Research On China's Shadow Banking Risk Early Warning Mechanism

Shadow banking is one of the important sources of risk in China's current economy.Therefore,it is crucial to prevent systemic financial risks caused by the shadow banking crisis.This paper establishes a shadow banking risk warning mechanism to warn the operation of shadow banking,provides a basis for policy formulation.Based on the previous studies,this paper analyzes the current situation of China's shadow banking,considers the influencing factors of China's shadow banking risk from the internal environment and external environment,and selects 16 quantitative indicators to construct the risk warning indicator of China's shadow banking system,then uses RBF neural network model to construct China's shadow banking risk warning mechanism,to draw corresponding conclusions and policy recommendations according to the training results of the model.The specific empirical analysis process is as follows: Firstly,in order to improve the accuracy of RBF neural network model operation,this paper uses principal component analysis method to reduce dimensionality of monthly data of 16 indicators from January 2011 to June 2018,and calculates the comprehensive safety score of the shadow banking system,and selects the mean value plus or minus one point five standard deviations as the risk warning line,at the same time,uses the data of the September 2008 financial crisis to verify the risk warning line.Secondly,using MATLAB to create RBF neural network model,by training and verifying RBF neural network model,this paper judges the feasibility of using RBF neural network model to predict the risk status of shadow banking,and by measuring whether the comprehensive safety score exceeds its threshold or not can judge whether the operation of shadow banking system is stable or not to make prediction on the security status of China's the shadow banking system security in July 2018.Finally,in order to verify the prediction accuracy of the RBF neural network model,the predicted mean square error values are compared with the BP neural network model and the Logit model.The empirical results show that the gap between the predicted value and the expected output value of the trained RBF neural network is small,which indicates that the trained RBF has high fitting precision and can be used to predict the risk of shadow banking system in China.Using RBF neural network to predict the risk status of shadow banking in July 2018,it was found that the comprehensive security score of shadow banking in July 2018 was more comprehensive than that in June.The score dropped by about 2%,but the risk situation is still at a relatively high level,which requires us to continuously improve the discovery and resolution mechanism of risk,improves the early warning capability of China and the prediction accuracy of early warning systems.