For nine consecutive years, the Department of Housing and Urban Development has known, by its own admission, that its largest rental subsidy program is “susceptible to significant improper payments.” For nine consecutive years, it has declined to say by how much. Not because the number is unknowable. Because in 2017 HUD fired the contractor whose job it was to know, and simply never got around to hiring a replacement.
That is the opening fact of HUD OIG Report 2026-FO-0006, and it is worth sitting with before the number arrives, because the number is the least interesting part. Any auditor can find a bad statistic. What OIG actually documented is an agency that spent nearly a decade in willful, well-funded ignorance of its own $38.4 billion program, then disputed the homework once someone else finally did it.
An Agency That Outsourced Its Own Curiosity
The Public and Indian Housing Tenant-Based Rental Assistance program, known more casually as Section 8 or the Housing Choice Voucher program, is the federal government’s largest direct rental subsidy. It moved $38.4 billion in fiscal 2025, 43 percent of HUD’s entire budget, through roughly 2,200 local Public Housing Agencies that calculate each household’s subsidy using a formula built from income, deductions, rent reasonableness, and utility allowances. It is, by design, decentralized, manual, and dense with judgment calls, which is precisely why the Payment Integrity Information Act requires HUD to estimate how often the formula gets applied wrong.
HUD used to pay a contractor to run that estimate every year. It discontinued the contract in 2017 and has not produced a figure since, citing “decentralized payment processes, dated and disparate systems” and a familiar chorus of complexity. OIG did not accept complexity as an excuse; it built the estimate HUD wouldn’t. Auditors pulled a stratified sample of 300 payments from the 2.3 million monthly housing assistance payments on the books in March 2025 and checked each one against the underlying documentation.
Seventy-four were wrong. That is 25 percent, one in four, extrapolated to a projected $1.43 billion a year in improper payments, a number OIG discounted downward for statistical caution before publishing it, meaning the real figure is probably higher, not lower.
Where the Money Actually Leaks
The errors sort into a fairly mundane taxonomy, which is itself the point; nothing here required fraud, only inattention. Income miscalculations accounted for the largest share (PHAs entering figures that didn’t match pay stubs, relying on stale documentation, or simply skipping verification against HUD’s own income-matching system). Utility allowance errors came next, PHAs applying the wrong schedule or the wrong unit size to a calculation most of them still do by hand. Then there is a category OIG calls “program compliance,” meaning the PHA skipped a required step entirely, no annual reexamination, no rent reasonableness analysis, no verified immigration status, and got the whole payment marked improper as a result, not because the dollar amount was necessarily wrong but because nobody could prove it wasn’t.
That distinction matters more than it sounds. OIG also logged 13 “technically improper” payments, cases where the PHA broke a rule but the dollar amount happened to land correctly anyway, pure luck rather than diligence. The audit treats those as a warning sign rather than a reprieve; the same shortcuts that produced a correct number this month are exactly the shortcuts that will produce a wrong one next month.
Then there is HOTMA, the Housing Opportunity Through Modernization Act, signed in 2016 with instructions to raise the medical and disability expense deduction threshold and the elderly and disabled household allowance. Ten years later, HUD’s own system still runs the old, lower thresholds, because the software meant to replace HUD’s aging PIC system, itself a replacement project HUD started in 2021, still has no launch date. A statute a decade old, unenforced, because the agency’s internal IT timeline has quietly outlived the urgency that produced the law in the first place.
The Monitoring That Wasn’t
If the errors were merely common, that would be one story. What makes it worse is how little anyone at HUD was watching for them. The agency ran 19 targeted “Calculation of HAP, Adjusted Income, and Rent” reviews in fiscal 2023 and has run zero since. In fiscal 2025, HUD completed 14 total Compliance Monitoring Reviews, covering nine separate program areas, for a portfolio of 2,200 housing agencies and 2.3 million households. That is not oversight so much as its rumor.
PHAs, for their part, were not shy about explaining why, once OIG actually asked. Understaffed offices process tens of thousands of annual reexaminations with no time to verify small, fluctuating income sources like child support or unemployment benefits. HUD’s own income-verification tool, the Enterprise Income Verification system, updates quarterly and simply doesn’t capture pensions, TANF, or state benefit income at all, meaning the tool HUD requires PHAs to use cannot see a meaningful share of the income it’s supposed to catch. Guidebooks on calculating rent and running reexaminations haven’t been updated since 2019 and 2020 respectively, years before several of the rule changes they’re supposed to explain.
The Part Where HUD Argues With the Math
Faced with all this, the Office of Public and Indian Housing did what agencies under audit reliably do, agreed with the spirit and quibbled with the specifics. PIH accepted the recommendation to work on simplifying HAP calculations and to develop better monitoring. It rejected the recommendation to embed a mandatory documentation checklist into its own system, for the reasonable-sounding but conveniently timed reason that its legacy IT platform can’t support it. And it declined to concur with the $1.43 billion figure itself, saying it was still waiting on additional information from OIG.
OIG’s rebuttal, printed in the report without much editorializing needed, lays out a timeline; final results shared with PIH in March 2026, three follow-up meetings in April to walk through methodology sample by sample, and no further request for information from PIH until July, three months after the walkthroughs concluded. The agency that spent nine years unable to produce its own improper payment estimate spent another four months sitting on someone else’s before deciding it needed more homework.
PIH’s written response includes one more line worth quoting in spirit if not verbatim, namely that the risk of error is “largely outside of PIH’s direct control,” a structural feature of a complex, decentralized statute rather than a management failure. That may even be half true. It is also the sentence every agency reaches for once the number gets big enough to need one, and it is a strange thing to say about a program HUD chose, for nine years running, not to measure at all.
Source: HUD Office of Inspector General, Report 2026-FO-0006, “HUD’s Public and Indian Housing Tenant-Based Rental Assistance Program Had Significant Improper Payments,” September 17, 2026.
