10T Comes to Credit Score Competition: Three‘s a Crowd?

Research evidence excerpt

10T Comes to Credit Score Competition: Three‘s a Crowd?

Nick Maciunas AC (1-212) 834-5671 North America Securitized J P M O R G A Nnicholas.m.maciunas@jpmorgan.com Products Research

J.P. Morgan Securities LLC

Ani Gelashvili AC (1-212) 834-2605 16 July 2026

ani.gelashvili@jpmorgan.com

10T Comes to Credit Score Competition

Three’s a Crowd?

• If our research has been helpful this year, please consider voting for us in the Extel

Survey in the “USA: Structured Securities” MBS categories! There are four: Resi-

dential Mortgage-Backed Securities Strategy/Nonagency, Residential Mortgage-

Backed Securities Strategy/Agency, USA: Prepayments Strategy, and USA: Agen-

cy CMOs & Derivatives. Thank you! Ballots can be requested here

• On July 1, the GSEs released historical FICO 10T data (covering 2013-2025 origi-

nations) as well as a new slug of historical VS4 scores (extending coverage from

2023 to 2025)

• The GSEs have not yet started the 10T pilot for originators, but the model has been

approved for use by FHFA; it remains unclear at this point whether the goal is to

have all three models (Classic FICO, 10T, and VS4) in use at the same time or

whether 10T will supplant Classic at some point

• In our analysis of the historical data, we find that 10T does a better job of classifying

credit risk than VS4 (which seemed to be a better differentiator than Classic

FICO). Of course, there are always caveats to this style of analysis—the key one

being that the historical record is populated with loans that were made with Classic

FICO above the old cutoff

• In any event, we find that the rough rule of thumb for converting from 10T to Clas-

sic is to take off 10 points, but that just like with VS4 (for which the rule of thumb

is 20 points), it’s not a one-to-one mapping