Datadory notebook
Fha Va Loan Level Disclosure Data: Dataset Structure and Field Coverage
Datadory delivers fha va loan level disclosure data covering comprehensive field definitions, entity mappings, and historical time series — structured for direct analytics and delivered on demand.
1,744 datasets. Pick your catch.
What is FHA VA loan level disclosure data?
FHA VA loan level disclosure data means the loan-level records Ginnie Mae publishes for its guaranteed book: MBS, HMBS, Platinum and multifamily securities whose collateral consists of FHA-insured, VA-guaranteed and USDA-backed mortgages. Because Ginnie Mae guarantees that government paper rather than the conforming loans Fannie Mae and Freddie Mac buy, its Ginnie Mae Disclosure Data and Reports program is the only public loan-level view of FHA and VA mortgage performance.
Which datasets cover FHA and VA loans, and how do they differ?
Four cataloged sources touch government-backed mortgage detail, and they divide by grain rather than by agency loyalty. Only one reaches loan level for FHA and VA collateral:
- Ginnie Mae Disclosure Data and Reports - the bulk tape. Daily new-issue, weekly, factor and monthly pool-, security- and loan-level files across MBS, HMBS, Platinum and multifamily programs, with monthly archives per file prefix back to January 2020.
- Ginnie Mae Investor Disclosure Portal - interactive CUSIP and pool-number search across MBS, HMBS, REMIC and Platinum securities, plus Disclosure Plus dashboards for CPR, delinquency, forbearance and cohort analysis. No account required, but no fixed bulk files either.
- Freddie Mac Single-Family Loan-Level Dataset - roughly 56 million conventional loans originated since January 1999 with monthly performance rows. Deeper history than Ginnie Mae, wrong collateral: no FHA or VA records.
- FHFA National Mortgage Database (NMDB) Aggregate Statistics - a nationally representative 5% sample of closed-end first-lien mortgages publishing aggregates by geography, loan type and borrower credit profile. Nationally representative, but aggregates only - microdata are not public at this level.
For fair-lending work on Enterprise purchases instead, the FHFA Public Use Database discloses annual Fannie Mae and Freddie Mac acquisitions at the 2020 census-tract level under 12 USC 4543.
What fields does the Ginnie Mae loan-level record carry?
Ginnie Mae documents every element. Pool and security records carry a nine-character CUSIP, a six-character Pool ID, issuer name, Pool UPB, RPB Factor and weighted averages for coupon (WAC), remaining maturity (WARM), loan age (WALA), credit score and LTV. Loan-level records run through numbered elements ending at LL-55, including property State (LL-35), MSA (LL-36), Credit Score (LL-27), Combined LTV (LL-25), Total Debt Expense Ratio (LL-26), Months Delinquent (LL-21) and Removal Reason (LL-39).
The MBS Loan Level Disclosure File is documented at Version 2.0 (August 2026), the first pipe-delimited release, adding Scheduled Unpaid Principal Balance (LL-54) and a Re-Performing Loan Indicator (LL-55). Every file type ships a published PDF layout plus a sample TXT, and the 32 supplemental record types stratify each pool.
How much does the FHA and VA tape weigh, and how far back does it reach?
Volume is the entry fee for loan-level coverage of the government book. Monthly portfolio files alone run large: llmon_202607.zip is about 430 MB for single family, the HMBS counterpart hllmon1_202607.zip near 469 MB, the quarterly Loan Performance archive roughly 782 MB, and the annual file approximately 2.3 GB.
Where to go next
Start the deeper dive at the commercial residential mortgage finance data guide, which ranks all 13 datasets Datadory pools for the industry and maps them to workflows. Then work the sources directly: the Ginnie Mae Disclosure Data and Reports bulk tape for FHA, VA and USDA loan-level files, and the Ginnie Mae Investor Disclosure Portal when one CUSIP needs an answer, not a gigabyte.
For applied patterns, the site's credit risk screening playbook (/use-cases/credit-risk-screening) walks screening workflows end to end, and the investors quants page (/for/investors-quants/commercial-residential-mortgage-finance) shows how MBS-focused analysts combine these sources.
Pick up where this leaves off
Every one of these ships with sample rows before you commit to anything.
Ginnie Mae Disclosure Data and Reports
FHFA Public Use Database (PUDB)
enterprise · record_num_sf_nfa · metro …+12 more
Ginnie Mae Investor Disclosure Portal
Freddie Mac Single-Family Loan-Level Dataset
PERIOD …+30 more
FHFA National Mortgage Database (NMDB) Aggregate Statistics
Want rows instead of a pitch? Name the datasets.
API, files, or your warehouse. Daily, weekly, or hourly.
Get a sampleQuestions worth asking
Does Ginnie Mae publish FHA and VA loans one by one?
Yes. Loan-level files carry one LL record per loan per reporting period, documented through element LL-55: Combined LTV (LL-25), Credit Score (LL-27), Months Delinquent (LL-21), property State (LL-35), MSA (LL-36) and Removal Reason (LL-39). Version 2.0 of August 2026 added Scheduled Unpaid Principal Balance (LL-54) and a Re-Performing Loan Indicator (LL-55).
Can you identify borrowers in Ginnie Mae loan-level files?
No, and trying violates the terms. Use is restricted to analysis of Ginnie Mae MBS credit performance; identifying any individual or linking records to other information to discern identity is expressly prohibited. Ginnie Mae also disclaims all warranties because the files are compiled from unverified issuer submissions.