Dynamic monitoring of financial security risks: A novel China financial risk index and an early warning system

Abstract

From the perspective of financial risk–a more macro level, this letter selected three levels of indicators which reflect emerging risks, synthesized seven dimensional indices, and developed a China financial risk index using two different methods, identifying the risk regime by Markov switching model. The convolution for neural network–long short-term memory model was used to construct an early warning system for financial risks. The model was optimized using regime-based prediction. The empirical results show that the composite dynamic monitoring system and the early warning system have good effects.