Why use the AFT Prepayment and Credit Model?
Gain reliable insights into voluntary and involuntary risk with a long-standing, trading-quality prepayment and credit model.
Unique data advantage
Uses unparalleled data assets, including our industry-leading servicer contributed loan-level database, to develop, validate, monitor and recalibrate, the model.
Sophisticated modeling approach
Realistically captures how borrowers behave in different scenarios by using an innovative technique that combines the best of statistical and behavioral modeling.
Single structure to project loss
Provides a more complete understanding of risk throughout the loan life cycle by projecting prepayment, default and loss severity in a single structure.
Reliable results in various economic scenarios
Used by lenders, servicers and capital markets professionals across the U.S., the AFT Prepayment and Credit Model evaluates multiple risk factors and provides detailed forecasts for residential and multifamily mortgages. The model’s predictive power is based on its unsurpassed data, proven methodology, flexible structure, and transparent and tunable parameters – all of which allow it to quickly adapt to market changes.
Our experienced team of modelers, data scientists and financial engineers test and recalibrate the model on a regular basis to help ensure it provides the best fit for today’s mortgage market. The model has also been extensively validated and tested by clients’ internal teams, third-party agencies, regulators and government agencies.
With ICE market-leading servicer contributed loan-level data, the model is able to cover a vast range of loan programs and products spanning all credit grades and geographies. ICE’s high quality data, combined with powerful, innovative modeling techniques that capture borrower behavior, enable AFT to deliver reliable results in various economic scenarios.
Model differentiators
Provides strategic, proactive and actionable analytics to the right people at the right time
- Built on industry-leading, servicer-contributed loan-level data
- Integrates with ICE’s hyper-local HPI including 23,000+ zip codes for precise and current LTV estimates
- Adapts quickly to changing interest rates and housing markets
- Offers an open structure for model tuning and customization
- Includes up-to-date and explicit modeling treatment to handle the impact of government programs
- Integrates with major third-party analytic platforms
- Delivers strong performance in both pre- and post- crisis environments
- Includes full collateral type coverage
- Allows users to tune the model to address the nuances of a particular specialized portfolio
- Extensive client support by experienced financial engineers and dedicated client support with decades of experience