white paper

Infima delivers superior prepayment predictions for Agency MBS

Kay Giesecke

Founder, Chairman and Chief Scientist at Infima, Professor at Stanford University

Accurate predictions of prepayment speeds are critical for security selection and portfolio formation in Agency MBS markets. Infima’s transformative, patent-pending deep learning technologies set new prediction standards, delivering performance boosting edges to MBS market participants including investors and dealers. This White Paper demonstrates the superior out-of-sample accuracy of Infima’s prepayment speed predictions during the January 2019—October 2021 period. Performance is shown to be robust across a range of market regimes including the 2020 pandemic. We validate that Infima’s speeds enable high-confidence investment and trading decisions, even in disruptive market environments.

Request Download