Original data · federal source
Mortgage Lender Transparency
HMDA filing year 2024 · Retrieved 2026-08-23T00:00:00Z · Source: FFIEC/CFPB HMDA Data Browser (public API)
Research question: for the major lenders we cover, what does the federal mortgage-disclosure record actually show — how many home-lending applications did each originate and deny, and how does each lender’s denial share compare with the national picture? Every number below was retrieved from the public HMDA API under each lender’s verified Legal Entity Identifier, and every figure links the exact query that produced it. Affiliate comparison sites do not publish this data — a denial rate cannot be monetized with a referral link.
Read this before the table: a higher denial share does not mean a stricter or worse lender
Decisioned home-lending applications, 2024
| Filing entity (LEI-verified) | Originated | Denied | Denial share* |
|---|---|---|---|
| JPMorgan Chase Bank, National AssociationLEI 7H6GLXDRUGQFU57RNE97 | 80,744 | 12,095 | 13.0% |
| Wells Fargo Bank, National AssociationLEI KB1H1DSPRFMYMCUFXT09 | 37,240 | 17,404 | 31.8% |
| Bank of America, National AssociationLEI B4TYDEB6GKMZO031MB27 | 83,165 | 80,920 | 49.3% |
| Citibank, National AssociationLEI E57ODZWZ7FF32TWEFA76 | 14,104 | 4,681 | 24.9% |
| U.S. Bank National AssociationLEI 6BYL5QZYBDK8S7L73M02 | 74,512 | 54,155 | 42.1% |
| Navy Federal Credit UnionLEI 5493003GQDUH26DNNH17 | 82,022 | 43,754 | 34.8% |
| Pentagon Federal Credit UnionLEI 549300FX7K8PTEQUU487 | 10,514 | 11,315 | 51.8% |
| National — all HMDA filers | 6,197,085 | 2,103,462 | 25.3% |
*Denial share of decisioned applications = denied ÷ (originated + denied). Withdrawn and incomplete files are excluded, so this is not a “denial rate of everyone who applied.”
How to actually use this
The honest uses: seeing how large each lender’s mortgage operation really is (originations), noticing when a lender’s funnel differs sharply from the national 25.3% benchmark — a prompt to ask why (product mix is the usual answer), and verifying that a lender genuinely is a major mortgage originator before applying. The dishonest use would be ranking “strictest to easiest” — the data cannot support that, which is why we don’t.
Methodology & limitations
- HMDA (Home Mortgage Disclosure Act) data covers mortgage / home-lending applications only. It says nothing about a lender's credit-card, auto-loan, personal-loan, or deposit-account decisioning.
- Denial-share disparities between lenders do NOT by themselves prove discrimination. HMDA's public LAR does not include credit score, debt-to-income ratio, or many other underwriting factors, so raw denial shares reflect differences in applicant pools, loan-purpose mix (purchase vs. refinance vs. HELOC), marketing/prequalification funnels, and product mix as well as underwriting standards.
- 'Applications originated' and 'applications denied' are COUNTS OF APPLICATIONS, not counts of unique customers. One person can submit multiple applications (e.g., shopping refinance offers at several lenders, or reapplying after a denial).
- 'Denial share of decisioned applications' = denied / (originated + denied). This intentionally excludes applications withdrawn by the applicant (action_taken=4), files closed for incompleteness (action_taken=5), loans purchased from another institution (action_taken=6), and preapproval-only outcomes (action_taken=7,8). It is therefore NOT the same as an 'overall approval rate' across all applications a lender received, and should not be read as one.
- Figures are for the SPECIFIC FILING LEGAL ENTITY identified by LEI (e.g., 'JPMorgan Chase Bank, National Association'), which is the entity that actually files HMDA data with regulators. This may differ from the consumer-facing brand, and does not include affiliated non-bank mortgage subsidiaries, if any, that file under a separate LEI.
- Large banks' HMDA denial shares can look elevated relative to smaller/specialist mortgage lenders because big banks solicit large volumes of preliminary/soft inquiries and HELOC or refinance applications from existing depositors that convert to formal 'applications' under Regulation C and are more likely to end in a coded denial rather than a quiet withdrawal; this is a data-collection/product-mix artifact, not proof of a stricter or discriminatory underwriting standard.
- LEI verification method: each candidate LEI was (1) matched against the GLEIF LEI registry for an exact legal-entity name match to the institution's National Association / credit-union charter, then (2) confirmed present in the CFPB HMDA 'filers' endpoint (states=DC, years=2024) with a large, plausible loan count, per the task's verification requirement. All 7 target lenders verified successfully.
- 2024 is the latest annual HMDA dataset available via the Data Browser API at time of retrieval (years parameter accepts up through 2025, but 2025 is a partial/in-progress filing year, not a completed annual dataset); 2024 was used as 'latest complete annual year.'
- National benchmark uses the dedicated nationwide aggregation endpoint (view/nationwide/aggregations), which sums all HMDA filers in the U.S. for 2024, not a sum of only the 7 lenders in this file.
LEI verification: each lender’s Legal Entity Identifier was matched against the GLEIF registry and confirmed against the HMDA filers list before any figure was attributed to it. Author: Credit Defense Hub Editorial Team; pending named subject-matter review. Corrections: via the contact page, logged with dates.
Download the dataset
Related
Educational information — not advice
This page provides general educational information about credit, debt, and consumer protections. It is not legal advice, financial advice, or credit repair services, and reading it does not create any professional relationship. Laws, procedures, deadlines, and dollar amounts vary by state and change over time.
For advice about your specific situation, consult a licensed attorney or qualified financial professional. See our full disclaimer.