Ginnie Mae Disclosure Data and Reports

It is the bulk disclosure program behind the only public loan-level view of US government-insured residential mortgage collateral: every pool and loan inside the mortgage-backed securities Ginnie Mae guarantees, backed by FHA, VA and USDA loans, published as structured files rather than interactive screens. The program organises into six categories - MBS Single Family, HMBS, Multifamily, Platinum, Factor Files and Other Files.

Single-family coverage is the deepest slice. Daily New Issue files land every business morning in four flavours: pool/security detail, pool supplemental detail, loan-level disclosure and a pool supplemental variant. Monthly cycles follow - Monthly New Issuance, Monthly Portfolio (pool/security, supplemental and loan-level for both Ginnie I and Ginnie II) and Loan Liquidations. HMBS carries the same architecture for Home Equity Conversion Mortgage-backed securities. Multifamily files disclose new issues and portfolio at pool and loan level, add prepayment penalty data and terminated pools, and Platinum covers the multiclass structures that repackage pools. Factor Files track every outstanding pool's factor monthly, while Other Files hold the long tail: annual WHFIT tax reporting, issuer directories, pool- and loan-level forbearance supplements, MBS loan payment history, quarterly (~780 MB) and annual (~2.3 GB) Loan Performance archives, and monthly Conditional Prepayment Rate files.

Every file type ships with a published PDF layout and a sample file, so nothing about the schema has to be inferred. The MBS Loan Level Disclosure File reached Version 2.0 in August 2026 - its first pipe-delimited release, adding Scheduled UPB (LL-54), ARM Adjustment Effective Date (LL-53) and Re-Performing Loan Indicator (LL-55). A Disclosure Data History section retains monthly archives back to January 2020 per file prefix, and a Legacy File Download keeps serving pre-2020-format outputs. Datadory delivers the entire surface as rows.

One row per pool or loan per reporting period, keyed on the CUSIP and pool ID pair, with header records marking each file's period and as-of date. Two illustrative rows in exactly the delivered shape:

record_type : PS            cusip        : 36295XBY0      pool_id     : 683255
pool_indicator : C           pool_type    : SF             issue_date  : 20191001
original_aggregate_amount : 3,544,349     remaining_security_rpb : 3,544,349.58   rpb_factor : 1.0000
issuer_number : 3536        issuer_name  : Usaa Federal Savings Bank
number_of_loans_in_pool : 74   wa_interest_rate_wac : 3.519%   wa_loan_age_wala : 349 mo
maturity_date : 20491020

header_record : HP          period       : 201910         as_of_date  : 20191030

The first row is an October 2019 single-family issuance fresh off the daily new-issue cycle: 74 FHA-guaranteed loans pooled into a $3.5 million security carrying a 3.519 percent weighted-average coupon, its factor still at 1.0000 because no payments have yet been applied. The second is a file header - the record that anchors every delivered extract to its period and as-of date, so downstream joins never guess which month a row belongs to. Live rows for whichever prefixes, programs and periods you name arrive in this identical shape - get a sample of this dataset and it lands cut to your list.

Thirty documented fields anchor the delivered row set below, spanning both levels the program publishes. At pool/security level the identity pair (cusip, pool_id) joins to issuer attribution (issuer_number, issuer_name), issuance context (issue_date, pool_indicator) and the economics trio of original aggregate amount, remaining principal balance and rpb_factor - the number prepayment models live on. Weighted averages summarise each pool: wa_interest_rate_wac, warm months to maturity, wala seasoning, wa_credit_score and wa_ltv.

At loan level the Version 2.0 layout exposes what no other public source shows for this collateral: the insuring agency per loan, purpose, unpaid_principal_balance and the newly added scheduled_upb, borrower credit_score, loan_to_value_ltv and combined_ltv_cltv, debt-to-income as total_debt_expense_ratio, FHA upfront_mip, the first_time_home_buyer flag, months_delinquent, the current_month_liquidation_flag, removal_reason codes that distinguish delinquent-loan buyouts from foreclosures, and the August 2026 addition of the re_performing_loan_indicator. Geography rides along on every loan as state and msa codes. The full attribute sets of the other five file categories are itemised below.

Geography - United States nationwide, because the guarantee follows FHA, VA and USDA lending wherever those programs operate. Geography arrives two ways: implicitly through the universe itself, and explicitly through the state and MSA codes carried on every loan-level record, which let a national book be cut to a metro or a state without any external join.

Temporal - the program runs on a fixed rhythm. Daily New Issue files publish every business morning; Monthly New Issuance posts on the first business day, Monthly Portfolio on the sixth business day, and Loan Liquidations on the fourth. Depth extends backward deliberately: the Disclosure Data History section retains monthly archives per file prefix back to January 2020, and a Legacy File Download continues producing pre-2020-format outputs for panels that need format continuity across the boundary.

Granularity - three altitudes from one key space. Pool/security records give one row per pool for fast full-universe scans. The pool supplemental family adds 32 record types of stratification. And the loan-level files give one row per loan per period - the finest public view of federally insured mortgage collateral anywhere.

API, files, or your warehouse. Daily, weekly, or hourly.

Your cadence is decoupled from the agency's publication rhythm: most teams load their named universe of pools, CUSIPs and programs once, then keep the tables current so each new period's balances, factors and delinquency states diff cleanly against last period's rows. Deliveries arrive normalised to the field dictionary above - fixed-width quirks and pipe-delimited changes absorbed upstream - keyed on cusip and pool_id, in JSON, CSV, spreadsheet-ready Excel tables or XML. A 430 MB monthly portfolio file becomes the same rows your BI layer already joins, and the quarterly and annual Loan Performance archives arrive pre-cut to the cohorts you actually model.

• Prepayment and CPR modelling - factor histories and the monthly CPR files feed speed curves per program, coupon and vintage rather than program-wide averages.
• Credit surveillance - months delinquent, liquidation flags and removal reasons track distress cohort by cohort, with the forbearance supplements separating pandemic-era accommodation from genuine default.
• Issuer and channel analysis - issuer directories and per-pool issuer numbers show which issuers are scaling the FHA, VA and USDA channels quarter over quarter.
• Multifamily due diligence - pool- and loan-level multifamily files expose property-level exposure, prepayment penalties and terminated pools before a position is taken.
• Policy and housing research - the only public measure of who the federal guarantee actually reaches: first-time buyers, debt-to-income bands, LTV distributions, state by state.
• Reconciliation - remaining principal balances and factors checked against custodian and trustee records on identical CUSIP keys.

Fixed-income investors and quant researchers (relevance 3/3) get the factor, CPR and delinquency inputs that government-backed MBS models run on, at whatever pool or program scope they name; see investors quants use cases. Journalists, academics and students (3/3) get a federal guarantee program's reach and borrower profile stated in auditable rows; see journalists academics use cases. Data scientists and ML engineers (3/3) get documented schemas, bulk scale and a stable feature source for CPR, delinquency and forbearance models; see data scientists use cases. Market researchers and consultants (2/3) size the securitised FHA/VA/USDA stock by program and collateral type; see market researchers use cases. Developers building data products (2/3) script against published layouts that change rarely and loudly - and when they changed in August 2026, the changelog said exactly which elements moved.

Within commercial and residential mortgage finance, this record owns the bulk layer: full-universe files at pool, security and loan grain, deep enough to answer portfolio-sized questions offline. The Ginnie Mae Investor Disclosure Portal owns the interactive layer - point queries at one CUSIP or pool number, plus Disclosure Plus dashboards - and shares these keys, so the two belong on the same project. Outside the agency, the Freddie Mac Single-Family Loan-Level Dataset discloses the Enterprises' cousin market but stops at acquisition-quarter performance, while Ginnie Mae's files carry ongoing monthly status to payoff. The FHFA National Mortgage Database samples the whole first-lien market but publishes only aggregates, and the FHFA Public Use Database covers Enterprise acquisitions categorically. If the question is what is happening inside every federally insured pool right now, this is the record. For how it differs from macro rate series, see the comparison with Federal Reserve Mortgage Debt Outstanding.

Three things worth settling upfront. First, name your scope: between daily new issues, six monthly file families and loan-level universes that run into millions of rows per period, the difference between a useful feed and a firehose is the pool list, program set and period range you specify - a sample can start at one coupon bucket in one state. Second, pick your time posture: current-period snapshots, January-2020-onward archives, pre-2020-format legacy continuity and the quarterly or annual Loan Performance cuts are different products from the same source, and the right choice depends on whether you are monitoring, backtesting or reconciling. Third, respect the design constraints: figures originate as issuer submissions the agency does not independently verify, and the loan-level terms exist to keep borrowers unidentifiable - so treat point estimates as reported values, lean on cohort trends, and expect privacy-preserving treatment to carry through every delivery.