| Although China’s bond market started late,but the development is relatively fast,at present China’s bond market is very huge.In March 2014,the default event of "Chaori Bond" broke the belief of "rigid payment".After 2018,the "bond default wave" broke out for the first time in China’s history,the number of defaulted bonds and the amount of defaulted bonds increased significantly,and a large number of various bonds issued by various companies defaulted one after another,especially the bonds issued by private enterprises.Therefore,it is very important to measure the credit risk of private enterprises.This paper first introduces the KMV model and BP neural network model used in the empirical study,then introduces the overall credit risk of private enterprises,analyzes the situation and causes of private enterprise bond default.Finally,the paper makes an empirical analysis of the default risk of private enterprise bonds.The main idea is to use KMV model to measure the bond default risk of private enterprises.However,considering that this model is only applicable to listed companies,the BP neural network model is considered in this paper to solve the problem that the asset value and volatility of non-listed companies cannot be directly calculated by KMV model.The financial index data,asset value and its volatility of listed private enterprises are used for neural network training,and the trained BP model is used to predict the asset value and its volatility of non-listed private enterprises,and the default distance and expected default rate of listed and non-listed private enterprises are calculated by combining the KMV model.The default distance is compared with the rating of the issuer by the authority,and the expected default rate is used for early warning of private enterprise bonds.Secondly,the default distance of private enterprises is analyzed and evaluated,and it is found that the default distance of different industries is different.Finally,the traditional PFM model is constructed to calculate the default distance and expected default rate of non-listed private enterprises,and the results of PFM model and BP-KMV model are compared and analyzed.The empirical results show that the BP-KMV model is more effective in predicting the default risk of private enterprises,and it is more accurate than the PFM model.The results of average default distance obtained by BP-KMV model are consistent with the results of the latest credit rating of private enterprises.At the same time,the default risk faced by listed private enterprises is smaller than that of non-listed private enterprises,and the default risk faced by different industries is also different. |
