Credit Risk Measurement Of Listed E-commerce Company Based On KMV-Logit Model

With the continuous development of the economy,the e-commerce industry is also booming,which not only greatly promotes the development of the digital economy,but also plays an important role in promoting rural revitalization,entrepreneurship and innovation.But at the same time,due to the rapid development of e-commerce industry,there have been a lot of "lack of integrity,fraud everywhere" events,its hidden credit risk is more hidden,and more destructive.Therefore,it is of great significance for the whole society,industry and enterprise managers and investors to construct reasonable and accurate risk measurement model and timely and effective early warning of credit risks.In this thesis,two models of KMV and Logit are selected for research when studying credit risk,and the two credit risk measurement models are combined to build a mixed model of KMV-Logit to test its credit risk recognition ability.In terms of sample selection,since Chinese default database is not perfect,and ST companies have higher credit risk and credit volatility,this thesis takes ST and *ST listed companies in e-commerce industry as default samples,and selects 20 non-ST listed companies as control samples according to the principles of similar size and similar business.In this thesis,2020 and 2021 data from Wind database and Orient Fortune choice database are used for research.Firstly,the default distance in the KMV model is calculated,and the difference in default distance between the default group and the control group is compared to test whether the default distance can identify the credit risk.Then,20 indicators of different dimensions are selected for research,and 8 indicators shared in 2020 and 2021 and passed the test are taken as explanatory variables to construct the traditional Logit model and test its credit risk forecasting ability.Finally,the default distance is added to the Logit model as a new explanatory variable,and the KMV-Logit mixed model is constructed to test its credit risk recognition ability.The research results show that: 1.The default distance in the KMV model can accurately reflect the gap between the default sample and the control sample,and the default distance of most ST and *ST companies is smaller than that of non-ST companies,which has a high credit risk;2.The traditional Logit model,its credit risk prediction accuracy is 85%,has a good forecasting ability;3.The KMV-Logit model has improved its credit risk prediction accuracy to 90%,and the model fitting effect and default prediction effect have been significantly improved and enhanced.Therefore,this thesis believes that KMV-Logit model is more accurate in the study of credit risk identification,and should be widely used.Finally,through empirical analysis,this thesis also puts forward relevant suggestions for the healthy and sustainable development of e-commerce industry,including optimizing the allocation of resources,strengthening the risk management;Improving laws and regulations,strengthening supervision;Establishing default database and improving credit evaluation system.