Buy-now, pay-later loans, which typically divide payments into equal portions over a six-week period, surged in popularity during the pandemic and have remained a preferred option among younger consumers and those without access to traditional credit. Yet, until now, the widespread use of these loans has largely gone uncounted in traditional credit models, raising concern among regulators and lenders about so-called “phantom debt” — liabilities invisible to underwriters.
The change comes at a time when signs of stress are beginning to show. According to the Federal Reserve’s annual report on household financial well-being, roughly one in four BNPL users reported missing a payment in 2024 — a significant jump from the year before.
FICO’s initiative follows moves by providers like Affirm Holdings, which earlier this year began submitting its loan data — including “pay in four” arrangements — to Experian. Although this reporting is not yet factored into legacy scoring models, the foundation is being laid for more robust credit tracking.
Historically, the credit industry has struggled with how to assess BNPL borrowing. Unlike revolving credit lines or fixed-term loans, BNPL plans tend to be short in duration and low in dollar value. What’s more, current models often penalize consumers for opening multiple accounts in a brief window — a behavior common among BNPL users but not necessarily indicative of risk.
To address these discrepancies, FICO’s new models group BNPL loans together, evaluating them as a collective pattern of behavior rather than penalizing each account individually. The firm trained the scoring system using a dataset of over 500,000 users, in collaboration with Affirm. Early findings suggest that borrowers with frequent BNPL usage — five loans or more — generally saw stable or improved scores under the new model.