统计与数据科学学院知识库

防范系统性金融风险是我国经济工作的重中之重，党的二十大报告明确要求，全党必须提高防范化解重大风险能力，守住不发生系统性风险底线。因此厘清我国主要金融行业间风险传染机制、立足不同视角发现各类因素对系统风险的影响对维护金融体系稳定具有重要意义。

本文为分析我国金融部门及相关行业风险传染机制和系统风险溢出的影响因素，首先构建时变参数向量自回归（TVP-VAR）模型，一方面解决了滚动窗口大小对溢出效应度量的影响，避免遗漏重要信息；另一方面，解决了传统的VAR模型无法捕捉经济结构的时变特征的问题，使研究更贴合实际经济情况。本文在该模型基础上获得基于广义方差分解的溢出指数，研究基于尾部风险溢出的跨市场风险传染。而后，通过条件在险价值（CoVaR）模型测度各金融部门对金融系统的风险溢出效应，基于混合面板数据模型，研究管理者过度自信及企业层面因素对系统风险的溢出效应。研究结果表明：第一，所选行业尾部风险总溢出指数达59%，表明极端损失具有较强的跨市场传染性。第二，黄金和地产行业是风险的主要接收方，银行和证券部门是风险的主要传播方。地产与银行、证券联系紧密，在风险波动剧烈时能较好吸收这两个金融部门传播的风险，黄金市场则在波动稳定时具有更好的风险吸收能力。第三，证券和保险公司系统风险稳定性更弱，更易向金融系统溢出风险。在危机期间，过度自信的企业管理者会加剧风险溢出。第四，证券公司或保险公司的过度自信管理者对系统性风险溢出的贡献，显著大于银行的过度自信管理者对系统性风险溢出的贡献。

综上所述，在危机期间金融部门及相关行业的跨市场尾部风险关联性会快速提高，政府部门要加强监管，并在制定政策时充分考虑对各部门间风险传递的放大效应。在防范系统性风险溢出时，需要更多地关注过度自信管理者对系统风险溢出的效应，并且注重不同行业对应系统性金融风险溢出效应的差异性。

Preventing systemic financial risks is the top priority of China's economic work. The report to the Party's 20 National Congress clearly states that the whole Party must improve its ability to forestall and defuse major risks and ensure that systemic risks do not occur. Therefore, it is of great significance to clarify the risk contagion mechanism among major financial industries and discover the impact of various factors on systemic risks from different perspectives to maintain the stability of the financial system.

In order to analyze the influencing factors of risk contagion mechanism and systemic risk spillover in Chinese financial sector and related industries, a time-varying parameter vector autoregression (TVP-VAR) model is firstly constructed. On the one hand, the influence of rolling window size on spillover effect measurement is solved to avoid missing important information. On the other hand, it solves the problem that the traditional VAR model cannot capture the time-varying characteristics of economic structure, so that the research is more in line with the actual economic situation. Based on this model, this paper obtains the spillover index based on the generalized variance decomposition and studies the cross-market risk contagion based on the tail risk spillover. Then, the conditional value at risk (CoVaR) model is used to measure the risk spillover effects of various financial sectors on the financial system. Based on the mixed panel data model, the spillover effects of managers' overconfidence and enterprise-level factors on the system risks are studied. The results show that: firstly, the total tail risk spillover index of selected industries is 59%, indicating that extreme losses have strong cross-market infectivity. Secondly, the gold and real estate industries are the main recipients of risks, while the banking and securities sectors are the main transmitters of risks. Real estate is closely connected with banks and securities, so it can better absorb the risks spread by these two financial sectors when the risks are highly volatile, while the gold market has better risk absorption capacity when the risks are stable. Thirdly, the stability of securities and insurance companies is weaker, which is more likely to spill risks to the financial system. During a crisis, overconfident managers exacerbate risk spillovers; Fourthly, The contribution of overconfident managers in securities companies or insurance companies to systemic risk spillover is significantly greater than that of overconfident managers in banks.

To sum up, the cross-market tail risk correlation of the financial sector and related industries will be enhanced rapidly during the crisis. The government departments should strengthen the supervision and take the amplification effect of risk transmission among various departments into full consideration when formulating policies. When preventing systemic risk spillover, we need to pay more attention to the spillover effect of overconfident managers on systemic risk, and pay attention to the difference of the spillover effect of systemic financial risk in different industries.