The Research On Prediction Model And Application Of Chinese Private Enterprise Credit Bond Default:Based On The Logit Method

Since the first substantial default of Chaori Bond in 2014,China’s bond market has experienced several rounds of defaults,and cyclical default risk exposure has accelerated.With the development of the financial market,it becomes more important to properly measure the credit risk.In contrast,due to the late occurrence and too small samples,there is a lack of systematic research on bond defaults in the academic areas of our country.Enhancing the ability to judge the potential default risk of bonds will be of great benefit to investors in improving risk response reserves and active structural adjustments,as well as the timely supervision and maintenance of market order by the supervisory authority.As private enterprises are more susceptible to the economic cycle,their financing capacity is weaker than state-owned enterprises,which leads to higher survival pressure than state-owned enterprises,therefore the proportion of default events is relatively high.Therefore,this article constructs a predicting model of bond default from the perspective of private enterprises.This article selects 317 credit bond issuers in private enterprises from 2016 to 2019,and establishes a predicting model through logit regression.Firstly,we extract several significant variables from current literature about default risk,and draw on the factors of the literature about financial distress pre-warning models and accounting misstatement models,and divide the independent variables from the external and internal perspectives,and select the effective variables through correlation analysis and stepwise regression method.Secondly,logistic regression was used to estimate the regression parameters and probability ratio.Then,we set the discrimination threshold and select the best decision point to build a model with higher prediction accuracy.Finally,an evaluation test is conducted on the established bond default predicting model,and select Huaxia Happiness as the newest default example to examine and analyse the applicability of model.The innovation point of this article is that we choose several variables beyond the scope of financial indicators,selecting more indicators that reflect the macroeconomic conditions,and introducing more comprehensive factors affecting default risk,so as to improve the overall forecasting ability of the model.The research finally got a higher prediction accuracy,which made a certain contribution to the further construction of a more systematic and complete bond default early warning model.