Module 3 · Execution Benchmarks · HUD published sale results
Best execution starts with observed prices ahead of models. HUD publishes results for every HECM Vacant Loan Sale, Non-Vacant sale, and healthcare-note sale: fourteen transactions of public price history on collateral directly comparable to segments of the 9281 book. SOO coverage: best-execution analysis · investor universe & capacity.
Published UPB excludes accrued interest, MIP and servicing advances. The economic denominators are Updated Loan Balance (ULB) and BPO value. On those, vacant due-and-payable HECMs cleared at 53.6–60.0% of ULB / 67.7–69.4% of BPO across the last two HVLS sales, and occupied collateral runs ~5–7 BPO points cheaper. Any hold-vs-sell comparison quoted "as % of UPB" should be treated as marketing.
The record
| Sale | Bid date | Collateral | Loans | UPB | Loan balance | BPO | Total bid | % UPB | % balance | % BPO |
|---|
Sources: HUD Office of Asset Sales, HVLS 2026-1, HVLS 2025-3 and HNVLS 2025-1 Sale Results Summaries; HLS 2026-1 Results Report (hud.gov, verified June 2026). "Loan balance" = ULB for HVLS, awarded loan balance for HNVLS.
Percent of loan balance and of BPO, by sale. Healthcare notes shown on UPB (no BPO published).
The complete record: 17 transactions, 2016–2025
Every HVLS/HNVLS transaction since the program began, taken from the original results documents, spans two structural eras (pool bidding through 2022-1; individual-loan bidding since) and a 350bp rate cycle. Clearing levels on loan balance have held in a tight 54–62% band since individual-loan bidding began, across every rate environment.
Bars: % of loan balance (gold) and % of BPO (navy). Line (right axis): 10-yr Treasury at bid date. Dashed divider marks the shift to individual-loan bidding.
After the sale: the Commissioner's report
HUD tracks every sold loan through resolution in its semiannual Report to the Commissioner (March 2026 edition: 20,707 settled loans across 14 reporting transactions). This is the oversight evidence any Ginnie Mae sale program will be asked to produce, and the outcome baseline it will be measured against.
All reporting transactions, count basis. "Non-REO" covers paid in full, short payoff/sale, third-party foreclosure sale.
Nonprofit participation went 4% → 50% → 63% as carve-outs expanded (2019–2022), then priority bidding was discontinued for the last three sales. Any Ginnie Mae program design inherits this policy debate, and this data settles parts of it.
Investor universe
The SOO asks its advisor to evaluate "the potential portfolio buyer and investor universe and their capacity to take on additional portfolios, factoring timing considerations." The published award tables answer the first half empirically.
| Repeat buyer | Observed behavior (published award tables) |
|---|
A handful of specialist buyers take most of each sale (GITSIT alone took 1,054 of 1,874 loans in HVLS 2025-3). Concentration cuts both ways: it proves a reliable bid exists, and it warns that pool structure, servicing-transfer terms, indemnifications and data quality determine whether sale pricing is set by one bidder or by competition. Subdividing pools by state, occupancy and balance/MCA band, using the stratifications this portal already computes, is how the SOO's "maximize proceeds while enabling an acquiring entity to sustain the portfolio" becomes an executable auction design.
Post-sale outcomes for all 14 HVLS/HNVLS transactions since 2017 are tracked in HUD's semiannual Report to the Commissioner (March 2026 edition), the compliance trail for any future Ginnie Mae program.
Jelani House served as program-financial-advisor engagement manager at HUD's Office of Asset Sales, working the qualification flows, diligence mechanics, bid economics and settlement operations behind the sale architecture these benchmarks come from, and HSG's loan-sale platform implements that architecture end to end. That combination is what turns published prices into executable strategy.
Next: the market monitor →