Banksalad carried out joint research with Sogang University to analyze the relationship between consumption behavior and credit risk using MyData. Based on the findings, Banksalad plans to develop an alternative credit scoring model to advance financial institutions' credit assessments and strengthen inclusive finance.
On the 14th, according to the financial sector, Banksalad and the research team of honorary professor Nam Ju-ha at Sogang University released a joint research paper titled "Consumption behavior and personal credit risk: MyData-based." This study is the first case in Korea and abroad to empirically analyze the relationship between consumption behavior and personal credit risk using actual transaction consumption data, and it used about 200,000 of Banksalad's card payment data points and a consumption category classification system.
According to the results, the more sustained the consumption in the medical and health sectors, the lower the default risk. In contrast, the higher the share of expenditure on telecommunications bills, convenience stores, and cafes and snacks, the higher the default risk.
Based on these findings, Banksalad, together with Honest AI and KCB, is developing the alternative credit scoring model "Banksalad Score" and pushing for its commercialization in the financial sector.
Banksalad Score reflects not only consumption data but also financial and behavioral data such as cash flow information by financial assets and platform usage patterns. It also used Machine Learning algorithms to finely distinguish between prime and high-risk borrowers and achieved 60% in the K-S statistic (Kolmogorov-Smirnov statistics) evaluation. This is considered an "excellent discriminatory power" level.
A Banksalad official said, "We will continue to pursue data-driven financial innovation so more customers can enjoy financial benefits and inclusive finance can be realized."