The Empirical Research Of Credit Risk Of China’s Listed Companies Based On Logit Model

In 2007, the Subprime Mortgage Crisis happened in US, and inspires the global financial crisis. Subsequently, the financial institutionsaround the world are paying more attention to credit risk management.The Institute of International Finance and the Ernst & Young Global Limited observed that the credit risk has gradually become the main concerns of the chief risk officer.In recent years, the size of credit bond in China increased from 1.59 trillion in 2010 to 14.6 trillion in 2015, with an average annual growth rate of 51.6%. With the slowdown in economic growth,the pressure on corporate debtis getting larger and larger.According to the statistics data of Haitong Securities, the defaultnumber of credit debts increased from 6 in 2014 to 24 in 2015. With the default incidents happened more frequently, the China’s credit risk exposures have increased rapidly.Therefore, the financial institutions must strengthen controls at the credit risk, and dig out more practical, more accurate and more convenient model for Credit Risk Assessment.At present,the credit risk management system of the western countries has been more perfect, and the credit risk quantitative research techniques such as Logit model, KMV model and CDS modelhave been utilized widely. The credit risk system of our country is still imperfect;theinnovative credit risk assessment technologies are seldom used because of data missing. Based onthe actual participation in the project and referring to a lot of literature, the paper uses the Binary logit regression model and the ordered logit regression modelto evaluate the credit risk of thelisted companies.The analysis datawere randomly sampledfrom Shanghai and Shenzhen Stock Markets. When using binary logit regression method, the stepwise regression and the principal componentanalysis method are used. When using ordered logit regression method, the factor analysis method and the maximum likelihood estimation method are used.The Empirical results show that:(1)the logit regression model possesses more perfect prediction performance than the linear probability model;(2)the indicators such asequity ratio, net profit rate, net profit margin, fixed asset turnover, Current asset turnover, and growth rate on equity can used to forecast credit risk of listed companies;(3)the factor analysis method has better applicability on listed companies credit rating.The paper believes that, the financial institutions should use logit regression model to evaluate the credit risk of listed companies, and timely understanding of the listed company’s credit rating and credit trend. This is beneficial not only to strengthen the credit supervision of listed companies, but also help to safeguard the interests of investors.