A Default Prediction Model In Installment Sales Using Non Financial Data

Abstract

In recent years, a credit scoring service that based on personal behavioral data has been providing in the world. Alibaba Group’s Ant Financial in China developed Sesame Credit and launched its service in 2015 as an additional function of Alipay. Credit score(risk) is mainly used in case of calculating loan limit and loan interest rate. On the other hand, a personal credit score affects not only loans but also to benefit public services, job hunting, and marriage hunting. In many countries, personal credit scores will be created as part of various services and become credit score society as in China. Personal credit scores have been already performed practically, but there is not so much research on academic at present. In particular, there is not research on the personal credit risk using only non-financial data without financial data. Generally, financial data is used principally to measure a company’s credit risk. However, the use of financial data to measure a personal credit risk is considered dangerous. Therefore, in the latter case, it’s important to calculate credit risk using only non-financial data without using financial data. In this study, we examine the method to calculate personal credit score using non-financial data.Firstly, we describe the flow of a customer who purchases a product using relation diagram until it defaults. Secondly, we performed a two-group discriminant analysis using the default variable as the objective variable and adding some default factors to the explanatory variables. As a result of the correct discrimination rate, discrimination was possible, albeit slightly. In addition, it turned out to be meaningful to analyze personal credit using only non-financial data.