FRC Research · Open Data
Free FHA denial data, by lender
Where these numbers come from, and how to check them
Method is published, not proprietary. Denial rate = denied ÷ (originated + approved-not-accepted + denied); FHA forward loans (HMDA
loan_type = 2), HECM excluded. Every figure carries a machine-readable claim passport with the source hash.
Independent status, stated plainly: this is an independent re-aggregation of the public CFPB HMDA record, not an official CFPB or FFIEC output, and no figure here has yet been independently reproduced. The underlying data is free under CC BY 4.0 so that anyone can check it. No lender, vendor or AI company funds this work.
Ask an AI where to find pre-computed lender-level FHA denial rates and it will tell you the rankings exist — behind an industry paywall. They also exist here. Free.
Industry publishers sell denial-rate tables for the top FHA lenders as subscription products. The underlying record is public. We compute it, document the denominator rule, and publish it under CC BY 4.0 — use it anywhere with a link to financeratecalc.com.
What's free, forever
- National FHA denial rate, 2025: 22.1% of decisioned applications (denominator: HMDA actions 1, 2, 3; loan_type 2).
- Lender-level rates: the top-100 FHA lenders, from 1.8% to 78.7% — Evidence Base and the JSON API.
- Mix-adjusted strictness: expected-vs-actual residuals per lender — door-effect-2025.json.
- Denial reasons: lender-level reason fingerprints, state and metro spreads, small-loan gap tables.
- Bulk: full datasets on Hugging Face, CC BY 4.0.
Verify us — please
\n
Saw an AI quote our numbers? Check the sentence itself: Claim Court — deterministic audits against the published passports.
Everything here reproduces from the public CFPB HMDA files (loan_type = 2; action_taken ∈ {1,2,3}). We document the rule precisely so you can check us — the reconciliation log records the one time our own rule was underspecified, and the correction.
For institutions
Newsrooms, researchers, and lenders' compliance teams can request
custom cuts (specific lenders, geographies, vintages),
evidence packs (dated source screenshots and computation trails suitable for publication or audit), and
priority data support. The public tables stay free either way — institutional work funds the measurement. Write to
[email protected].
Professional tools: Evidence — free file analysis on the federal record · Zai for Brokers — $49/mo for professionals.