themortgagellm™LLPA reality-check: market vs Fannie base grid
“For a 720 FICO / 80 LTV conforming 30-year purchase, what is the median observed LLPA in the last 14 days vs the currently-effective published Fannie base grid?”
Response
For Conforming 30YR Fixed Purchase, FICO 720-739 / LTV >75-80, trailing 14 days:
| Source | LLPA (bps) |
|---|---|
| Optimal Blue observed median | −50 |
| Fannie base grid (currently effective) | +125 |
| Market vs grid | −175 bps (market prices below) |
Market is pricing 175 bps below the Fannie base grid on this cell. A negative observed LLPA in a cell where the grid says +125 means the borrower's all-in points came out as a net credit after seller concessions, promotional pricing, aggregator override credits, or lender-paid LLPA buydowns. This is common in high-competition cells (720-780 FICO, 75-95 LTV) where several aggregators compete aggressively.
What to do with it:
- Rate-sheet engineers: this is your competitive headroom / risk. If your desk holds LLPA closer to grid, you're above market and losing conversion.
- Capital markets: net-credit pricing shows up as a persistent negative in fallout / gain-loss math — feeds into best-execution routing.
- Overlay designers: cells where market already prices well below grid are where an additional lender overlay burns dry economics faster than the model shows.
See Beyond the grid for the fuller within-cell residual-risk analysis and how the LLPA overlay model was calibrated against exactly this kind of divergence.