Empirical Study On Credit Risk Measurement Of Listed Companies Based On KMV And Logit Models

Under the background of globalization,the scale of international securities market is getting larger and larger.The Chinese government is also opening up some regulations of the financial industry.At the same time,Chinese listed companies are facing huge opportunities and challenges.A set of effective credit risk measurement methods to help the early warning of financial distress will be very necessary,Banks,company management and the majority of market investors all need to assist their management and investment decisions to better manage and invest.From now on,credit risk measurement models mainly include financial indicator ratio model logistic model and quantitative credit risk model KMV model.The calculation characteristics of financial statements and financial indicators themselves have some shortcomings such as hysteresis and quiescence at time points.Therefore,although the former model is relatively simple in practical application,the calculation results also have the disadvantage of hysteresis,failing to capture the rapid changes of the capital market.Mainly through the capital market theory after a model of information data analysis,rather than the book financial information of the company,more timely analysis the situation of listed companies,and that financially troubled risk size,this model has also been a lot of companies abroad application,and formed a default database effectively,but also at home is still in the study,has not yet formed a default database,finally it is concluded that the KMV model variables EDF in the absence of a default database can't exactly,can only roughly estimate.Based on 90 listed companies as samples,this paper on the basis of KMV model,on first according to the situation of our country to adjust some of the parameters of the KMV model,and then calculate the samples of listed companies by using the model of default distance DD,at the same time,the descriptive statistical analysis of the DD,preliminary judging DD of financial distress on the discrimination ability of the company;In the second step,after screening out some appropriate financial indicators,the default distance obtained from KMV model was introduced into the logistic model as an indicator to judge its ability to distinguish companies with high credit risks.Finally,the ROC curves of the above two groups of models were compared and their test effects were evaluated.Empirical results show that after the inclusion of default distance DD,the logistic model's discrimination accuracy of ST and non-st companies is much higher than that of the individual default distance DD.Finally,it is concluded that although the default distance DD can better distinguish financial distressed companies,the logistic model with the default distance index has better discriminating ability,and both the discriminating ability and the the continuous expansion of the global securities market and the deepening of the opening up of China's financial industry,Chinese listed companies are facing more fierce competition and greater risks while meeting the opportunities of reform and development.Banks,corporate management and market investors all need an effective financial distress warning system to help them make decisions.The current financial crisis warning model can be divided into two categories:the financial ratio model represented by logistic model and the credit risk quantification model represented by K.MV model.Although the former model is easy to operate,the characteristics of financial statements themselves determine that it has defects such as quiescence and hysteresis,which cannot reflect the rapid and subtle changes in the capital market.The latter model relies on capital market information rather than book information.Theoretically,it can timely,scientifically and dynamically reflect the current operating status of listed companies,and it has been widely used abroad.However,due to the special national conditions of the absence of default database,EDF,the final output variable of the model,cannot be obtained in China.This paper takes 90 listed companies as samples and expands the model in two steps.First,after adjusting the KMV model according to the actual situation in China,it estimates the default distance DD of the samples according to the KMV model and makes descriptive statistical analysis to preliminarily judge the ability of this index to distinguish financial crisis companies from financial health companies.Secondly.DD of default distance was taken as a variable and other screened variables were introduced into the logistic model constituted by the logistic model.Finally,ROC curves of the above two groups of models were compared and analyzed for their test effects.Empirical results show that after DD is included in the default distance,logistic mode]has a much higher accuracy in identifying ST and non-st companies than that of DD alone.Therefore,it can be seen that the default distance variable can better reflect the financial situation 'of listed companies in China.After adding the default distance variable,the financial distress early warning model has been significantly improved in terms of its discrimination ability and model goodness of fit.