Empirical Research On Credit Risk Of Listed Companies Based On Jump Diffusion KMV-Logit Hybrid Model

The issuance of credit bonds by enterprises is an important way of financing.With the acceleration of the reform of China's financial system and the improvement of the openness of the financial market,the frequent defaults of corporate credit bonds in recent years have caused great impact on the stability and development of China's capital market,and also greatly damaged the interests of the majority of investors.The credit bond contains the huge credit risk,among which the investor pays most attention is the default risk.Based on the importance and difficulty of credit risk management,it is of great practical significance and value to study how to accurately measure the default risk of credit bonds and how to detect the measured results dynamically and timely.Starting from the traditional KMV model for measuring credit risk,this paper points out that the basic assumption that corporate asset value follows lognormal distribution and the value changes continuously fail to explain the observed features of "peak and thick tail" and "implied volatility smile" of return on assets.The traditional KMV model does not take into account the jumping behavior of asset values and will underestimate credit risk.To further introduce assets value fluctuate with jump KMV model to measure credit risk,expounds its theoretical basis,and simplify the calculation expression of default distance,is derived based on the maximum likelihood method and least squares estimate for parameters estimation and equation,comparing with jump diffusion KMV model and the traditional KMV credit risk measurement results of the model.The jump diffusion KMV model was combined with the Logit model,and the default distance was added to the traditional Logit model as an explanatory variable to construct the mixed model.The credit risk measurement effect of the mixed model and the traditional Logit model was compared through out-of-sample testing.Through empirical analysis of 20 listed companies and 100 control companies that actually defaulted on corporate bonds,the results show that the traditional KMV model does not take into account the jumping behavior of asset value,which will underestimate the credit default risk of companies.The jump diffusion KMV model has better risk identification and measurement ability,and the asset value of the defaulting company is more vulnerable to the impact of emergencies,will jump more frequently,and has greater jump risk.The KMV-Logit hybrid model well combines the advantages of the two models,comprehensively utilizes the company's stock market data and financial data,and dynamically monitors the default risk of listed companies' credit bonds.Jump diffusion KMV-Logit hybrid model can significantly improve the ability to identify the actual default companies,and more sensitive to credit default risk.