Don’t Incinerate HUD’s Reforms to Homelessness Assistance. Strike Provisions that Maintain Non-Performance.
Congress is about to undo the most serious attempt in a decade to fix federal homelessness policy.
Congress is about to undo the most serious attempt in a decade to fix federal homelessness policy. It is doing so in three short provisions buried in the Transportation-HUD Continuing Resolution.
Since 2005, Washington has spent $36.6 billion through HUD’s Continuum of Care (CoC) Program. In 2005, the national homelessness count was 763,010. In 2024, it was 771,480, the highest on record. This year, HUD finally changed course. Its FY 2026 competition called Housing First “a profound failure,” opened 40 percent of renewal funding to merit-based competition, and set aside $1.3 billion for transitional housing and services tied to treatment and work.
The provider lobby unsuccessfully sued to stop those reforms. Now it is asking appropriators to finish the job. Republicans should strike these provisions before the bill moves.
Renewals regardless of results
Page 192 directs HUD to adjust renewals so projects “operate at substantially the same levels,” with cost-of-living increases for services. In plain terms, every incumbent gets last year’s grant plus inflation, regardless of results.
That is how the CoC Program has worked for years, and it is why it failed. HUD found that in FY 2024 roughly 90 percent of each CoC’s renewal demand was awarded “without any connection to CoC score or project merit.” Cicero’s analysis of 1,458 CoC-funded nonprofits shows what that bought. From 2013 to 2024, their compensation grew 117.8 percent and revenue 123.9 percent, against 34.7 percent inflation. The homeless count rose 30.7 percent. Only 6 percent of people in CoC housing increased their earned income.
The outcomes are not neutral, either. HUD reports nearly twice as many permanent supportive housing exits due to death as to unsubsidized housing (forthcoming meta-analysis). In Los Angeles, 29 percent of recorded exits from one major supportive housing program were deaths. Guaranteeing these projects their funding is not compassion. It is a promise to keep paying for failure.
A floor that protects the worst performers
On the same page, a provision guarantees every state at least 95 percent of last year’s CoC total. A competition in which no state can lose more than 5 percent is not a competition. It locks in the existing map of awards, the one built under Housing First, and protects them from competition and accountability.
The biggest beneficiaries are the systems with the worst records. Los Angeles has received more than $2 billion in CoC awards since 2005. A court-ordered assessment of about $2.3 billion in the city’s homelessness programs could not trace substantial funds, and Los Angeles still has more people living on the street than any other American city. Meanwhile, states like Florida, Tennessee, and Utah that are pursuing treatment and accountability gain little from competing, because the dollars cannot move to them.
The CoC structure already shields incumbents: a new provider must ask its own competitors for permission to apply. A statewide hold-harmless provision adds a federal guarantee on top that pre-empts innovation, accountability, and better outcomes.
Freezing the data that should expose them
Section 244 bars HUD from making “methodological changes” to the Annual Homeless Assessment Report, the Point-in-Time count, or Homeless Management Information System (HMIS) data standards. These are the systems Congress relies on to judge whether the money works, and they are badly broken. This section pre-emptively denies reform opportunities.
The GAO found in 2020 that HUD’s quality checks of Point-in-Time data are limited. Portland’s auditor found housing placements overstated by more than 20 percent. California’s statewide system held test records named “Mickey Mouse” and a shelter reporting 1,100 enrollments for 300 beds. In Los Angeles, a court-appointed monitor found self-reported numbers never verified on site, counts never deduplicated, and major programs tracked in Google Forms. Site visits found 30 to 40 percent vacancy in buildings reported as substantially full.
Cicero has urged HUD to require projects to report participant deaths by site and to stop letting deaths count as “planned exits,” as Denver’s social impact bond did. Section 244 would block both. It protects bad data because bad data protects bad programs.
What Congress should do
These provisions are short, but their effect is not. Together they turn a competitive grant into an entitlement, freeze the money where it has failed, and blind the agency charged with oversight. HUD’s efforts to increase accountability, oversight, and outcomes will be trashed.
Republican appropriators should strike the renewal-level proviso (page 192, lines 6–11), the 95 percent state floor (page 192, lines 18–23), and Section 244 (page 251). Striking them cuts no homelessness funding. It keeps HUD’s power to pay for results: people leaving the street for treatment, work, and housing they can keep.
Taxpayers have spent two decades and tens of billions of dollars on the theory that more money on the same terms would end homelessness. It has not. Congress should not write that theory back into law.

Stay Informed
Sign up to receive updates about our fight for policies at the state level that restore liberty through transparency and accountability in American governance.