| Line of credit is the maximum amount of loans made by banks to specific subjects over a certain period of time,and more than 70% of commercial bank loans from major international commercial banks are withdrawals under credit lines.In recent years,China’s economic volatility,credit risk management problems are increasingly prominent,the increase of residential leverage,the proportion and importance of retail credit business is increasing,the increase of digital financial products,inclusive financial comprehensive promotion,the identity and behavior characteristics of retail customers make risk control management more complex.How to evaluate the maximum amount of credit that an individual can bear objectively from the perspective of the growth of individual economic value is a forward-looking and exploratory problem.The traditional measurement of credit line is mainly based on static personal income,asset status,the risk limit modeling method of this paper has three main innovations:First,the application of human capital theory,to predict the future value of individual customers’ labor income,on the basis of describing income growth,reflecting the future labor income value uncertainty risk factors.Applied behavioral economics theory incorporates customer behavior(such as consumption habits,travel frequency,etc.)into the model system,which significantly enhances the scientific nature of revenue forecasting.Apply machine learning methods for model training and iteration to improve model accuracy.On the theoretical basis,based on the theory of labor economics,according to the relevant research,the influence of personal income by the industry,current residence area,working years,educational background,etc.is quantified as an income index to measure the value of human capital and form quantitative indicators.In modeling method,this paper starts from the income limit strategy,takes the customer as the core,combines the theory of human capital theory and behavioral economics and other theories,designs the three categories of indicators,:income index,labor force index and behavior index,first of all,divides the customer group,predicts the future development of homogenized group from the customer group,and reduces the model deviation in the statistical sense.On the basis of the group,we consider the value of human capital,measure the future value of labor income from the three dimensions of income growth,income volatility and tail risk,predict the change of customer’s income,measure the risk and willingness of individual repayment,form quantitative indicators,and determine the maximum amount of debt that customers can bear.Of the amount measured by the model,95.96% can be covered by the measured amount of the Ministry of Housing(Personal Credit Line Measurement Department) and 98.64% by the actual credit limit.Through the analysis of variable characteristics,the growth rate of the reverse working life decreases year by year,there is no absolute correspondence between academic qualifications and risk limits,and there are differences in various regions of the trend of risk limit coefficient change.Through this modeling analysis,the paper puts forward some suggestions on the time series change of customer behavior,dynamically adjusting credit lines,using data to form differentiated credit methods and policies,and increasing the dimension and quality of customer characteristic data collection. |