of Derivatives,1999,(7):64-78.
[11]Koenker R,Bassett.Regression quantile[J].Econometrica,1978,46:33-50.
[12]Chen S.An integrated smoothed maximum score estimator for generalized censored quantile regression model[J].Journal of Econometrics,2010,155(1):90-98.
[13]Taylor J W.A quantile recession approach to estimating the distribution of multi-period returns[J].Journal of Derivatives,1999,(7):64-78.
[14]Kupiec P H.Thevniques for verifying the accuracy of risk measurement models[J].Journal of Derivatives,1995,(3):73-84.
[15]张琼,黄旭东,林雪勤.参数法、半参数法的动态VaR模型风险度量[J].统计与决策,2016,23(3):15-20.
[16]董然,郑慧,卢志伟.基于GARCH族模型的国债市场收益特征识别及风险测度[J].金融理论与实践,2016,(1):42-46.
[17]Meng Wang,Zhao Chen,Christina Dan Wang.Composite quantile regression for GARCH models using high-frequency data[J/OL].Econometrics and Statistics,2017:1-19.http://dx.doi.org/10.1016/j.ecosta.2016.11.004.
[18]张锡.基于VaR和CVaR的金融风险测度研究[J].科技与管理,2016,18(4):62-68.
[19]杨彬彬,徐庆娟.基于GARCH-VaR模型的汇率风险实证研究[J].广西师范学院学报(自然科学版),2017,(1).
[20]黄崇珍,曹奇.基于GARCH-VaR模型的开放式基金风险度量[J].统计与决策,2017,(1):152-155.
[21]Cathy W S.Chen Zona Wang,Songsak Sriboonchitta,Sangyeol Lee.Pair trading based on quantile forecasting of smooth transition GARCH models[J].North American Journal of Economics and Finance,2017,(39):38-55.
[22]许启发,徐金菊,蒋翠侠.基于神经网络分位数回归的多期CVaR风险测度[J].数理统计与管理,2017,36(4):715-730.
[23]赵丽丽,张波.基于改进ICA模型的高维波动率估计[J].数理统计与管理,2017,36(1):38-50.
[24]杨爱军,蒋学军,林金官,刘晓星.基于MCMC方法的金融贝叶斯半参数随机波动模型研究[J].数理统计与管理,2016,35(5):817-825.
[25]黄金波,李忠飞,姚海祥.条件VaR和条件CVaR的核估计及其实证分析[J].数理统计与管理,2016,35(2):232-242.
[26]张所地,范新英.基于面板分位数回归模型的收入、利率对房价的影响关系研究[J].数理统计与管理,2015,34(6):57-65.
[27]田茂茜,虞克明.银行业如何影响房地产业?Copula分位数回归及预测方法[J].数理统计与管理,2015,34(1):150-161.
[28]朱涛,卢健,江孝感.MGARCH模型对金融时间序列刻画能力的研究[J].数理统计与管理,2014,33(5):860-868.
[29]王理同,王晓叶,原俊青,周明华.基于半参数EGARCH模型的VaR和CVaR度量与实证研究[J].数理统计与管理,2014,33(4):655-659.
[2]Bollerslev T.Generalized autoregressive conditional heteroscedasticity[J].Journal of Econometrics,1986,(31):307-327.
[3]Engle R,Manganelli S.CAViaR:Conditional autoregressive Value at Risk by regression quantiles[J].Journal of Finance,2004,(22):367-381.
[4]Chen Hua,Kang Yixue.An empirical analysis of quantile regression based risk measurement[J].Management Engineering&Technology of Statistics,2012:37-45.
[5]刘新华,黄大山.中国股市风险CAViaR方法的稳定性分析及其时变建模[J].系统工程理论与实践,2005,(3):1-6.
[6]陈耀辉,朱盼盼.基于分位数回归模型的人民币汇率风险测度方法研究[J].长江大学学报,2015,(22):22-29.
[7]苟红军,陈迅,花拥军.基于GARCH-EVT-COPULA模型的外汇投资组合风险度量研究[J].管理工程学报,2015,(1):183-193.
[8]解其昌.分位数回归方法及其在金融市场风险价值预测中的应用[M].北京:中国农业科学技术出版社,2014.
[9]Artzner P,Delbaen F,Eber J M.Health D.Coherent measures of risk,mathematical finance[J].Mathematical Finance Theory Modeling Implementation,1999,9(3):203-228.
[10]Taylor J W.A quantile regression approach to estimating the distribution of multi-period returns[J].Journal of Derivatives,1999,(7):64-78.
[11]Koenker R,Bassett.Regression quantile[J].Econometrica,1978,46:33-50.
[12]Chen S.An integrated smoothed maximum score estimator for generalized censored quantile regression model[J].Journal of Econometrics,2010,155(1):90-98.
[13]Taylor J W.A quantile recession approach to estimating the distribution of multi-period returns[J].Journal of Derivatives,1999,(7):64-78.
[14]Kupiec P H.Thevniques for verifying the accuracy of risk measurement models[J].Journal of Derivatives,1995,(3):73-84.
[15]张琼,黄旭东,林雪勤.参数法、半参数法的动态VaR模型风险度量[J].统计与决策,2016,23(3):15-20.
[16]董然,郑慧,卢志伟.基于GARCH族模型的国债市场收益特征识别及风险测度[J].金融理论与实践,2016,(1):42-46.
[17]Meng Wang,Zhao Chen,Christina Dan Wang.Composite quantile regression for GARCH models using high-frequency data[J/OL].Econometrics and Statistics,2017:1-19.http://dx.doi.org/10.1016/j.ecosta.2016.11.004.
[18]张锡.基于VaR和CVaR的金融风险测度研究[J].科技与管理,2016,18(4):62-68.
[19]杨彬彬,徐庆娟.基于GARCH-VaR模型的汇率风险实证研究[J].广西师范学院学报(自然科学版),2017,(1).
[20]黄崇珍,曹奇.基于GARCH-VaR模型的开放式基金风险度量[J].统计与决策,2017,(1):152-155.
[21]Cathy W S.Chen Zona Wang,Songsak Sriboonchitta,Sangyeol Lee.Pair trading based on quantile forecasting of smooth transition GARCH models[J].North American Journal of Economics and Finance,2017,(39):38-55.
[22]许启发,徐金菊,蒋翠侠.基于神经网络分位数回归的多期CVaR风险测度[J].数理统计与管理,2017,36(4):715-730.
[23]赵丽丽,张波.基于改进ICA模型的高维波动率估计[J].数理统计与管理,2017,36(1):38-50.
[24]杨爱军,蒋学军,林金官,刘晓星.基于MCMC方法的金融贝叶斯半参数随机波动模型研究[J].数理统计与管理,2016,35(5):817-825.
[25]黄金波,李忠飞,姚海祥.条件VaR和条件CVaR的核估计及其实证分析[J].数理统计与管理,2016,35(2):232-242.
[26]张所地,范新英.基于面板分位数回归模型的收入、利率对房价的影响关系研究[J].数理统计与管理,2015,34(6):57-65.
[27]田茂茜,虞克明.银行业如何影响房地产业?Copula分位数回归及预测方法[J].数理统计与管理,2015,34(1):150-161.
[28]朱涛,卢健,江孝感.MGARCH模型对金融时间序列刻画能力的研究[J].数理统计与管理,2014,33(5):860-868.
[29]王理同,王晓叶,原俊青,周明华.基于半参数EGARCH模型的VaR和CVaR度量与实证研究[J].数理统计与管理,2014,33(4):655-659.