统计与数据科学学院知识库

在防范化解系统性金融风险的底线思维下，尤其结合疫情冲击下我国宏观经济增速显著下滑的背景，需要增强预警意识，一方面要及时洞察金融市场的实时变化状况，以预警出现可能出现的金融风险，另一方面要深入分析金融市场状况对实体经济的影响，防范金融风险传染至宏观经济。

本文首先利用混频动态因子模型测度出月度GDP，然后结合金融市场的多维数据构建时变参数因子增广的向量自回归（TVP-FAVAR）模型，通过各因子对通货膨胀的冲击效应确定出其在金融状况指数（FCI）中的动态权重，以此得到我国月度动态FCI。在此基础上，利用谱分析测度FCI的预测能力对其做稳健性检验。最后，构建分位数回归模型和马尔科夫区制转换向量自回归（MS-VAR）模型研究在不同视角金融市场对实体经济的非对称性冲击。

研究结果表明：第一，本文构建的动态FCI能有效地跟踪各潜在因子对中国金融状况的差异性影响；第二，动态FCI较常系数FCI能更有效地捕捉到我国金融市场的变化特征，对实体经济的预测能力更强，可为中国货币政策的制定提供科学依据；第三，动态FCI和GDP均存在39个月耦合震荡周期，包含了实体经济的前瞻性信息，具备优良的预测实体经济的能力，可作为实体经济的先行指标；第四，在不同的经济发展水平下，金融市场对实体经济的影响存在显著差异性。经济下行时期金融市场对实体经济的影响力大于经济上行时期金融市场对实体经济的影响力；第五，经济系统维持在某个状态的可能性较高，在不同状态之间跳跃的能力比较弱，区制表现出来稳定的特征。而且随着经济金融状态的不同，金融市场对实体经济的冲击存在差异性。

综上所述，在不同的状态金融市场对实体经济的影响是有显著差异的；相比常系数FCI，动态FCI更能有效地捕捉到我国金融市场的变化特征，对实体经济的预测能力更强，可作为实体经济的先行指标。本文的研究为政府部门实时监测实体经济和金融市场的状况提供了政策制定依据，进而制定政策来防范金融市场的风险以便更好地为实体经济服务提供了科学有效的建议。

Under the bottom-line thinking of preventing and resolving systemic financial risks, especially combined with the background of the significant decline in China's macroeconomic growth rate under the impact of the epidemic, it is necessary to enhance the awareness of early warning, on the one hand, it is necessary to timely understand the real-time changes in the financial market to warn of possible financial risks, on the other hand, it is necessary to deeply analyze the impact of financial market conditions on the real economy to prevent financial risks from contagious to the macroeconomy.

In this thesis, the monthly GDP is measured by the mixed dynamic factor model, and then the vector autoregressive (TVP-FAVAR) model with time-varying parameter factor expansion is constructed by combining the multidimensional data of the financial market, and the dynamic weight in the financial condition index (FCI) is determined by the impact effect of each factor on inflation, so as to obtain the monthly dynamic FCI in China. On this basis, the robustness of the FCI is tested by the predictive ability of the spectral analysis measure. Finally, a quantile regression model and a Markov zone system transformation vector autoregressive (MS-VAR) model are constructed to study the asymmetric impact of financial markets on the real economy from different perspectives.

The results show that: first, the dynamic FCI constructed in this thesis can effectively track the differential impact of potential factors on China's financial conditions. Second, the dynamic FCI can more effectively capture the changing characteristics of China's financial market than the constant coefficient FCI, and has a stronger ability to predict the real economy, which can provide a scientific basis for the formulation of China's monetary policy. Third, both dynamic FCI and GDP have a 39-month coupling shock cycle, which contains forward-looking information of the real economy, has excellent ability to predict the real economy, and can be used as a leading indicator of the real economy; Fourth, under different levels of economic development, there are significant differences in the impact of financial markets on the real economy. The influence of the financial market on the real economy during the economic downturn period is greater than the influence of the financial market on the real economy during the economic upswing; Fifth, the probability of the economic system maintaining in a certain state is high, the ability to jump between different states is relatively weak, and the regional system shows the characteristics of stability. Moreover, with the different economic and financial conditions, the impact of financial markets on the real economy is different.

In summary, the impact of financial markets on the real economy is significantly different in different states; Compared with the constant coefficient FCI, dynamic FCI can more effectively capture the changing characteristics of China's financial market, have stronger ability to predict the real economy, and can be used as a leading indicator of the real economy. This thesis provides a policy formulation basis for government departments to monitor the real economy and financial market in real time, and then formulate policies to prevent risks in the financial market so as to better serve the real economy.