主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2025, Vol. 33 ›› Issue (6): 14-26.doi: 10.16381/j.cnki.issn1003-207x.2022.1453cstr: 32146.14/j.cnki.issn1003-207x.2022.1453

• • 上一篇    下一篇

中国金融机构系统性风险回测与关联研究

欧阳资生1, 周学伟2,3()   

  1. 1.湖南师范大学商学院,湖南 长沙 410081
    2.上海财经大学金融学院,上海 200433
    3.上海国际金融与经济研究院,上海 200433
  • 收稿日期:2022-07-04 修回日期:2023-02-04 出版日期:2025-06-25 发布日期:2025-07-04
  • 通讯作者: 周学伟 E-mail:zhouxuewei@mail.shufe.edu.cn
  • 基金资助:
    国家社会科学基金项目(23BTJ043);湖南省研究生科研创新项目(CX20220466)

Systemic Risk Backtesting and Connectedness of Chinese Financial Institutions: Evidence from MES and ΔCoVaR

Zisheng Ouyang1, Xuewei Zhou2,3()   

  1. 1.Business School,Hunan Normal University,Changsha 410081,China
    2.School of Finance,Shanghai University of Finance and Economics,Shanghai 200433,China
    3.Shanghai Institute of International Finance and Economics,Shanghai 200433,China
  • Received:2022-07-04 Revised:2023-02-04 Online:2025-06-25 Published:2025-07-04
  • Contact: Xuewei Zhou E-mail:zhouxuewei@mail.shufe.edu.cn

摘要:

有效测度金融机构的风险水平,是防范化解系统性风险的关键。本文运用无条件覆盖测试(unconditional coverage test)方法,回测检验我国35家上市金融机构的系统性风险,并借助分位数相干性(quantile coherency)方法,构建分位数频域关联网络,以此探讨金融机构系统性风险关联的频域机制。研究发现:(1)在金融市场危机时期,MES、CoVaR和ΔCoVaR等常用的系统性风险度量指标,未能有效测度我国金融机构的系统性风险。(2)相较于短期关联层和中期关联层,我国金融机构在长期关联层存在明显的极端风险联动效应。(3)在短期、中期和长期三个关联层,宁波银行、兴业证券和东方能源等中小金融机构均具有较高的关联重要性,相关部门要注意防范我国中小金融机构的“太关联而不能倒”风险。

关键词: 系统性风险, 回测检验, 金融机构

Abstract:

The premise of preventing and defusing systemic risk is to accurately measure systemic risk. Through literature review, it is known that most of the existing research focuses on the measurement of systemic risk, and there is little literature evaluating the validity and accuracy of systemic risk. In addition, it is found that most of the literature mainly discusses the risk connectedness of financial institutions in the time domain, ignoring the connecting mechanism in the frequency domain. As a result, it focusy on two issues: (1) Backtesting the systemic risk of Chinese financial institutions to examine the effectiveness of MES and ΔCoVaR. (2) Based on the backtesting results of systemic risks, the risk connectedness among financial institutions is investigated by the quantile connectedness network in frequency.In response to the above problems, the following work is done: First, the MES and ΔCoVaR of 35 financial institutions in China are backtested by the unconditional coverage test, aiming to evaluate the effectiveness of MES and ΔCoVaR. Second, the early warning system is calulated to analyze the risk evolution of the financial system. Third, the quantile connectedness networks in the frequency domain through the quantile coherence method is proposed to explore the risk connectedness among financial institutions. Finally, the connectedness importance of financial institutions on specific frequency bands is measured.It is found that (1) During the period of financial market crisis, commonly used systemic risk indicators such as MES, CoVaR, and ΔCoVaR fail to measure the systemic risk of financial institutions. (2) Compared with the short-term and medium-term connectedness layer, Chinese financial institutions have obvious extreme risk linkage effects in the long-term connectedness layer. (3) In the short-, medium-, and long-term, small- and medium-sized financial institutions such as Bank of Ningbo, Industrial Securities, and Dongfang Energy have connectedness importance. Therefore, the regulatory authorities should pay attention to preventing the risk of "too connected to fail" of small- and medium-sized financial institutions.

Key words: systemic risk, backtesting, financial institutions

中图分类号: