Imbalanced Chinese Listed Corporate Overseas Bond Default Risk Early Warning And Analysis

Under the background of deleveraging and the National Development and Reform Commission gradually opening up overseas financing of Chinese enterprises,more and more domestic enterprises have turned to overseas bond market for financing.However,in recent years,overseas bond thunderstorms have occurred frequently.The reasons are market supervision,market development,investor behavior,economic slowdown and other factors.On the one hand,frequent defaults of overseas bonds will affect the reputation of Chinese companies in overseas markets,lead to the rise of financing cost,and then cause enterprises to find it difficult and expensive to finance;On the other hand,it will make an impact on our exploration of establishing overseas bond market with Chinese characteristics by integrating with international standards and improving market governance effect.Therefore,it has a certain reference significance for the relevant regulatory authorities to improve the overseas bond market system or for the bond issuing bodies to strengthen the prevention and resolution of risk awareness.In view of the above situation,this thesis concludes the influencing factors and model selection of the bond default risk by reviewing the existing literature on the bond default risk,and analyzes the present situation of the overseas bond default risk in our country.First,according to the characteristics of the sample,use SMOTE over-sampling technique to balance the original sample;Then,using logistic regression as the benchmark model,support vector machine,random forest and neural network three machine learning models were used to predict the default risk of post-treatment sample and original sample,respectively.Finally,based on the overseas bond default risk forecast result,this article constructs our country overseas bond default risk early warning system.If the default risk forecast result exceeds the risk threshold set by the early warning system,the default risk early warning system will show the risk and give a warning.The innovation of this thesis is mainly reflected in the following aspects: Firstly,A default risk early warning model applicable to China’s overseas bonds is constructed,which can monitor the default risk and early warning situation of companies issuing overseas bonds,help Chinese enterprises fully understand their own bond risk,expand China’s overseas bond market,and alleviate the financing difficulties of enterprises;Secondly,The model applicable to China’s overseas bond default risk early warning can help China establish and improve the management system of the overseas bond market default risk prediction,early warning and disposal,and improve the risk prevention ability of China’s overseas bond market;Thirdly,Because of the particularity of the bond sample,SMOTE over-sampling technique applied in this thesis to balance the bond sample,greatly improve the prediction and warning effect of the model;At last,For the first time,this thesis conducts an empirical study on the overseas bond samples of Chinese listed companies,comprehensively predicts the default risk of overseas bonds of Chinese listed companies,and builds an early warning model to warn them,which makes up for the blank of empirical research in this aspect in the present academic.The following conclusions are drawn from this study: Firstly,The use of SMOTE over-sampling technique in the balance bond sample could significantly improve the forecast and warning effect of the model,and smote should be most effective as the sample plus-minus ratio approaches 1;Secondly,Based on the accuracy rate and accuracy rate index,the effect of the three machine learning models of support vector machine,random forest and neural network is significantly better than that of logistic regression model;At last,The early warning system based on the prediction results of overseas bonds in this thesis can well identify the default risk of bonds and send out early warning signals.Finally,the suggestions in this article are mainly divided into the following points: Firstly,companies need to improve their own risk management awareness and identify their own risk status as early as possible according to the early warning model;Secondly,Investors match their own risk preference with bond risk and choose the appropriate investment target;At last,The relevant supervisory authorities of our country will strengthen the mechanism of information disclosure,control risk with early warning model at macro level,and accelerate the establishment and perfection of our overseas bond market system.