Asset-Depletion and Retirement-Income Qualification for the Parents' Purchase: the $150k Gift Does Not Clear, and Co-Signing Costs More Than Writing a Bigger Check
The question
Verbatim: "What asset-depletion or retirement-income-based mortgage qualification programs exist, and what would they imply for how much cash the founder needs to front for his parents' $450-600k home purchase?"
Context: this is Critical Open Question #1 in [[readme]] - the one that decides whether the founder co-signs (with DTI consequences for his own 2027 build) or gifts cash instead. The sibling brief written tonight covers the jumbo construction-to-perm envelope for his OWN build; this brief stays on Project B and on the co-sign-vs-gift fork.
⚠️ THE VAULT SUPPLIES NO REAL FIGURES FOR THE PARENTS. A full-text sweep of
01-projects/,04-finance/, and06-reference/research/returns only planning references to "parents" - no income, no Social Security amount, no pension, no account balances, no ages, no location. [[index]] has recorded "Parents' mortgage qualification clear - Unknown" since a June 30 due date that passed seven weeks ago. Every number attributed to the parents below is a stated placeholder assumption, listed in the Assumptions block and never presented as fact. The structure of the answer - which program, which divisor, which cash gap - is robust; the specific dollar figures move when the founder supplies real numbers.
What we already know (from the vault)
- The current plan is $600k house, $150k gift, $450k mortgage carried by the parents on retirement income, founder co-signs "if needed." [[readme]] The Q1 2027 gate in [[milestones]] passes on either "pre-approved on retirement income alone" or "willing for founder to co-sign" - the two are treated as interchangeable. They are not. That is the core finding here.
- The sequencing in [[milestones]] is fatal to the co-sign path. Parents close 2027-04-30; the founder's construction loan closes 2027-06-15. Six weeks apart. Fannie's contingent-liability exclusion needs twelve months of documented payment history (below), so a co-signed PITI would sit on his application at full weight.
- The founder's own build is income-constrained, not capital-constrained, and it is jumbo-portfolio. Practical mortgage ceiling ~$1.0-1.2M; sources and uses close with near-zero margin at the bottom of the build budget and are ~$400k short at the top. [[2026-08-14-construction-to-permanent-loans-florida]] [[readme]]
- The vault already warns against funding both in the same window. "Don't go full stretch and fund the parents' home in the same window" and "Don't fund the parents' $150k gift, a $2M build, AND keep a full reserve in the same 18 months - pick two." [[2026-06-04-home-affordability-build-vs-buy]]
- The $150k gift figure was never derived from the parents' finances. [[milestones]] records the rationale as "keeps founder's lifetime gift exemption use modest; leaves ~$450k mortgage for parents to carry on retirement income (typical FL retiree mortgage qualification ceiling)." That parenthetical is an unsourced heuristic, and the arithmetic below does not support it.
What the web says
- The section number in the question's framing is dead. Fannie's B3-3.1-09 was reorganized out of existence in the March 4, 2026 income-assessment restructure (Announcement SEL-2026-02, mandatory for lenders no later than June 1, 2026). The live section is B3-3.4-06, Employment-Related Assets as Qualifying Income (03/04/2026); retirement/pension/annuity income is B3-3.4-03; Social Security is B3-3.4-15; nontaxable gross-up lives in B3-3.1-01.
selling-guide.fanniemae.com/sel/b3-3.1-09/...now 301-redirects. Anything citing B3-3.1-09 - including most of the blog corpus and, absent this check, my own priors - is stale. - Fannie B3-3.4-06: divide by the loan's amortization term. Net Documented Assets ÷ 360 for a 30-year. Net = eligible assets minus any early-distribution penalty that would apply on a complete distribution, minus funds allocated to down payment, closing costs, and required reserves. Max LTV/CLTV 70%, or 80% if the asset owner is at least 62 at closing. Purchase and limited cash-out only; principal residence and second home only. No income-continuance documentation required. The old "70% of stocks/bonds/mutual funds" haircut is not in the current text - it has been replaced by penalty subtraction, which for a 59.5+ borrower is zero. Retirement accounts qualify here only if a distribution is not already set up, or the existing distribution is not enough to qualify.
- Freddie §5307.1 has two live versions and they differ sharply. Current (effective 08/05/2026): net eligible assets ÷ 240, LTV/TLTV ≤80%, primary or second home, no minimum asset threshold, no percentage haircut anywhere, 62+ required only for depository accounts and securities (not retirement), and the asset must not already be used as a source of income. Crypto excluded. Under Bulletin 2026-10 (issued 08/05/2026, mandatory 02/03/2027, early adoption permitted today) the section is renamed "Accumulated assets as income" and the divisor drops 240 → 180, a $30,000 minimum is added, the 80% LTV cap and the 62+ age gate are removed, investment properties become eligible, reserves become a subtraction, and new 12-month seasoning plus a ±20% balance-volatility test apply. Two Freddie lenders can legitimately quote materially different qualifying income in the same week right now.
- Social Security gross-up is smaller than folklore says. B3-3.4-15 gives a safe harbor of 15% of the benefit treated as nontaxable with no documentation, grossed up at the B3-3.1-01 rate of 25% - a net multiplier of 1.0375x, not 1.25x. Grossing up more than 15% of the benefit requires documenting the actual nontaxable portion. Continuance need not be verified for retirement benefits on the borrower's own work record. For annuity/pension/IRA distribution income (B3-3.4-03), 3-year continuance must be documented, and eligible 401(k)/IRA/Keogh balances may be combined to satisfy it; fixed distributions need no history, variable distributions need 12 months and are averaged. New in the 2026 text: where an asset account is the sole or majority income source, the lender must affirmatively assess repayment ability after the asset depletes.
- Non-QM asset-utilization is a different animal: much shorter divisors, real gates. A&D Mortgage: ÷60, 100% of cash and public securities, 80% of retirement, 3-month seasoning, 3-month reserves, DTI to 55%, up to $4M, 80% CLTV. LoanStream (OCMBC): ÷60, 3-month seasoning, 80% LTV purchase, explicitly blendable with other income. Angel Oak: ÷84, 70% securities, 70% non-401(k) retirement, 50% of 401(k), 700 FICO, and a hard $500,000 minimum remaining liquid assets after down payment, closing, and reserves - plus no other income allowed, which disqualifies it for anyone leaning on Social Security. None of the three program sheets imposes a 59.5 or 62 age minimum; they substitute haircuts instead. Rate premium over agency is directionally 50-150bps but I did not retrieve a rate sheet - treat as unverified.
- HECM for Purchase (H4P) is the option the vault has never considered. FHA-insured, borrower 62+, buys a primary residence with a large down payment and no monthly principal-and-interest obligation for life. 2026 maximum claim amount $1,249,125 (up from $1,209,750). Required down payment runs roughly 55-60% at age 62 and 47-52% at age 70, driven by the youngest borrower's age and the expected rate. Financial assessment reviews credit, residual income, and ability to keep paying taxes and insurance - no minimum income and no minimum credit score, and a LESA (set-aside) can be imposed instead of a denial.
Convergences and contradictions
- Convergence: every agency divisor is, read plainly, a stated annual drawdown rate. Fannie ÷360 = 3.33%/yr. Freddie ÷240 = 5.0%/yr. Freddie's incoming ÷180 = 6.67%/yr. Non-QM ÷84 at a 70% haircut = 10%/yr effective. Non-QM ÷60 at 80% = 16%/yr effective. The agencies and the non-QM shops do not disagree about the arithmetic; they disagree about how fast it is acceptable to burn a retiree's portfolio. That reframes "which program qualifies them" into "at what drawdown rate am I willing to underwrite my own parents."
- Contradiction with [[milestones]]: the Q1 2027 gate treats "pre-approved on retirement income alone" and "founder co-signs" as equivalent pass conditions. They differ by roughly $400-600k of the founder's own 2027 borrowing capacity (quantified below). The gate needs to be split into two criteria with different consequences.
- Contradiction with the $150k sizing rationale in [[milestones]]: "typical FL retiree mortgage qualification ceiling" is doing load-bearing work with no source behind it. Under the placeholder assumptions here, Social Security alone supports roughly a $186-230k mortgage, not $450k.
Assumptions (all placeholders - none of these come from the vault)
| # | Assumption | Value used | Risk if wrong |
|---|---|---|---|
| A1 | Both parents are 62+ at closing | Yes | If under 62: Fannie's 80% LTV carve-out and H4P both disappear |
| A2 | Both are 59.5+ (no early-distribution penalty) | Yes | If under 59.5, Fannie subtracts a 10% penalty from net documented assets |
| A3 | Combined gross Social Security | $4,300/mo | Linear driver of everything below |
| A4 | Pension / annuity income | $0 | Any pension is the single cheapest fix - it is 1:1 qualifying income |
| A5 | Non-housing monthly debts (car, card minimums) | $0 | Every $500/mo of car payment costs ~$75k of their loan capacity |
| A6 | Retirement / investment assets | Unknown - solved for | This is the output variable, not an input |
| A7 | Rate: 30-yr fixed conventional | 6.75% (sensitivity at 7.25%) | +50bps costs ~7% of loan capacity |
| A8 | Florida property: tax 1.2% of price; HOI $4,200/yr at $450k, $5,400/yr at $600k; no HOA, no flood | - | An HOA or a Zone AE flood premium lands directly in PITI and directly in DTI |
| A9 | Gift is $150k per [[readme]]; parents' own assets are NOT consumed by the down payment | - | If the parents fund any of the down payment, that money leaves the depletion base |
The arithmetic
Step 1 - PITI at each price (gift = $150k).
| Price | Loan | LTV | P&I @6.75% | Tax | Ins | PITI | PITI @7.25% |
|---|---|---|---|---|---|---|---|
| $450,000 | $300,000 | 66.7% | $1,946 | $450 | $350 | $2,746 | $2,847 |
| $600,000 | $450,000 | 75.0% | $2,919 | $600 | $450 | $3,969 | $4,120 |
Note the LTV: at $600k with a $150k gift the loan is 75% LTV, which breaks Fannie B3-3.4-06's 70% base cap. That deal only exists under the 62+ carve-out to 80%, or under Freddie's 80% (current) - so assumption A1 is not a detail, it is a gate.
Step 2 - qualifying income required.
| Price | @45% DTI | @50% DTI | @36% DTI (conservative) |
|---|---|---|---|
| $450,000 | $6,102/mo | $5,492/mo | $7,628/mo |
| $600,000 | $8,820/mo | $7,938/mo | $11,025/mo |
Social Security grossed at the safe harbor: $4,300 x 1.0375 = $4,461/mo. (If they can document that essentially all of the benefit is nontaxable - plausible if SS is nearly their whole income - the full 25% gross-up gives $5,375/mo. That documentation is worth ~$914/mo, which is ~$137k of loan capacity. Cheapest single lever on the board.)
Step 3 - the asset gap, at 45% DTI. $450k needs $1,641/mo from assets. $600k needs $4,359/mo.
Monthly qualifying income per $100,000 of net eligible assets, and the assets required to close each gap:
| Program | Divisor | Per $100k | Implied annual burn | Assets needed: $450k deal | Assets needed: $600k deal |
|---|---|---|---|---|---|
| Fannie B3-3.4-06 (30-yr) | ÷360 | $278 | 3.3% | $591k | $1.57M |
| Freddie §5307.1 current | ÷240 | $417 | 5.0% | $394k | $1.05M |
| Freddie post-Bulletin 2026-10 (adoptable now) | ÷180 | $556 | 6.7% | $295k | $785k |
| Non-QM (A&D / LoanStream), retirement @80% | ÷60 | $1,333 | 16.0% | $123k | $327k |
| Non-QM, taxable brokerage @100% | ÷60 | $1,667 | 20.0% | $98k | $262k |
| Angel Oak asset depletion | ÷84 @70% | $833 | 10.0% | n/a - no other income permitted | n/a |
Angel Oak is out of scope for this profile on two counts: it forbids blending other income (so the Social Security disappears), and it requires $500k of liquid assets remaining after closing - a household with $500k left over is not the household that needs a $150k gift.
Step 4 - if the assets are not there, what does the gift have to be? Social Security only, no assets counted, no pension:
| Price | Max loan @45% DTI | Gift required | Max loan @50% DTI | Gift required |
|---|---|---|---|---|
| $450,000 | ~$186k | ~$276k (incl. ~$12k closing) | ~$220k | ~$242k |
| $600,000 | ~$148k | ~$470k (incl. ~$18k closing) | ~$182k | ~$436k |
The perverse result at $600k is real and worth naming: taxes and insurance scale with price, so a more expensive house eats the fixed Social Security DTI budget from both ends and the supportable loan actually shrinks. On these placeholders the parents can carry a bigger mortgage on a $450k house than on a $600k one.
Step 5 - H4P, for comparison. No income qualification beyond the financial assessment, no monthly P&I ever, non-recourse:
| Price | Down @ age 62 (55-60%) | Down @ age 70 (47-52%) |
|---|---|---|
| $450,000 | $248k - $270k | $212k - $234k |
| $600,000 | $330k - $360k | $282k - $312k |
At $600k, H4P at age 70 costs the founder $282-312k against $436-470k for the all-cash-gift conventional path - roughly $130-190k less, with the monthly payment risk removed entirely. The cost is the parents' terminal home equity, which is the founder's eventual inheritance. That is a real trade, not a free lunch, but it has never been on the table in [[readme]] and it should be.
Step 6 - what a co-sign costs the founder. At 7% on a 30-year, $1,000/mo of P&I buys $150,300 of loan. A co-signed mortgage puts the full PITI on his back-end ratio:
| Scenario | Parents' PITI | Founder's borrowing capacity displaced | As % of his $1.0-1.2M ceiling |
|---|---|---|---|
| $450k house | $2,746/mo | ~$413k | 34-41% |
| $600k house | $3,969/mo | ~$597k | 50-60% |
Compare the two levers directly at the $600k price: gifting the extra ~$286-320k (from $150k up to the ~$436-470k the cash-only path needs) costs him cash he arguably does not have spare after the build. Co-signing costs him ~$597k of build financing against a ceiling the sibling brief already shows closing with near-zero margin. On the vault's own numbers, co-signing is the more expensive option, and it is expensive in exactly the currency the build is short of.
Step 7 - can the co-signed debt be excluded later? Fannie B3-6-05, Monthly Debt Obligations (08/05/2026) - the "Contingent Liabilities" heading no longer exists; the material now sits under "Debts Paid by Others." Full PITIA may be excluded if (1) the party making the payments is themselves obligated on the mortgage debt, (2) no delinquencies in the most recent 12 months, and (3) the borrower is not using rental income from that property - and the lender must obtain 12 months of canceled checks or bank statements from the paying party. The parents are on the note, so condition (1) is satisfied - this is the one arrangement where the exclusion is actually available. But [[milestones]] has them closing 2027-04-30 and the construction loan closing 2027-06-15: six weeks of history against a twelve-month requirement. The exclusion is unavailable on the current timeline. To use it, the parents would have to close by roughly May 2027 minus twelve months (already past) or his construction application slips to mid-2028. Two further caveats: the property still counts against his financed-property limit under B2-2-03 regardless, and his build loan is jumbo portfolio, not agency - per [[2026-08-14-construction-to-permanent-loans-florida]] the GSE products are out of range entirely, so B3-6-05 is persuasive doctrine his actual underwriter is free to ignore.
Synthesis for RDCO
The $150k gift number does not survive contact with the underwriting math, and the vault should stop treating it as settled. It was sized in April 2026 to keep the founder's lifetime-exemption use modest, justified by an unsourced "typical FL retiree qualification ceiling," and it has been carried forward through [[milestones]], [[index]], and the Q1 2027 decision gate without ever being tested against a formula. Tested, it works only if the parents hold serious investable assets: $591k under Fannie, $394k under current Freddie, $295k under Freddie's incoming rules for the $450k house, and $1.57M / $1.05M / $785k respectively for the $600k house. A household sitting on $1M in an IRA does not need a $150k gift to buy a $600k house. The gift and the asset-depletion path are, to a first approximation, mutually exclusive answers to the same problem. The honest read is that if the parents needed help enough to make this a project, the assets probably are not there, and the operative column is Step 4: the cash-only gift is ~$242-276k at $450k and ~$436-470k at $600k, not $150k. That is a $90-320k hole in the plan's cash line, sitting on top of a build whose sources and uses already close with essentially no margin.
The divisor is the real finding, and it should change how the founder reads any lender's "yes." Fannie's ÷360 underwrites a 3.3%/yr drawdown - roughly the classic safe-withdrawal number, which is why it qualifies almost nobody. Non-QM's ÷60 underwrites a 16-20%/yr drawdown. A non-QM asset-utilization approval is not a lender telling him his parents can afford the house; it is a lender agreeing to be repaid out of a portfolio that will be gone in five to six years, at which point the only remaining source of the payment is him. He would be paying a 50-150bp rate premium to convert a visible cash gift today into an invisible contingent obligation in 2032. Every structural fact in this brief points the same direction: the failure mode of manufactured qualifying income is that the founder ends up funding it anyway, later, without having planned for it. Freddie's Bulletin 2026-10 change is the one piece of genuinely good news - the 240 → 180 divisor is a ~39% increase in qualifying income from the same assets, it is adoptable today at a lender's option, and it becomes mandatory 02/03/2027, which is on the right side of the Q2 2027 closing target. Asking each lender "have you adopted Bulletin 2026-10 yet" is a free question with a five-figure answer.
Co-signing is the most expensive option on the board and it should come off the table as a default. [[milestones]] currently passes the Q1 2027 gate on "pre-approved on retirement income alone OR willing for founder to co-sign," which prices those two outcomes as equivalent. They are not within an order of magnitude of each other. A co-signed $600k purchase loads $3,969/mo onto his back-end and vaporizes ~$597k of his own borrowing capacity - half the $1.0-1.2M ceiling that [[2026-08-14-construction-to-permanent-loans-florida]] shows is already the binding constraint on the build. The B3-6-05 escape hatch exists and the co-borrower structure genuinely satisfies its hardest condition, but the twelve-month seasoning requirement collides head-on with a six-week gap between the two closings, and his jumbo portfolio underwriter is not bound by Fannie doctrine anyway. If co-signing is going to happen at all, the sequencing has to invert: parents close a full year before his construction application, not six weeks before. That is a calendar decision available today at zero cost, and it is the single highest-leverage change to the plan.
Two options the vault has never priced deserve a look before the Q1 2027 gate. First, H4P: if both parents are 62+, a reverse purchase eliminates the monthly payment entirely, requires no income qualification beyond a financial assessment with no minimum income or FICO, and at age 70 needs $282-312k down on a $600k house versus $436-470k for the cash-only conventional path. It is cheaper in cash and removes the payment-default risk the founder is implicitly backstopping. What it costs is the parents' terminal equity, i.e. his own inheritance - a trade he may well take, but one he should make explicitly rather than by omission. Second, an intra-family purchase-money note at the applicable federal rate: the founder holds the paper instead of a bank, no co-sign exists to load his DTI, the interest is taxable to him rather than lost to a lender, and the note is an asset on his balance sheet rather than a gift against his lifetime exemption. Both of these are worth a serious look precisely because the conventional path's real answer - "write a check for $276k or $470k" - is a worse deal than either. The blocking input is unchanged and now seven weeks overdue: the parents' actual ages, Social Security amounts, any pension, non-housing debts, and account balances. Every number in this brief is a placeholder until that conversation happens, and it is a thirty-minute conversation gating a six-figure decision.
Why this is in the vault
This resolves Critical Open Question #1 in [[readme]] in structure if not in dollars, and it forces two concrete edits to [[milestones]]: the Q1 2027 gate criterion "pre-approved on retirement income alone OR willing for founder to co-sign" must be split into two criteria with different consequences (~$597k of build capacity apart), and the 2027-04-30 parents' closing / 2027-06-15 construction-loan closing sequence must either invert to a twelve-month gap or the co-sign branch must be abandoned. It also puts a documented range under the $150k gift line in the project's cash-outlay table, which feeds the [[index]] savings target.
Open follow-ups
- Do Florida jumbo portfolio construction lenders (Seacoast, TD, Goldwater, and the national OTC jumbo programs) honor Fannie's B3-6-05 "Debts Paid by Others" exclusion for a co-signed mortgage, or do portfolio overlays count the full contingent PITI regardless of payment history? This determines whether the twelve-month sequencing fix has any value at all for his specific lender class, and no existing brief addresses portfolio overlay treatment of contingent liabilities.
- What is the current long-term applicable federal rate, and what does a Florida-recorded intra-family purchase-money note require to be respected as bona fide debt rather than a disguised gift? This is the one option that avoids both the DTI hit and the outright cash outlay, and it is entirely unresearched in the vault.
Related
- [[readme]] - Home Rebuild 2027 + Parents' Retirement Home, the parent project and source of Critical Open Question #1
- [[milestones]] - the Q1 2027 decision gate and the closing sequence this brief argues must change
- [[2026-08-14-construction-to-permanent-loans-florida]] - sibling lane; the founder's own jumbo build envelope and the $1.0-1.2M income ceiling the co-sign would eat into
- [[2026-06-04-home-affordability-build-vs-buy]] - the "don't fund the gift and the full-stretch build in the same window" constraint
- [[index]] - Finance Pulse longitudinal index, where "Parents' mortgage qualification clear" has been Unknown since June 30
- [[2026-08-04-flood-zone-ae-construction-loan-requirements]] - Florida insurance and flood cost context that flows into the PITI assumptions here
Sources
Vault
~/rdco-vault/01-projects/home-rebuild-2027/README.md~/rdco-vault/01-projects/home-rebuild-2027/milestones.md~/rdco-vault/04-finance/index.md~/rdco-vault/04-finance/2026-06-04-home-affordability-build-vs-buy.md~/rdco-vault/06-reference/research/2026-08-14-construction-to-permanent-loans-florida.md
Agency guides (primary)
- Fannie Mae Selling Guide B3-3.4-06, Employment-Related Assets as Qualifying Income (03/04/2026) - https://selling-guide.fanniemae.com/sel/b3-3.4-06/employment-related-assets-qualifying-income
- Fannie Mae Selling Guide B3-3.4-03, Annuity, Pension, or Retirement Income (03/04/2026)
- Fannie Mae Selling Guide B3-3.4-15, Social Security Income (03/04/2026)
- Fannie Mae Selling Guide B3-3.1-01, General Income Information - nontaxable income gross-up (03/04/2026) - https://selling-guide.fanniemae.com/sel/b3-3.1-01/general-income-information
- Fannie Mae Selling Guide B3-6-05, Monthly Debt Obligations (08/05/2026) - https://selling-guide.fanniemae.com/sel/b3-6-05/monthly-debt-obligations
- Fannie Mae Announcement SEL-2026-02 (Chapter B3-3 restructure, 03/04/2026) - https://singlefamily.fanniemae.com/media/44976/display (landing page and PDF return 403 to automated retrieval; the restructure is verified from the live Guide sections' own date stamps and redirects, not from the announcement text)
- Freddie Mac Single-Family Seller/Servicer Guide §5307.1, Assets as a basis for repayment of obligations (current, eff. 08/05/2026) - https://guide.freddiemac.com/app/guide/section/5307.1 (JS-rendered; plain fetch returns an empty shell)
- Freddie Mac Bulletin 2026-10 (Selling), 08/05/2026 - "Accumulated assets as income," divisor 240 → 180, eff. 02/03/2027 with early adoption permitted - https://guide.freddiemac.com/app/guide/bulletin/2026-10
Lender program sheets (primary for the program, secondary for market norms)
- A&D Mortgage, Asset Utilization - https://admortgage.com/programs/asset-utilization/
- LoanStream Wholesale (OCMBC), Non-QM Asset Utilization - https://loanstreamwholesale.com/non-qm-asset-utilization/
- Angel Oak Mortgage Solutions, Asset Depletion - https://angeloakms.com/programs/asset-depletion-mortgage-program/
HECM for Purchase
- HUD, FHA 2026 loan limits announcement (HECM maximum claim amount $1,249,125) - https://www.hud.gov/news/hud-no-25-145
- HUD HECM Financial Assessment, updated 01/2026 - https://entp.hud.gov/pdf/mp_sfo12_hcm_fin_assessmnt.pdf
- Reverse.mortgage, 2026 HECM limits and purchase down-payment tables (secondary; down-payment percentage bands by age are directional and must be re-priced at the parents' actual ages and the then-current expected rate) - https://reverse.mortgage/purchase
Explicit non-verifications
- No primary source retrieved for the non-QM rate premium over agency (the 50-150bp figure is trade-press directional only).
- Which Fannie announcement removed the former 70% asset haircut could not be verified; only the current text is verified.
- Freddie §5307.1's reserve requirement is not stated in-section and was not chased into the reserves chapter.
- FHA/HUD Handbook 4000.1 forward-mortgage retirement-income rules were not researched; if the parents end up on an FHA forward loan the continuance and gross-up rules differ.