BlackLeafwatch the watchmen
HUD's HECM reverse-mortgage safety-net program (Life Expectancy Set-Asides)

Reverse-Mortgage Safety Nets Are Running Out Years Early

Summary

A HUD Inspector General audit found that federally insured reverse-mortgage borrowers are draining the government's own safety-net accounts years earlier than the agency calculated, threatening up to $258 million in defaults -- a number HUD's own housing office disputes, citing internal loss-rate data less than half the auditor's.

By Marcus Aurelius · July 16, 2026

HUD's Office of Inspector General -- 's independent statutory watchdog, distinct from the housing office it audits -- found that an estimated 1,237 borrowers in the Home Equity Conversion Mortgage (HECM) program, the federally insured reverse mortgage that lets homeowners 62 and older convert home equity into cash, are watching their built-in safety net run dry years before the government's own math said it would. That safety net is the Life Expectancy Set-Aside, or LESA: for borrowers 's own screening flagged as financially vulnerable, the lender sets aside part of the loan to pay property taxes and insurance directly, for as long as the borrower is expected to live. When it runs out, the borrower has 30 days' notice before those bills become their own -- and missing them triggers default on the loan that holds their home.

Auditors opened 80 loan files. 72 were already in trouble.

didn't audit the whole HECM portfolio. It built a targeted universe: 1,462 active loans with fully funded LESAs, closed between 2018 and 2022, worth $357 million in combined maximum claim amount, that already showed signs of running down faster than projected. From that universe, auditors pulled a statistically representative sample of 80 loans and opened the files -- payment histories, tax and insurance records, the original LESA calculation itself. Seventy-two of the 80 -- 90% -- had LESAs that were either already fully depleted or on track to run out an average of six years before 's own formula estimated. Projected across the full universe, that rate implies 1,237 borrowers headed for the same cliff.

Projected loss to HUD
$258M
OIG's estimate if all 1,237 borrowers default -- HUD's own Single Family office disputes the loss-rate assumption behind it
Borrowers facing early payments
1,237
Estimated HECM borrowers whose safety-net fund will deplete significantly sooner than HUD calculated
Sample deficiency rate
90%
72 of 80 sampled loans had LESAs already depleted or on track to deplete an average of 6 years early
What HUD OIG found when it opened 80 HECM loan files
Results of the statistically representative sample of reverse-mortgage loans with fully funded Life Expectancy Set-Asides
LESA already fully depleted
25
On track to deplete early
47
Not deficient
8
Source: HUD OIG Report 2026-KC-0003, Figure 2 (p.3-4)
View data as table
Of the 80 sampled HECM loans with fully funded LESAs, 25 had already fully depleted and 47 more were on track to deplete years earlier than HUD projected -- 72 of 80 (90%) in total. The remaining 8 were not deficient: 2 on track for the full life expectancy, 5 paid in full, and 1 created in error with no distributions.
LESA already fully depleted25
On track to deplete early47Accelerated depletion -- not yet exhausted but running years ahead of HUD's projection
Not deficient82 on track for full life expectancy, 5 paid in full, 1 created in error

HUD built in a 20% cushion. Some borrowers needed 192%.

's LESA formula isn't naive about rising costs -- it already applies a 120 percent multiplier to a borrower's current property tax and insurance bill, banking a 20% cushion against future increases. That cushion wasn't close to enough for a third of the deficient loans: 31 of the 72 borrowers with problem LESAs saw their property charges rise by an average of 192%, concentrated in California, Florida, Colorado, Texas, and Arizona -- states where property values, and the taxes and insurance premiums tied to them, climbed hardest in exactly this window. One California borrower's combined tax-and-insurance bill went from $2,103 in 2021 to $12,262 by 2024 -- a 483% increase, more than 24 times the buffer 's formula assumed.

When the deadline arrives
Of the 47 sampled borrowers with accelerated LESA depletion, the year each would need to start paying property charges out of pocket
2026
15
2027
9
2028 and later
23
Source: HUD OIG Report 2026-KC-0003, p.4
View data as table
Of the 47 sampled borrowers whose LESA is depleting on an accelerated schedule, 15 needed to start paying property charges out of pocket in 2026 alone, 9 more in 2027, and 23 in 2028 or later -- every one of them years before HUD's own formula said they would need to.
202615
20279
2028 and later23

The formula behind the shortfall was never rechecked

The deeper problem, per , is that nobody at was watching. HUD's LESA formula calculates a borrower's life expectancy using the U.S. Decennial Life Table for 1979-1981 -- published in August 1985 by the Department of Health and Human Services, and never swapped out for a newer table even as has updated the underlying tables several times since. periodically checks HECMs that have already defaulted or foreclosed for portfolio-wide risk patterns, found, but never built a process to check active LESAs against their own depletion timeline -- there simply was no requirement to look. Auditors separately flagged 27 of the 80 sampled loans for a rounding step in the age calculation that 's own Handbook 4000.1 governs -- a finding 's housing office disputes as reasonable practice rather than an error (below).

HUD agrees to look. HUD disputes the price tag.

's single recommendation is procedural: have the Deputy Assistant Secretary for Single Family Housing periodically evaluate whether the LESA formula still works and whether active accounts are depleting early. In its written reply -- reprinted in full in Appendix B -- Single Family Housing agreed to do that evaluation, without attaching a target date.

It's the dollar figure the two sides can't agree on. converted depleted LESAs into a dollar loss using a 72.3% loss rate, pulled from 's own Single Family Acquired Management System records covering 1,144 properties. Single Family's Office of Risk Management countered with its own numbers -- a 33.2% loss rate for foreclosed properties later resells, and 26.9% for note sales -- both less than half of 's figure, and both grounds for calling the $258 million estimate overstated. 's rebuttal, also in the report: the Risk Management figures leave out foreclosure and holding costs -- acquisition, repairs, taxes, maintenance, sales expenses -- that its own SAMS data captured.

Two HUD offices, two numbers for the same risk
The loss rate each side applied to convert depleted LESAs into a dollar estimate -- the entire gap between OIG's $258 million and a lower HUD-cited figure
OIG (SAMS data, 1,144 properties)
72.3%
HUD Risk Mgmt. -- REO properties
33.2%
HUD Risk Mgmt. -- note sales
26.9%
Source: HUD OIG Report 2026-KC-0003, pp.6-9, Appendix B (p.12)
View data as table
OIG applied a 72.3% loss rate, drawn from HUD's own Single Family Acquired Management System, to translate depleted LESAs into a dollar loss. HUD's Office of Single Family Housing disputed that rate, citing its own Office of Risk Management data: 33.2% for real-estate-owned properties and 26.9% for note sales -- both less than half the OIG figure.
OIG (SAMS data, 1,144 properties)72.3%
HUD Risk Mgmt. -- REO properties33.2%
HUD Risk Mgmt. -- note sales26.9%

Run 's $258 million through Single Family's own disputed numbers instead of its own, and the range moves a lot: applying the 33.2% rate scales the estimate to roughly $118 million; the 26.9% rate brings it to roughly $96 million -- as low as 37% of 's figure, using inputs 's own risk office supplied. Neither office published a recalculation using the other's rate; this is the arithmetic each side's own cited number implies. Single Family went further than disputing the rate: it cited data as of February 28, 2026 showing that 90% of non-terminated HECM borrowers with an already-depleted LESA balance had not reported a tax-and-insurance default -- its evidence that a drained LESA doesn't automatically mean a lost loan, whatever the eventual dollar exposure turns out to be.

The audited loans are a fraction of what's exposed

's 1,462-loan universe was deliberately narrow -- loans already flagged as running more than 10% ahead of 's projected depletion schedule. HUD's current LESA portfolio holds 41,002 HECMs -- about 28 times the size of the group sampled from -- and every one of those loans was built on the same 1985-vintage life table and the same 20% cushion that failed a third of the deficient loans in this sample. 's own framing: as property taxes and insurance keep climbing, more of that larger pool will hit the same wall, on the same formula nobody's required to recheck. Nothing in 's response commits to fixing the formula itself -- only to evaluating whether it needs fixing, with no completion date yet on record.

  • 1,237 HECM borrowers are projected to run out of their tax-and-insurance safety net years earlier than calculated, opening as much as $258 million in exposure for -- a figure 's own Office of Risk Management disputes as overstated.
  • The formula behind the shortfall dates to 1985 and was never rechecked. 's LESA life-expectancy math still runs on the 1979-81 Decennial Life Table, and there was no requirement to periodically check whether active LESAs were tracking their own depletion schedule -- only defaulted and foreclosed loans got that scrutiny.
  • agreed to evaluate the formula. It did not agree on the price of not fixing it sooner. Applying 's own disputed loss rates (33.2% and 26.9%) instead of 's (72.3%) to the same exposure moves the estimate from $258 million down to roughly $96-118 million -- the entire swing traces to which office's loss-rate assumption you use.

The $258 million figure is 's statistically projected "funds to be put to better use" estimate -- a lower-confidence-limit projection from the 80-loan sample to the 1,462-loan universe (Appendix E), not a count of loans that have already defaulted; as of the audit, 25 sampled LESAs had fully depleted and none of the 80 had yet defaulted outright. The $96 million and $118 million comparison figures in this piece are this article's own illustrative rescaling of 's total using 's cited loss rates in place of 's -- linear arithmetic shown to bound the dispute, not an alternate official estimate from either office. hudoig.gov's own report page and PDF return HTTP 403 to direct fetches; this piece sources the report through Oversight.gov, the federal cross- portal that mirrors the identical document.

Sources(2) ▾
  • U.S. Department of Housing and Urban Development, Office of Inspector General (via Oversight.gov), HUD Underestimated Rising Property Charges for an Estimated 1,237 HECM Borrowers -- Report Summary Page (2026-KC-0003) (2026-05-05) 's own audit, report number 2026-KC-0003, published May 5, 2026 by the Office of Audit and Evaluation. hudoig.gov's own report page returns HTTP 403 to direct fetches, so this article uses Oversight.gov -- the Council of the Inspectors General on Integrity and Efficiency's federal mirror, which hosts the identical report and links the same accessibility-remediated PDF -- as the capture. Used here for the report's title, number, publish date, and top-line framing (What We Audited, What We Found, What We Recommend). oversight.gov · original document
  • U.S. Department of Housing and Urban Development, Office of Inspector General, HUD Underestimated Rising Property Charges for an Estimated 1,237 HECM Borrowers -- Full Report (2026-KC-0003) (2026-05-05)The full 21-page audit report (Section 508-remediated PDF), linked from the Oversight.gov summary page above, fetched directly and converted to text with pdftotext -layout. Sourced from it: the Results of Audit section (pp.3-6) on LESA depletion and the $258 million projected loss; the Scope and Methodology (pp.8-9) on the 1,462-loan universe and 80-loan statistical sample; Appendix A's Schedule of Funds to Be Put to Better Use (p.10); Appendix B's Management Response, including Single Family Housing's dispute of the loss-rate methodology (pp.11-12); Appendix D's per-loan property-charge-increase detail (pp.16-18); and Appendix E's sample-projection methodology (pp.19-21). One Internet Archive Save Page Now request for this PDF was attempted and did not return a snapshot in time (no prior Wayback capture exists either, per the Wayback availability API) -- consistent with the archive-availability gaps noted elsewhere in current collection; the capture link points at Oversight.gov's own permanent government-hosted copy. oversight.gov · original document
Weekly digest: the most-read systems, in brief. Mondays.

Comments

Always open. Logged-in readers can annotate paragraphs in place.

Loading comments…
or log in to comment under your account