BD Consumer Finance Corporation Consumer Instalment Loan Risk Control Research

Being a new type of non-bank financial institutions,consumer finance companies,compared with banks,small loan companies,financing guarantee corporation and other traditional industries,have credit products that are closer to the consumer scene.These kind of products reach deeper to medium and lower customer level and the amount of loans is smaller.Consumer credit approval time and stage scene profitability are key requirements in consumption scene.New challenges and opportunities come along with credit risk control for companies to solve the problems of low-end customer level,small and scattered loans as well as inefficient consumption scenario.As a new format,consumer finance companies is still lacking in systematic and complete research,especially the study of risk control.Scientific and reasonable risk control system is the core element in the stable development of consumer finance companies.This paper use the past consumer credit data of BD Consumer Finance Company for analysis,construct the model of consumption credit decision in order to provide BD Consumer Finance Companies with risk control decisions and improve its risk management system.This paper introduces and studies the development process and risk control of China’s consumer finance industry and companies,interprets the development of the BD Consumer Finance Company,its current business situation and risk management status.The traditional credit risk control mode of artificial audit is not suitable for the developmental needs of BD Consumer Finance Company.Quantitative analysis is drew based on the authentic credit customer data collected by BD Consumer Finance Company in the past.Among 30,000 randomly extracted customer data,about 2/3 of data sets is used to train credit decision analysis model,the rest is used as model test sets for testing their analytic results.By using Logit regression,Decision Tree regression,Boosting regression,Bagging regression,Random Forest regression and other methods,a credit decision model based on statistical modeling,data mining and improved statistical modeling methods is constructed.This paper provides empirical analysis and comparison of three credit decision models.Then,based on the statistical modeling method,the improved credit decision model is established and simple cost accounting is used in the model.Based on the model of consumer credit decision and comparison of empirical result,this paper concludes as following:A.Through the three analytic models,companies can identify the risk of loan defaults and also the important variables that affects loan defaults.In the empirical process,it is clear that variables like the borrower’s gender,education level,marital status,fertility situation,family income,living expenditure,have a greater impact on the performance of the borrower’s repayment,of which family income and living expenditure variables are most important.B.Among the three models,improved credit decision model based on statistical modeling stands out in empirical results.According to the empirical results of the model test,through the effective model selection,the probability of repayment is increased from 72% to 80%.The loan default losses borne by consumer finance companies are reduced as well.C.Through comparison,it is clear that although the adoption of effective credit decision model will narrow the scope of the borrower’s access and may limit the size of the loan,the cost of loan default is much lower than the cost of the development of the loan industry,which provides help and guidance to the BD Consumer Finance Company to improve the risk control system and profitability.