Real-time Monitoring of China’s Financial Stress—Evidence Based on Mixed Frequency Big Data Dynamic Factor Model

Preventing and resolving systemic financial risk is of great significance to maintaining financial security. The key issue is to accurately describe and monitor systemic financial risk in real time. Based on the mixed frequency dynamic factor model, this paper constructs a daily frequency China’s Financial Stress Index（FSI） by comprehensively using traditional financial statistics and Internet search big data to monitor China’s systemic financial risks in real time. The research shows that the constructed FSI can accurately measure and monitor domestic systemic financial risk, and the phased change characteristics of the index and the identification results of regime state are highly consistent with the actual evolution of systemic financial risk. The uncertainty of expectation and risk perception of economic subjects are the main sources of financial stress, and China’s financial stress has strong sustainability. The introduction of big data indicator can significantly improve the estimation of FSI model and promote the prediction effect of FSI on macroeconomic variables. The FSI on “high level” regime state is better in predicting month-on-month CPI growth, while the FSI on “low level” regime state is better in predicting GDP growth. In the future, regulatory authorities need to continue to optimize the financial risk monitoring and early warning system, improve monitoring capabilities, and reasonably guide market expectations.