中国上市银行增量长期条件风险价值估算研究  

Research on the Estimation of Incremental Long-term Conditional Value at Risk of Listed Banks in China

作　　者：王周伟[1] 魏鹏飞 WANG Zhouwei;WEI Pengfei(School of Finance and Business,Shanghai Normal University,Shanghai 200234,China;School of Finance,Central University of Finance and Economics,Beijing 100098,China)

机构地区：[1]上海师范大学商学院,上海200234 [2]中央财经大学金融学院,北京100098

出　　处：《统计与信息论坛》2023年第6期81-101,共21页Journal of Statistics and Information

基　　金：国家自然科学基金面上项目“结构变化中银行系统性金融风险的多维多重传染研究”(71973098)。

摘　　要：防控系统性金融风险的前提在于风险趋势的动态监测,该趋势决定于长期条件风险价值溢出传染。为有效测度中国上市银行的增量长期条件风险价值指标并评价其系统重要性和系统脆弱性,选用GARCH-MIDAS模型估算其长期波动率、短期波动率及总波动率,使用DCC-MIDAS模型测度长期相关系数和总相关系数。将所测算出的长期波动率与长期相关系数相结合,计算出增量长期条件风险价值、长期风险溢出指数和长期风险吸收指数。对比分析增量长期条件风险价值和以往研究中所使用的增量条件风险价值之间的差异和联系。考察剔除短期扰动下的风险价值溢出的长期趋势特征,并根据构建的长期风险溢出指数和长期风险吸收指数,稳健地识别出剔除短期扰动下的系统重要性银行和系统脆弱性银行。研究表明:其一,金融资产波动可以显著分解为长期波动和短期波动,长期波动决定着银行收益率波动趋势;金融资产间的相关系数则主要由其长期相关系数决定,并受到短期因素的干扰;其二,长期风险价值确实能够有效测度银行风险价值变化趋势,且增量长期条件风险价值决定增量条件风险价值的变化趋势,是银行风险溢出传染的核心要素;其三,通过增量长期条件风险价值构造的长期风险溢出指数LSRE与长期风险吸附指数LSRR可以稳健地识别系统重要性机构和系统脆弱性机构,从而为审慎监管精准施策提供指导。The key to preventing systemic financial risk is to dynamically monitor its risk trend,which depends on the long-term conditional value at risk spillover.Therefore,in order to effectively measure the incremental long-term conditional value at risk of listed banks in China and evaluate their systematic importance and vulnerability,the GARCH-MIDAS model is selected to measure their long-term volatility,short-term volatility and total volatility.And the DCC-MIDAS model is used to measure long-term correlation coefficient and total correlation coefficient.According to the long-term volatility obtained by variance decomposition,the long-term value at risk and total value at risk of listed banks can be calculated.In fact,the conditional value at risk,which is used in the other previous research paper,is the total conditional value at risk,which not only includes low-frequency factors such as macroeconomic indicators and enterprise operating conditions,but also includes the interference of random shocks.Then,the long-term conditional value at risk can be further calculated from long-term value at risk and long-term correlation coefficient.Similarly,the total conditional value at risk can be further calculated from total value at risk and total correlation coefficient.Therefore,the total conditional value at risk is not only affected by low-frequency factors such as macroeconomic index and enterprise operation index,but also disturbed by random shocks.However,the long-term conditional value at risk,would not be disturbed by random shocks,and it’s the advantage of the index.The characteristics of long-term conditional value at risk are analyzed through data and images,and fortunately,the advantage of long-term index is verified by comparing and analyzing the difference and relationship between long-term incremental conditional value at risk and incremental conditional value at risk.In addition,systemic important financial institutions are characterized by large scale,high degree of association with other institutions an

关 键 词：GARCH-MIDAS模型 长期波动 DCC-MIDAS模型 长期相关 增量条件风险价值 

分 类 号：F832.59[经济管理—金融学] O212.2[理学—概率论与数理统计]