Research On The Credit Scoring Of Personal Loan Fromauto Finance Companies Based On LOGISTIC Model

Auto industry has now become one of the pillow industries to boom China’s very fast developing economy.As estimated,auto financial business accounts for 20%of the total profit from the Auto Industry Chains(AIC).As profits increase,auto financial corporations have emerged like mushrooms after spring rain.According to incomplete statistics,the penetration rate of China’s auto consumption now accounts roughly for 25%to 35%while it was about 5%ten years ago,which still falls substantially behind some developed countries(70%).The gap indicates a great potential for China’s auto market.As China’s Auto Industry Society estimated,the increasing number of the 1990ers will surely trigger the fast development of auto financial markets,which will undoubtedly become a new growth point.The penetration rate of China’s auto finance business will reach 50%by 2020,i.e.market scale reaching 200million Yuan RMB,according toDeloitte&Touche.However,auto financial markets in China have been left unstable.There is still loan default.There are some main reasons for increased loan default risks such as poorly-established credit system,ill transparency of individual income,frequent interregional population migration,particularly,the incomplete individual credit rating system.Therefore,it is of great feasibility and significance in theory and in practice to study how auto financial companies keep under control loan default risk by improving individual credit rating.The thesis has presented an overview of the closely related research of auto finance at home and abroad,and has undertaken to analyze how to establish an efficient individual credit rating to keep a balance between checking efficiency and default risk.Based upon the analysis of 25,000 sampled clients and their information about loan and credit rating set in one large-sized auto company,by means of questionnaires,the research has screened eight substantial variables about the individual auto loan risk.The 18,592 clients have been sampled to build an individual credit rating model by using Logistic Model Regression,and the 7,970 clients have been sampled to check the function of individual credit rating model.The analysis of inspection has shown that all the regression coefficients of variables were negative.When WALD Inspection P was less than 0.05,variables of the model agreed with actual volume of business;when VIF was less than 10,there was no multicollinearity in the model;when K-S value was 32.59 and GINI coefficient was 44.82,the model had a capacity of separating good accounts from bad accounts.One the other hand,the study is intended to established a cut-off line in accordance with rating and default,so as to rate default risks from clients according to a finalized individual credit rating.The thesis has pointed out that the authentic and complete individual data is vital for a well-established rating model.It has provided some possible suggestions about well-coordinated prevention and control of auto finance loan risk,for instance,the strengthening of trainings of staff,improved competence of workers,the establishment of accreditation system of auditors,established risk-sharing system as well as improved individual auto loan system.It has concluded that anti-fraud rating model is very necessary.