MBA Analysis: Single Credit Score Has Minimal Pricing Impact

An analysis published Friday by the Mortgage Bankers Association (MBA) suggests that adopting a single, randomly selected credit bureau score instead of the current multiscore "decisioning" method would have a negligible effect on loan pricing and guarantee fee revenue for government-sponsored enterprises (GSEs). The MBA's research, which analyzed approximately 105,000 mortgage applications from the first half of 2025 using Intercontinental Exchange (ICE) McDash loan application data, found that a randomly chosen credit score landed in the same Fannie Mae loan-level price adjustment (LLPA) bucket as the current decisioning score approximately two-thirds of the time. Furthermore, about 90% of these randomly selected scores fell within one pricing bucket above or below the decisioning score.

These findings emerge amid ongoing discussions within the mortgage industry regarding potential changes to credit scoring requirements. Proposals have included moving away from the established practice of requiring multiple credit bureau scores for mortgage underwriting. For its analysis, the MBA employed the methodology detailed in Fannie Mae’s Selling Guide to determine decisioning credit scores. When three borrower credit scores were available, the middle score was used; for two scores, the lower score was selected; and if only one score was reported, that score served as the decisioning score. The study excluded loans with co-borrowers and applications where credit scores were below 500.

Researchers then simulated a single-score approach by randomly selecting one available bureau score for each application. This simulated score was then compared to the decisioning score to determine its corresponding LLPA pricing bucket. The analysis revealed that among borrowers with decisioning credit scores between 700 and 719, nearly 68% of randomly selected scores fell into the identical pricing bucket. An additional 91% landed either in the same bucket or one bucket higher or lower.

The MBA's report indicated that movements between adjacent pricing buckets, both upward and downward, occurred with roughly equal frequency. This suggests that the net change in LLPA revenue for GSEs would be minimal if a single credit score approach were implemented. The study provides quantitative data to inform the ongoing debate about credit scoring practices in the mortgage market.