Can Big Data Defeat Traditional Credit Rating?
51 Pages Posted: 11 Jan 2019
Date Written: January 6, 2019
This paper examines the impact of large-scale alternative data, or big data, on predicting consumer loan delinquency for a traditional lender. Using a unique dataset containing over 3,000 variables, we construct a big data credit score by applying machine learning techniques to deal with high dimensionality and massive missing values. We find that the big data score predicts a borrower’s delinquency likelihood with 18.4% greater accuracy than a lender’s internal rating. The combined model (using both the big data score and internal rating) predicts with 22.6% greater accuracy. Our simulation shows that using the big data score increases the net present value per applicant by $220.
Keywords: Big Data, Fintech, Personal Credit, Alternative Data, Credit Evaluation
JEL Classification: G10, G21, G23
Suggested Citation: Suggested Citation