中国金融部门间系统性风险溢出的监测预警研究——基于下行和上行ΔCoES指标的实现与优化
|
摘要
为了对我国金融部门间的系统性风险溢出进行实时监测和有效预警,本文基于Adrian and Brunnermeier (2016)的Co ES指标构想,在左尾视角的基础上进一步引入右尾视角,构建下行和上行ΔCo ES分别作为系统性风险的同期度量指标和前瞻预警指标,并提出了更为有效合理且同时适用于下行和上行ΔCo ES的计算方法。本文一方面采用下行和上行ΔCo ES对我国银行、证券、保险三个金融部门间的系统性风险溢出进行监测预警研究,另一方面还基于我国的经验数据检验上行和下行ΔCo ES的性质。研究结果显示,我国金融部门间具有显著的系统性风险溢出效应,且三个部门间的风险溢出存在非对称性,银行部门是系统性风险的主要发送者,证券部门是系统性风险的主要接收者;三个部门两两间的风险溢出水平表现出明显的协同性和周期性,且上行的风险溢出水平高于下行。同时,基于我国的经验数据发现,上行ΔCo ES对下行ΔCo ES具有显著的先导性、前瞻性,上行ΔCo ES可以作为系统性风险的前瞻预警指标。此外,下行ΔCo ES能够引领ΔCo VaR和基于MES估计方法计算的短期ΔCo ES指标,表明本文构建的下行ΔCo ES实时性更强,更适合作为系统性风险的实时监测指标。
China's comprehensive financial reforms were accompanied by the rapid integration of financial institutions including banks,securities,and insurance companies and the continual introduction of cross-businesses and cross-products such that mixed financial operations have become a trend. This not only exacerbates the cross-sector and cross-market spread of risk,but increases the possible cross-contagion of financial risk. When a crisis occurs in one industry,the cross-contagion and resonance of risk may further induce systemic financial risk,threatening China's financial security. At present,preventing and reducing systemic risk and safeguarding financial security are the focus of the whole society. However,it is not easy to prevent systemic risk,and the accurate measurement of risk is a prerequisite for scientific prevention. Only through real-time monitoring and effective forewarning can we hold the bottom line for systemic risk and national financial security. Research on monitoring and early warning signs of systemic risk spillover among China's banking,securities,and insurance sectors will not only contribute to the prevention of cross-sector risk,but also help to prevent systemic financial risk and defend the security of the entire financial system.To carry out real-time monitoring and effective early forewarning of systemic risk spillover in China's financial sector based on the Co ES indicator( Adrian and Brunnermeier,2016) and the traditional left-tail perspective,we also include the right-tail perspective. We subsequently build upper ΔCo ES and lower ΔCo ES measures as real-time monitoring and forward-looking warning indicators,respectively,of systemic risk under a unified Co ES framework. We also take advantage of Co VaR and LRMES to propose a more effective and accurate calculation method for these indicators. We select the banking,securities,and insurance industry index in the Shenwan second-class industry index to proxy for these three financial sectors. The modified Co ES indicators are used to monitor and predict the systemic risk spillover among China's financial sectors. Based on the data,this paper tests the forward-looking and early warning characteristics of the upper ΔCo ES and the real-time characteristics of the lower ΔCo ES.The results suggest that there are significant and asymmetric systemic risk spillover effects among the three financial sectors. While the banking sector is the main sender of systemic risk,the securities sector is the main recipient. Risk spillover among China's financial sectors shows significant co-movement and cyclicality,and the upside risk spillover is higher than that of the downside. At the same time,the upper ΔCo ES is significantly ahead of the lower ΔCo ES,and can therefore be used as a forward-looking warning indicator of systemic risk.In addition,the lower ΔCo ES can lead the ΔCo VaR,and the short-term ΔCo ES calculated based on the MES estimation method indicates that the lower ΔCo ES is more suitable as a real-time indicator of systemic risk.There are two main policy implications of these results. In the cross-sectional dimension,according to the role and status of the different financial sectors in systemic risk transmission,the supervisory authority should select regulatory objectives and policy tools to carry out differentiated monitoring and prevention. As the banking sector is the main sender of systemic risk,the focus of China's systemic risk prevention should be the banking sector. Reducing risk spillover from the banking sector is the key to preventing and defusing systemic financial risk and maintaining the security and stability of the financial system. As the securities sector is the main recipient of systemic risk,we should pay attention to improving this sector's ability to resist risk and reducing its systemic vulnerability. In the time dimension,the supervisory authority should not only monitor the real-time dynamics of the systemic risk spillover,but also build effective early-warning indicators of systemic risk and improve the mechanisms of risk monitoring,early warning,and early intervention. The supervisory authority can use the lower and upper ΔCo ES constructed in this paper as real-time monitoring and forward-looking warning indicators to further improve the systemic risk monitoring and early warning system in China.
China's comprehensive financial reforms were accompanied by the rapid integration of financial institutions including banks,securities,and insurance companies and the continual introduction of cross-businesses and cross-products such that mixed financial operations have become a trend. This not only exacerbates the cross-sector and cross-market spread of risk,but increases the possible cross-contagion of financial risk. When a crisis occurs in one industry,the cross-contagion and resonance of risk may further induce systemic financial risk,threatening China's financial security. At present,preventing and reducing systemic risk and safeguarding financial security are the focus of the whole society. However,it is not easy to prevent systemic risk,and the accurate measurement of risk is a prerequisite for scientific prevention. Only through real-time monitoring and effective forewarning can we hold the bottom line for systemic risk and national financial security. Research on monitoring and early warning signs of systemic risk spillover among China's banking,securities,and insurance sectors will not only contribute to the prevention of cross-sector risk,but also help to prevent systemic financial risk and defend the security of the entire financial system.To carry out real-time monitoring and effective early forewarning of systemic risk spillover in China's financial sector based on the Co ES indicator( Adrian and Brunnermeier,2016) and the traditional left-tail perspective,we also include the right-tail perspective. We subsequently build upper ΔCo ES and lower ΔCo ES measures as real-time monitoring and forward-looking warning indicators,respectively,of systemic risk under a unified Co ES framework. We also take advantage of Co VaR and LRMES to propose a more effective and accurate calculation method for these indicators. We select the banking,securities,and insurance industry index in the Shenwan second-class industry index to proxy for these three financial sectors. The modified Co ES indicators are used to monitor and predict the systemic risk spillover among China's financial sectors. Based on the data,this paper tests the forward-looking and early warning characteristics of the upper ΔCo ES and the real-time characteristics of the lower ΔCo ES.The results suggest that there are significant and asymmetric systemic risk spillover effects among the three financial sectors. While the banking sector is the main sender of systemic risk,the securities sector is the main recipient. Risk spillover among China's financial sectors shows significant co-movement and cyclicality,and the upside risk spillover is higher than that of the downside. At the same time,the upper ΔCo ES is significantly ahead of the lower ΔCo ES,and can therefore be used as a forward-looking warning indicator of systemic risk.In addition,the lower ΔCo ES can lead the ΔCo VaR,and the short-term ΔCo ES calculated based on the MES estimation method indicates that the lower ΔCo ES is more suitable as a real-time indicator of systemic risk.There are two main policy implications of these results. In the cross-sectional dimension,according to the role and status of the different financial sectors in systemic risk transmission,the supervisory authority should select regulatory objectives and policy tools to carry out differentiated monitoring and prevention. As the banking sector is the main sender of systemic risk,the focus of China's systemic risk prevention should be the banking sector. Reducing risk spillover from the banking sector is the key to preventing and defusing systemic financial risk and maintaining the security and stability of the financial system. As the securities sector is the main recipient of systemic risk,we should pay attention to improving this sector's ability to resist risk and reducing its systemic vulnerability. In the time dimension,the supervisory authority should not only monitor the real-time dynamics of the systemic risk spillover,but also build effective early-warning indicators of systemic risk and improve the mechanisms of risk monitoring,early warning,and early intervention. The supervisory authority can use the lower and upper ΔCo ES constructed in this paper as real-time monitoring and forward-looking warning indicators to further improve the systemic risk monitoring and early warning system in China.
引文
[1]白雪梅和石大龙,2014,《中国金融体系的系统性风险度量》,《国际金融研究》第6期,第75~85页。
[2]卜林和李政,2015,《我国上市金融机构系统性风险溢出研究---基于CoVaR和MES的比较分析》,《当代财经》第6期,第55~65页。
[3]陈建青、王擎和许韶辉,2015,《金融行业间的系统性金融风险溢出效应研究》,《数量经济技术经济研究》第9期,第89~100页。
[4]范小云、王道平和方意,2011,《我国金融机构的系统性风险贡献测度与监管---基于边际风险贡献与杠杆率的研究》,《南开经济研究》第4期,第3~20页。
[5]方意、赵胜民和王道平,2012,《我国金融机构系统性风险测度---基于DCC-GARCH模型的研究》,《金融监管研究》第11期,第26~42页。
[6]高国华和潘英丽,2011,《银行系统性风险度量---基于动态CoVaR方法的分析》,《上海交通大学学报》第12期,第1753~1759页。
[7]李志辉和樊莉,2011,《中国商业银行系统性风险溢价实证研究》,《当代经济科学》第6期,第13~20页。
[8]梁琪、李政和郝项超,2013,《我国系统重要性金融机构的识别与监管---基于系统性风险指数SRISK方法的分析》,《金融研究》第9期,第56~70页。
[9]刘吕科、张定胜和邹恒甫,2012,《金融系统性风险衡量研究最新进展述评》,《金融研究》第11期,第31~43页。
[10]肖璞、刘轶和杨苏梅,2012,《相互关联性、风险溢出与系统重要性银行识别》,《金融研究》第12期,第96~106页。
[11]严伟祥、张维和牛华伟,2017,《金融风险动态相关与风险溢出异质性研究》,《财贸经济》第10期,第67~81页。
[12]杨子晖、陈雨恬和谢锐楷,2018,《我国金融机构系统性金融风险度量与跨部门风险溢出效应研究》,《金融研究》第10期,第19~37页。
[13]张冰洁、汪寿阳、魏云捷和赵雪婷,2018,《基于CoES模型的我国金融系统性风险度量》,《系统工程理论与实践》第3期,第565~575页。
[14]周天芸、杨子晖和余洁宜,2014,《机构关联、风险溢出与中国金融系统性风险》,《统计研究》第11期,第43~49页。
[15]Acharya,V.V.,L.H.Pedersen,T.Philippon,and M.Richardson.2017.“Measuring Systemic Risk”Review of Financial Studies,30(1):2~47.
[16]Acharya,V.V.,R.F.Engle,and M.Richardson.2012.“Capital Shortfall:A New Approach to Ranking and Regulating Systemic Risks”American Economic Review,102(3):59~64.
[17]Adrian,T.,and M.K.Brunnermeier.2016.“CoVaR”American Economic Review,106(7):1705~1741.
[18]Benoit,S.,J.Colliard,C.Hurlin,and C.Pérignon.2017.“Where the Risks Lie:A Survey On Systemic Risk”Review of Finance,21(1):109~152.
[19]Bhattacharya,S.,C.A.Goodhart,D.P.Tsomocos,and A.P.Vardoulakis.2015.“A Reconsideration of Minsky's Financial Instability Hypothesis”Journal of Money,Credit and Banking,47(5):931~973.
[20]Brownlees,C.,and R.F.Engle.2017.“SRISK:A Conditional Capital Shortfall Measure of Systemic Risk”Review of Financial Studies,30(1):48~79.
[21]Brunnermeier,M.K.,and Y.Sannikov.2014.“A Macroeconomic Model with a Financial Sector”American Economic Review,104(2):379~421.
[22]Danielsson,J.,M.Valenzuela,and I.Zer,2016.“Learning From History:Volatility and Financial Crises”SRCDiscussion Paper,No.57.
[23]Gauthier,C.L.,C.Graham,and L.Ying.2004.“Financial Conditions Indexes for Canada”Bank of Canada Working Paper,No.22.
[24]Girardi,G.,and A.Tolga Ergün.2013.“Systemic Risk Measurement:Multivariate GARCH Estimation of CoVaR”Journal of Banking&Finance,37(8):3169~3180.
1当然除了CoVaR和MES等基于金融市场数据的尾部依赖方法,系统未定权益分析SCCA等联合违约方法、基于Granger因果检验和广义方差分解的网络分析方法、基于金融机构业务数据的金融网络模型等也都得到我国学者的广泛关注与应用。鉴于本文主要沿着尾部依赖方法这一条线进行拓展性研究,故对其他方法的具体细节以及相关文献不做详细介绍,感兴趣的读者可参阅刘吕科等(2012)等相关研究。
2系统性风险同期度量方法并不影响截面维度上金融机构系统性风险水平的横向比较,即不会影响到系统重要性金融机构的识别与监管。
1目前,AB-CoES对CoVaR方法的两个改进之处已得到国内外部分学者的认同与应用。Girardi and Tolga Ergün(2013)采用了第二个改进之处,将金融困境的定义从机构的收益率等于其风险价值VaR变成小于等于其VaR;张冰洁等(2018)采纳了第一个改进之处,将风险度量指标由VaR改成了ES,采用CoES这一指标名称,提出了自己的CoES指标。但遗憾的是,Girardi and Tolga Ergün(2013)仍采用VaR作为风险度量指标,张冰洁等(2018)中CoES的条件事件仍然是Xi=VaRi p。
1ΔCoESLi b、ΔCoESLb i、ΔCoESLs b、ΔCoESLb s、ΔCoESLi s和ΔCoESLs i等6个指标给出了三个部门两两间下行的系统性风险溢出水平。
2比如,银行对证券的风险溢出从2009年3月份的0.1474降至2010年4月的0.0346,证券对银行的风险溢出也由0.1174降至0.0337,下降幅度均超过70%。
32010年6月三个部门两两间的风险溢出水平上升至0.1135、0.0910、0.1199、0.0884、0.1096和0.1214。
1比如,ΔCoESUt与ΔCoESLt-6的相关系数,实际上也是ΔCoESLt与ΔCoESUt+6的相关系数。
1限于篇幅,上行与下行ΔCoES在滞后1~6期下的Granger因果检验结果在正文中未列出,感兴趣的读者可向作者索取。
1限于篇幅,下行ΔCoES对上行ΔCoES的脉冲响应结果在正文中未列出,感兴趣的读者可向作者索取。
1限于篇幅,上行ΔCoES对下行ΔCoES的预测能力评价结果在正文中未列出,感兴趣的读者可向作者索取。
2本文将基于Acharya et al.(2017)中MES估计方法计算的下行系统性风险度量指标ΔCoES称为短期ΔCoES,CoESLj i q为在i的日收益率小于其q%分位数条件下,j的日收益率均值。
1限于篇幅,下行ΔCoES与短期ΔCoES的实时性检验结果在正文中未列出,感兴趣的读者可向作者索取。
[2]卜林和李政,2015,《我国上市金融机构系统性风险溢出研究---基于CoVaR和MES的比较分析》,《当代财经》第6期,第55~65页。
[3]陈建青、王擎和许韶辉,2015,《金融行业间的系统性金融风险溢出效应研究》,《数量经济技术经济研究》第9期,第89~100页。
[4]范小云、王道平和方意,2011,《我国金融机构的系统性风险贡献测度与监管---基于边际风险贡献与杠杆率的研究》,《南开经济研究》第4期,第3~20页。
[5]方意、赵胜民和王道平,2012,《我国金融机构系统性风险测度---基于DCC-GARCH模型的研究》,《金融监管研究》第11期,第26~42页。
[6]高国华和潘英丽,2011,《银行系统性风险度量---基于动态CoVaR方法的分析》,《上海交通大学学报》第12期,第1753~1759页。
[7]李志辉和樊莉,2011,《中国商业银行系统性风险溢价实证研究》,《当代经济科学》第6期,第13~20页。
[8]梁琪、李政和郝项超,2013,《我国系统重要性金融机构的识别与监管---基于系统性风险指数SRISK方法的分析》,《金融研究》第9期,第56~70页。
[9]刘吕科、张定胜和邹恒甫,2012,《金融系统性风险衡量研究最新进展述评》,《金融研究》第11期,第31~43页。
[10]肖璞、刘轶和杨苏梅,2012,《相互关联性、风险溢出与系统重要性银行识别》,《金融研究》第12期,第96~106页。
[11]严伟祥、张维和牛华伟,2017,《金融风险动态相关与风险溢出异质性研究》,《财贸经济》第10期,第67~81页。
[12]杨子晖、陈雨恬和谢锐楷,2018,《我国金融机构系统性金融风险度量与跨部门风险溢出效应研究》,《金融研究》第10期,第19~37页。
[13]张冰洁、汪寿阳、魏云捷和赵雪婷,2018,《基于CoES模型的我国金融系统性风险度量》,《系统工程理论与实践》第3期,第565~575页。
[14]周天芸、杨子晖和余洁宜,2014,《机构关联、风险溢出与中国金融系统性风险》,《统计研究》第11期,第43~49页。
[15]Acharya,V.V.,L.H.Pedersen,T.Philippon,and M.Richardson.2017.“Measuring Systemic Risk”Review of Financial Studies,30(1):2~47.
[16]Acharya,V.V.,R.F.Engle,and M.Richardson.2012.“Capital Shortfall:A New Approach to Ranking and Regulating Systemic Risks”American Economic Review,102(3):59~64.
[17]Adrian,T.,and M.K.Brunnermeier.2016.“CoVaR”American Economic Review,106(7):1705~1741.
[18]Benoit,S.,J.Colliard,C.Hurlin,and C.Pérignon.2017.“Where the Risks Lie:A Survey On Systemic Risk”Review of Finance,21(1):109~152.
[19]Bhattacharya,S.,C.A.Goodhart,D.P.Tsomocos,and A.P.Vardoulakis.2015.“A Reconsideration of Minsky's Financial Instability Hypothesis”Journal of Money,Credit and Banking,47(5):931~973.
[20]Brownlees,C.,and R.F.Engle.2017.“SRISK:A Conditional Capital Shortfall Measure of Systemic Risk”Review of Financial Studies,30(1):48~79.
[21]Brunnermeier,M.K.,and Y.Sannikov.2014.“A Macroeconomic Model with a Financial Sector”American Economic Review,104(2):379~421.
[22]Danielsson,J.,M.Valenzuela,and I.Zer,2016.“Learning From History:Volatility and Financial Crises”SRCDiscussion Paper,No.57.
[23]Gauthier,C.L.,C.Graham,and L.Ying.2004.“Financial Conditions Indexes for Canada”Bank of Canada Working Paper,No.22.
[24]Girardi,G.,and A.Tolga Ergün.2013.“Systemic Risk Measurement:Multivariate GARCH Estimation of CoVaR”Journal of Banking&Finance,37(8):3169~3180.
1当然除了CoVaR和MES等基于金融市场数据的尾部依赖方法,系统未定权益分析SCCA等联合违约方法、基于Granger因果检验和广义方差分解的网络分析方法、基于金融机构业务数据的金融网络模型等也都得到我国学者的广泛关注与应用。鉴于本文主要沿着尾部依赖方法这一条线进行拓展性研究,故对其他方法的具体细节以及相关文献不做详细介绍,感兴趣的读者可参阅刘吕科等(2012)等相关研究。
2系统性风险同期度量方法并不影响截面维度上金融机构系统性风险水平的横向比较,即不会影响到系统重要性金融机构的识别与监管。
1目前,AB-CoES对CoVaR方法的两个改进之处已得到国内外部分学者的认同与应用。Girardi and Tolga Ergün(2013)采用了第二个改进之处,将金融困境的定义从机构的收益率等于其风险价值VaR变成小于等于其VaR;张冰洁等(2018)采纳了第一个改进之处,将风险度量指标由VaR改成了ES,采用CoES这一指标名称,提出了自己的CoES指标。但遗憾的是,Girardi and Tolga Ergün(2013)仍采用VaR作为风险度量指标,张冰洁等(2018)中CoES的条件事件仍然是Xi=VaRi p。
1ΔCoESLi b、ΔCoESLb i、ΔCoESLs b、ΔCoESLb s、ΔCoESLi s和ΔCoESLs i等6个指标给出了三个部门两两间下行的系统性风险溢出水平。
2比如,银行对证券的风险溢出从2009年3月份的0.1474降至2010年4月的0.0346,证券对银行的风险溢出也由0.1174降至0.0337,下降幅度均超过70%。
32010年6月三个部门两两间的风险溢出水平上升至0.1135、0.0910、0.1199、0.0884、0.1096和0.1214。
1比如,ΔCoESUt与ΔCoESLt-6的相关系数,实际上也是ΔCoESLt与ΔCoESUt+6的相关系数。
1限于篇幅,上行与下行ΔCoES在滞后1~6期下的Granger因果检验结果在正文中未列出,感兴趣的读者可向作者索取。
1限于篇幅,下行ΔCoES对上行ΔCoES的脉冲响应结果在正文中未列出,感兴趣的读者可向作者索取。
1限于篇幅,上行ΔCoES对下行ΔCoES的预测能力评价结果在正文中未列出,感兴趣的读者可向作者索取。
2本文将基于Acharya et al.(2017)中MES估计方法计算的下行系统性风险度量指标ΔCoES称为短期ΔCoES,CoESLj i q为在i的日收益率小于其q%分位数条件下,j的日收益率均值。
1限于篇幅,下行ΔCoES与短期ΔCoES的实时性检验结果在正文中未列出,感兴趣的读者可向作者索取。