Module 2 · Hold vs. Sell · Interactive

Hold or sell — decided per segment, not per slogan.

The SOO asks for "insight into the valuation and marketability of Ginnie Mae's portfolios regarding hold or sell." That answer is different for a loan at 96% of MCA than for one at 60%. This engine runs the real May-2026 segments through both paths — with every assumption exposed on a slider. A scenario tool built on public benchmarks, not an appraisal.

The economics in one paragraph

Hold rides to FHA assignment. Sell takes the market's price today.

The HOLD path

Balances accrue at note rate + 0.5% MIP until they hit 98% of MCA; Ginnie Mae funds the mandatory buyout, then assigns eligible loans to FHA and recovers the claim — effectively par on the balance, capped at MCA. Hold value = those proceeds, discounted for time, minus servicing and advance carry, with a haircut for loans that can't assign (due-and-payable, deceased-borrower, title issues) and resolve through foreclosure instead.

The SELL path

A sale converts the segment to cash now at the market's clearing level and transfers the servicing burden, buyout funding obligation, and unfunded LOC draws to the buyer. The public anchors: HUD's own HECM sales clear at ~54–60% of loan balance for due-and-payable vacant collateral and ~5–7 points lower occupied — while assignment-eligible, claim-backed loans command far stronger executions. Where each segment prices between those poles is exactly what structuring is for.

The FCRA wedge: government hold-value is computed under Federal Credit Reform Act conventions (cohort single-effective-rate discounting — FY2025 MMI actuarial review), while the market prices at risk-adjusted yields. The same cash flows produce two different "values." A disposition advisor's first job is to make that wedge explicit, segment by segment — which is what the sliders below do.

New input: the collateral cushion

This book runs at an estimated 41% current LTV — price severity accordingly.

HSG's FHFA roll-forward of all 53,780 origination appraisals (see the equity module) estimates $31.8B of current property value behind $12.9B of balances — and zero loans underwater. The non-assignable recovery slider below defaults to a conservative 60%; against a 41% WA LTV book, resolutions through foreclosure or payoff should test materially higher. Equity depth is also the strongest argument for challenging deep market-bid discounts loan by loan rather than accepting them book-wide.


The engine

The real book, in five segments.

Segments are the actual May-2026 balance/MCA bands. Defaults are anchored to public benchmarks (shown beside each slider); drag anything.

Global assumptions

5.50%
6.75%
$350
12%
60%

Anchors: WA note rate 6.28% (+0.5% MIP); HVLS 2025-3 cleared 53.6% of loan balance (vacant D&P); FHA assignment ≈ par to MCA. FY2026 CJ: >$500M monthly transaction volume.

Hold NPV vs. sale proceeds, by segment

Bars recompute live. The verdict strip below totals the book under your assumptions.

Hold NPV (assignment path)Sell proceeds (at slider price)
total hold NPV under current assumptions
total sale proceeds under current assumptions
best-execution mix — take the better path per segment

What the engine consistently shows

The near-buyout cohorts are hold-shaped. The deep book is where sale competes.

Loans at 90–98% of MCA are months from a claim-backed par recovery — selling them forfeits nearly-realized value unless a buyer pays through it. The long-dated, sub-80% book is the opposite: a decade of carry, ops cost and LOC funding stands between today and assignment, so a competitive market bid can beat the discounted hold path — if structuring, data quality and indemnification terms bring real buyer depth. That segmentation — not "sell the book" or "keep the book" — is the disposition strategy the SOO is asking for.

Engine math, simplifications and every default's source: Method. This page is a decision-structuring tool, not a valuation or investment advice.