Mortgage Data Explained

Polygon Research uses a proprietary, rules-based algorithm that applies the actual CFPB Ability-to-Repay (ATR) and Qualified Mortgage (QM) logic to every loan record. While some tools simply guess based on the lender's name, we analyze loan-level attributes—like interest rate spreads, DTI ratios, and specific product features—to accurately classify the loan status.

Most lenders labeled as "Non-QM" actually originate a mix of QM and Non-QM products. Relying on a lender list creates "noisy" data. Polygon’s approach looks at the DNA of the individual loan, allowing you to see the true Non-QM volume even at traditional banks that are quietly expanding their portfolio business.

Yes. Transparency is a core pillar of our data science. Our flag identifies the specific trigger for the Non-QM status—whether it was a pricing threshold (Rate Spread), a prohibited product feature (Interest-Only or Balloon), or a points-and-fees violation. This detail is critical for lenders refining their own product strategy.

Our algorithm is dynamic and year-specific. We account for the transition from the old DTI-based General QM rule to the current price-based General QM rule. This ensures that when you benchmark your performance against 2021 or 2022 data, the classifications are compliant with the regulations that were actually in effect at that time.

Yes. Unlike reports that only look at the securitization market, Polygon Research analyzes the full universe of over 67 million closed loans in the HMDA microdata. This provides a total view of the market, including the significant volume of Non-QM loans that are held in private portfolios and never reach the secondary market.