The FY26 NOFO Is Vacated. That Is Not the Same as Winning.

By Common Ladder · August 3, 2026 · updated August 14 · 20 min read

Common Ladder — Learn / Evidence. Practical guidance for Continuum of Care executives and boards. Current as of August 14, 2026 — rebuilt after the August 7 order in State of Washington v. HUD and National Alliance to End Homelessness v. HUD setting aside the FY2026 CoC NOFO in its entirety, and Judge Carter’s August 13 partial preliminary injunction in LAHSA v. HUD. The August 26 deadline is no longer in force. HUD cannot accept applications.

A Continuum of Care director in a mid-sized county spent July assembling an application she now cannot submit. Project applications went into e-snaps on the evening of July 21. The CoC Application itself never arrived. On August 7 a federal court set the whole notice aside, and HUD's page went from a countdown to a notice that it is "unable to accept applications at this time."

The instinct is to file the binder and wait. That is the wrong instinct, and the reason is in the opinion. The court did not rule that HUD may not do this. It ruled that HUD did not do it correctly. The distance between those two sentences is the entire strategic question for the next six months.

The argument

Treat the vacatur as a procedural reprieve, not a substantive win, and use the pause on the two things a deadline made impossible: securing renewal continuity under the appropriations proviso, and preparing to contest the policy on the record when HUD reopens it through notice and comment. The evidence case for permanent housing has not changed and does not need to be re-argued. What changed is that there is finally time to make it somewhere it will count.

What the court actually held

The two FY2026 challenges — the nonprofit-and-local-government action filed July 2 and the states' action filed July 7 — were decided together on August 7. All parties agreed to forgo interim relief and move straight to summary judgment. The cases are related but were not consolidated.^[6]

The holding is narrow and the language is not. The $1.3 billion set-aside for new transitional-housing and supportive-services-only projects is an "incentive" under 42 U.S.C. § 11386b(d). It falls in the residual category at § 11386b(d)(2)(C), which requires "notice and comment to the public." HUD did not provide it. The court rejected all three of HUD's counter-arguments and called its informal-notice theory "specious," lacking "even marginal support from the text, structure, or history" of the Act. The remedy was categorical: the court "must set aside the 2026 NOFO in its entirety."^[1]

Then the sentences that most of the coverage has skipped. The court denied the request for a permanent injunction. It declined to reach the challenge to HUD's departure from Housing First. It declined to reach the DEI, gender-identity, and immigration certifications. It declined to reach the nonprofit plaintiffs' challenge to the two OMB apportionment footnotes conditioning HUD's access to the funds on compliance with executive orders. And it said plainly that HUD may try again with the same conditions after doing the process properly.^[1]

So: do not tell your board that a court vindicated Housing First. A court held that HUD skipped a required step. That is a real and consequential win — it is the third time in three months that this agency's homelessness funding decisions have been held unlawful — and it buys time. It does not settle what the money will be used for.

What went down with the NOFO. Because the order sets aside HUD's issuance of the 2026 NOFO in its entirety, the July 24 modification and its Appendix III — the § 578.13 direct-to-HUD route that would have bypassed local CoC ranking in a geographic area HUD found non-compliant — went with it. That is the reading the text supports and it is the operative assumption for planning. It is an inference from the scope of the remedy rather than a holding the court stated separately, so confirm it against HUD's next issuance before relying on it in a legal posture. Note also what did not disappear: HUD's remedial-action authority under 42 U.S.C. § 11360a and 24 CFR § 578.13 is statutory and regulatory. It was not struck. The mechanism can return in the next NOFO. Only its current instantiation is gone.^{[1] [7]}

Where the money actually stands

Three separate questions get collapsed into one in most conversations this month. Keep them apart.

Renewals. This is the one that determines whether your projects operate in 2027, and it does not run through the competition at all. Section 244 of the Consolidated Appropriations Act, 2026 directs HUD to renew CoC projects that expired or expire in the first quarter of calendar 2026, with two escalating triggers attached to HUD's own delay: if funding was not awarded by April 1, 2026, projects expiring in Q2 are renewed; if funding was not awarded by July 1, 2026, all remaining projects are renewed. With the FY2026 competition now vacated and no FY2026 awards made, that final trigger is the provision to press HUD on. Read it yourself, confirm your own project expirations against it, and put the question to your HUD field office in writing.^[4]

Obligation. The FY2025 money was announced in three tranches between March 31 and May 21, 2026 — 6,689 awards. As of May 21, 48 of those 6,689 had been fully executed. Not forty-eight percent. Forty-eight. NAEH reported in late July that nearly 700 already-expired grants still had no grant agreement, which is the basis for the claim added against OMB and its director in the July 2 complaint. Vacating unlawful conditions removed a barrier; it did not move a dollar.^[5]

Appropriations. The House passed a clean continuing resolution through December 4 on July 21 and adjourned. The Senate passed its own version through December 11 in the early hours of August 8, 90–6, with anomalies the House did not include — among them a provision that would extend the availability of some homeless assistance grants that would otherwise expire. Because the Senate amended the text, it returns to the House when members come back August 31. Nothing is enacted. If that expiring-grant extension matters to your portfolio, it is a specific, nameable ask for an August recess meeting, and the recess is happening now.^[8]

The disbursement question is the binding one, and it has been for four months. That belongs on the September board agenda regardless of what HUD does with the NOFO.

Los Angeles: the § 578.13 threat is defused for now, and the reasoning cuts everyone

HUD's determination on whether CA-600 — Los Angeles City and County — meets the requirements of the Act was due August 10 or whenever a court ruled on LAHSA's emergency relief, whichever came first. HUD never made it. Its page still carries the pre-deadline holding language.^[7]

On August 13, Judge David O. Carter of the Central District of California granted LAHSA a partial preliminary injunction against HUD's suspension, restoring LAHSA's access to FY2026 federal homelessness funding. Reported scope covers roughly $241 million and 140 projects. Carter found HUD's June 11 action "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law" and "flying in the face of Congressional mandates."^[7]

Two cautions, both load-bearing. First, "partial" has not been resolved. No available source establishes which claims were enjoined and which were denied, or whether the order reaches HUD's June 18 § 578.13 letter as distinct from the June 11 suspension. The order itself is not yet public; the account above rests on wire coverage. Do not characterize its scope in anything you publish or file.^[7]

Second, the opinion is not a vindication of LAHSA either. Carter wrote that HUD has been "complicit," that "for decades, LAHSA and HUD have been in a joint partnership in failure," and that the City-County-LAHSA structure "enabled a cycle of blame-shifting… allowing both entities to evade accountability, with no single party willing to take responsibility." A court that grants you relief while describing your governance that way has told you something. Any CoC whose governance would not survive that paragraph should read it as addressed to them.^[7]

Worth noting for the record: LAHSA's Collaborative Applicant designation was never formally stripped. It was noticed on June 11 and June 18 and then overtaken by events, because the route HUD built to replace it lived inside a NOFO that no longer exists.^[7]

Why Housing First is still not the thing to trade

The scoring language is gone for now. The temptation it created will return with the next NOFO, and the answer to it is the same. It is worth restating while there is time to absorb it rather than three weeks before a deadline.

The finding that Housing First produces housing stability is among the most replicated in homelessness research. In the At Home/Chez Soi five-city randomized trial, Housing First participants spent 73% of follow-up time stably housed against 32% for treatment-as-usual — a roughly 41-point gap.^[9] The systematic-review evidence is candid about the limits: Housing First does not reliably outperform treatment-as-usual on mental health or substance use outcomes.^[10] That honesty is the point. Housing First promises housing stability, and on housing stability it delivers.

A note on discipline, because it matters in a cycle where every claim will be scrutinized: do not claim Housing First reduces mortality. A 2025 secondary analysis of the At Home/Chez Soi trial — 2,255 randomized, 2,108 linked to administrative health data, mortality followed nearly a decade — found no survival difference between Housing First and treatment as usual (adjusted hazard ratio 0.83, 95% CI 0.43–1.22). The point estimate is directionally favorable and the interval is wide, so this is an absence of demonstrated effect rather than a demonstration of absence. But it is the best available answer, and overclaiming on survival is a credibility liability in front of exactly the reviewers who read the literature. Claim what is well-supported: housing stability, reduced emergency system use, cost offsets in high-utilizer populations.^[11]

The national performance data make the case from the other direction. Across the 2020–2024 System Performance Measure series, permanent-housing retention has run at about 96% — the strongest measure in the system, though the series has drifted down slightly across its last three years rather than holding perfectly flat. Once people reach permanent housing, they overwhelmingly stay.^[12] One caveat worth knowing before you cite national comparisons: the federal measurement layer has thinned. FY2024 is the latest verified SPM vintage, and AHAR Part 2 has not been published for the 2023, 2024, or 2025 cycles, so annualized national prevalence figures in circulation are roughly four years old. Cite your own current data where you can, and be accurate about the vintage of anything federal.^[13]

The cost argument has not changed either, though it should be cited carefully. PSH cost offsets outweigh program costs in most U.S. studies measuring both. The figure most often quoted — roughly $16,300 in averted health, crisis, and corrections spending per occupied unit per year — comes from Culhane and colleagues' analysis of New York placements in the 1990s and is stated in 1999 dollars, so it is an estimate rather than a current-cost or causal figure, and the same analysis found Medicaid outpatient spending went up by about $2,000 per unit as people connected to routine care. Use it as an order of magnitude, disclose the vintage, and let the offset argument rest on the direction of the finding rather than the precision of the number.^[14]

What the newest evidence says about the services that NOFO was buying

The set-aside is vacated. The argument for it will be back, and the evidence that answers it got substantially stronger this summer.

Not all housing supports do the same work. A peer-reviewed study of Florida's Section 1115 housing-assistance pilot — 1,300 Medicaid enrollees with serious mental illness or substance use disorder, tracked 2017 to 2024 — compared four distinct service types. Tenancy-sustaining support and peer support were associated with better outcomes, including roughly 20% lower emergency department use for peer support. Transitional-housing support was associated with roughly 15% higher ED use. The study is observational, selection into service types is its central limitation, and its mortality finding should be read against the randomized null described above rather than quoted on its own. But the directional message is consistent with everything else in the evidence base: the value concentrates in the ongoing work of helping someone stay housed — lease compliance, landlord mediation, eviction prevention, benefits navigation — not in one-time transitional assistance.^[15]

That distinction has a specific implication for whatever the set-aside becomes. It covered two different things: new transitional-housing projects and new supportive-services-only projects. The evidence cuts differently across the two halves. Transitional housing is the component the outcome evidence does not support. Services-only is a category that can fund exactly the tenancy-sustaining and peer-support work the same evidence favors — and it can be attached to permanent housing that already exists. If a set-aside returns, the services-only side is where the scoring incentive and the evidence point the same direction. That is worth having worked out in advance rather than discovered under deadline.

Rapid Re-Housing now has a randomized estimate, and the picture is better than it was. Until this summer the strongest evidence was quasi-experimental: a linked-administrative-data study of Los Angeles County that compared RRH enrollees who leased up to otherwise similar enrollees at the same provider in the same month. That study — now published in the Journal of Urban Economics — finds four-year reductions in homeless service use of 30% for individuals and 25% for families, with the authors bounding the true effect between roughly 3 and 20 percentage points. What differs by household type is persistence: effects hold for families and fade for individuals by year three, health and criminal-justice benefits appear for families only, and lease-up produced no increase in later PSH enrollment and no effect on employment or earnings.^[16]

(A correction worth naming, because the figure is still circulating: there is no pooled "28% reduction" in that paper. It appears to have originated in an earlier abstract and survives in search-engine caches. Cite 30% and 25% separately, or cite the bounds.)

Alongside it there is now a genuine randomized trial. A Santa Clara County RCT of temporary rental subsidies for single homeless adults — n=733, enrolled 2018–2023 — finds a 12 percentage-point reduction in homelessness between 6 and 18 months after random assignment, a 32% decline, with most of the effect apparently persisting after the subsidy ends, at roughly $262 per homeless day averted. It is a draft and should be cited as one. But the direction and magnitude converge with the Los Angeles estimate from a completely different design, and that convergence is worth more than either study alone.^[17]

And the strongest new result is about the delivery mechanism, not the program. A three-arm Philadelphia RCT — 301 cash households, 170 voucher households, 725 controls — found that households receiving a monthly cash rental subsidy reported homelessness at 3.0 per 100 at two years against 6.9 for controls, a 57% reduction. The sentence to carry into every landlord conversation is the lease-up one: 100% of households offered the cash subsidy were able to use it, while 25% of households offered a voucher still could not lease a unit — in a city with an above-average voucher use rate. The trial ran entirely outside the pandemic-era rental assistance window. The binding constraint on tenant-based assistance is not household capability. It is landlord acceptance and administrative friction.^[18]

The practical reading for a portfolio: prevention for imminent-risk households, RRH for families and now — with randomized support — for single adults as well, PSH for chronically homeless adults with high service needs. And on the supply side, the constraint to work on is the landlord, not the tenant.

Pull the lever that actually moves performance

If you have capacity freed up by a suspended competition, the evidence says spend it on governance before programs. Multivariate analyses of more than 300 CoCs find that System Performance Measure outcomes vary systematically with governance composition, board structure, federal funding intensity, and local housing-market conditions — not primarily with individual program quality.^[19]

This finding is PROVISIONAL and associational, not causal, so hold it loosely. But it points somewhere useful, and this month it points somewhere urgent. The § 578.13 mechanism is dormant, not dead. HUD's remedial-action authority survived the ruling untouched, and Judge Carter's LAHSA opinion — the most sympathetic judicial treatment any CoC has received this year — still describes a governance structure in which "responsibility is routinely deflected." Documented governance is no longer only a scoring input. It is the record that establishes your CoC is functioning as the statute contemplates, and the next time HUD reaches for § 578.13 it will be reaching at whoever has the weakest record.

Sharpen the by-name list. Tighten coordinated entry. Get conflict-of-interest procedures current and board minutes that show the community actually decided. This is the work a deadline always crowds out, and there is no deadline.

And do not let the litigation distract from the people about to enter your system. Landlords filed more than 1.2 million eviction cases in 2025, essentially back to pre-pandemic volume, with several metros running well above local baselines — Nashville roughly 46% above its 2023–24 average.^[20] On the link from filings to shelter inflow, be precise, because the field routinely overstates it: the best available estimate finds that a one-percentage-point increase in the county eviction filing rate is associated with 0.205 additional sheltered homeless people per 10,000 the following year. Against a mean homelessness rate of 22.3 per 10,000, that is a 0.9% increase — real, measurable, and considerably smaller than "eviction filings drive homelessness" implies. The result with teeth is the asymmetry: filings predict, judgments do not, which the authors read as tenants leaving in anticipation of a judgment. The intervention point is before the courtroom, not in it.^[20]

What is well-evidenced is the response. Targeted prevention and emergency financial assistance have large causal effects and a strong benefit-cost profile, and a randomized Santa Clara prevention trial now puts a number on it: homelessness cut from 4.1% to 0.9% at six months for households reached before housing loss, with a marginal value of public funds of 2.5. Cash delivered after housing loss shows no comparable public-cost offset. Three literatures, three designs, one conclusion.^[21]

What to do now

Eight things, in the order a CoC director should take them.

1. Confirm your renewal position under § 244 of the Consolidated Appropriations Act, 2026 — in writing. This is the money question, and it is now separate from the competition. Identify every project by expiration quarter, map it against the proviso's triggers, and put the resulting question to your HUD field office on paper. Do this first and this week.

1. Do not demobilize the application work. Preserve every draft, ranking record, and project application as it stands. A re-issued NOFO after notice and comment will draw on most of the same material, and the CoCs that kept their binder assembled will be the ones able to respond on a short reopened clock.

1. Prepare to comment. This is the single most consequential shift in what the field should be doing. The ruling turned on HUD's failure to take public comment on an incentive under § 11386b(d), which means the lawful path back runs through a comment docket. Decide now who drafts your CoC's comment, what evidence it carries, and which of your board members and provider partners will file their own. A year of litigation produced a procedural win; the substantive fight happens on that docket, and it will have a short window.

1. Watch two appeal clocks. No notice of appeal of the June 29 or June 30 FY2025 judgments has been found, and that roughly sixty-day federal-defendant window closes around August 28–29, 2026. HUD has said it is "considering all legal options, including appeal to the First Circuit" on the August 7 order, whose own window runs into early October. Neither silence is informative yet.^{[1] [8]}

1. Build the contingency plan for a late or unobligated award — and raise its priority, not lower it. Forty-eight of 6,689 FY2025 awards were fully executed as of May 21. A vacated competition does not fix that, and a re-issued NOFO pushes any FY2026 award further out. Model a scenario in which funds do not arrive on time. Identify which projects need bridge financing, from which local or philanthropic sources, and for how long. That costs one board meeting.

1. Document the treatment and self-sufficiency services your permanent housing projects already deliver, and do it now. The scoring that would have rewarded them is vacated; the version that comes back will want the same thing. These services are routinely undercounted in supportive housing portfolios because they were never the thing being scored. Inventory them, quantify them, and hold the contingency line: connect tenants to services, never make continued housing conditional on accepting them. The randomized evidence is specific — conditioning risks the outcome that is demonstrably at stake without a demonstrated gain on the one it is meant to serve.^[10]

1. Fix governance while there is no deadline. By-name list, coordinated entry, conflict-of-interest procedures, board composition, minutes that show a real community decision. § 578.13 is dormant, not repealed.

1. Defend the front door. Targeted prevention is among the cheapest performance protections available, it is now randomized-trial evidenced, and rising eviction filings make it more important rather than less.

A pause is a resource. It is also a short one.

The FY2026 NOFO is gone. The policy behind it is not, the money still has not moved, and the court that struck the notice wrote an invitation to reissue it into the same paragraph that struck it.

That is a narrower win than the field's relief suggests, and it comes with something the last eighteen months have not offered: time that is not being consumed by a deadline. What that time is spent on is the actual decision in front of CoC leadership right now. Spent on securing renewal continuity, preparing a comment record, and fixing the governance documentation that a remedial-action authority could reach next year, it is the most valuable quarter this field has had since 2024. Spent waiting for HUD to announce something, it is just a gap.

Win the pause. There will not be another one.

Every factual claim in this piece traces to the Common Ladder evidence base (KB IDs in the notes below), a named primary source, or a dated Field-Signal Brief. CANONICAL findings are cited directly; PROVISIONAL findings and contested or unconfirmed items are flagged as such in the body text as well as the notes. Three items are carried with explicit uncertainty and are flagged where they appear: the scope of Judge Carter's August 13 partial preliminary injunction (the order is not public); whether the vacatur of the NOFO in its entirety carries the July 24 modification and Appendix III with it (an inference from the remedy's scope, not a stated holding); and the appeal posture, where silence is not yet informative.

Companion piece: A Federal Court Just Called HUD's Housing First Reversal "Unreasoned." — the June rulings, and what the August 7 order added to them.

Notes

1. ↩ KB W22-1 [DOCUMENTARY]. State of Washington, et al. v. HUD and National Alliance to End Homelessness, et al. v. HUD, C.A. Nos. 26-cv-436-MSM-AEM and 26-cv-439-MSM-AEM (D.R.I.), Judge Mary S. McElroy, Memorandum and Order filed August 7, 2026 (26-cv-436 ECF No. 34, 10 pp.). The slip opinion was read in full. Verbatim holdings: "The Court's review of the record and the parties' arguments reveal that HUD's issuance of the 2026 NOFO violates the APA based on HUD's failure to engage in the notice-and-comment process required by the MVA. See 5 U.S.C. § 706(2)(D)…"; "…this Court must set aside the 2026 NOFO in its entirety."; "The Court, however, denies Plaintiffs' request for a permanent injunction." Mechanism: the $1.3B set-aside is an "incentive" under 42 U.S.C. § 11386b(d), residual category § 11386b(d)(2)(C), which expressly requires "notice and comment to the public." The court called HUD's informal-notice theory "specious" and said it "lacks even marginal support from the text, structure, or history of the MVA." Not decided: the DEI / gender-identity / immigration certifications; the departure from Housing First; the nonprofit plaintiffs' challenge to OMB's two apportionment footnotes conditioning HUD's access to funds on compliance with two executive orders (Vought named in his official capacity). The opinion states on its face that HUD "may attempt to issue yet another NOFO that contains these conditions after undergoing the requisite notice-and-comment process." Procedural notes: both plaintiff groups had alternatively sought preliminary injunctions but "all parties agreed to forgo interim relief and instead move for summary judgment"; the 436 plaintiffs sought only partial summary judgment; the cases are related but not consolidated; a pro-rata-estimate claim under § 11386a(b)(2)(A) was withdrawn after HUD supplied the information. Source: homelessness-knowledge-base.md, W22-1; slip opinion via democracyforward.org.
2. ↩ KB W22-1. HUD, Continuum of Care Program Competition page, fetched August 14, 2026, verbatim: "Due to the court's order, the 2026 CoC NOFO and its August 26, 2026 application submission deadline are no longer in force as HUD is unable to accept applications at this time." And: "HUD is currently evaluating the order and considering all legal options, including appeal to the U.S. Court of Appeals for the First Circuit. Despite the order, HUD stands by the lawfulness of the 2026 CoC NOFO…" ⚠️ The same page has not been reconciled — it still displays "FY 2026 Application Submission Deadline: August 26, 2026, 8:00 PM ET" further down, still frames the competition in the present tense ("more than $4 billion"), still carries the pre-deadline CA-600 holding language, and posted a CoC Application Detailed Instructions PDF for an application that was never loaded into e-snaps. The e-snaps portal remains open for Project Applications and the CoC Priority Listing. LAHSA's competition page, by contrast, renders the deadline in strikethrough under the banner "Please be advised the Court set aside the FY 2026 NOFO in its entirety." The deadline is suspended, not extended; no replacement date exists. Source: hud.gov CoC Program Competition; homelessness-knowledge-base.md, W22-1.
3. ↩ KB W22-1, W20-1, W16-1. The vacated FY2026 NOFO's terms, for reference: $4.04 billion total, of which $1.3 billion was set aside for new transitional-housing and supportive-services-only projects (CRS Insight IN12709, July 15, 2026); the Tier 1 renewal guarantee cut from 90% to 60% of Annual Renewal Demand; scoring rewarding treatment, self-sufficiency, service engagement and law-enforcement partnerships over Housing First. Plaintiff states characterized the set-aside as "a de facto 68% cap on permanent housing funds." The at-risk figure used by every plaintiff and AG in this litigation is 97,000 residents of CoC-funded permanent housing (NAEH estimate, cited in the opinion at ECF 30-1 at 9). ⚠️ Two figures previously carried in this brief are struck: "$3.847 billion" could not be located in any source and has no provenance; and "6,689" is the count of FY2025 awards announced, not FY2026 projects at risk. Source: homelessness-knowledge-base.md, W22-1.
4. ↩ Consolidated Appropriations Act, 2026 (Pub. L. 119-75), § 244 proviso. Congress directed HUD to renew any CoC projects that expired or are set to expire within the first quarter of calendar 2026 (before April 1), with triggers for further delay: if funding is not awarded by April 1, 2026, projects expiring in Q2 2026 are to be renewed; if funding is not awarded by July 1, 2026, all remaining projects are to be renewed. Source: National Association of Counties, "As Litigation Delays New Funding, Congress Directs HUD to Renew Expired Continuum of Care Projects" (published February 4, 2026; updated February 15, 2026). ⚠️ This is read through NACo's summary, not from the enacted statutory text. The proviso is the operative continuity mechanism now that the FY2026 competition is vacated, and it is important enough to your renewals that you should confirm the section text and your own project expiration dates directly rather than rely on this description.
5. ↩ KB W22-1. HUD announced FY2025 CoC renewals in three tranches between March 31 and May 21, 2026 — 6,689 awards in total. NAEH, verbatim: "As of May 21, all awards for FY2025 have been announced. Yet only 48 out of the 6,689 awards have been fully executed." As of July 21–22, 2026, NAEH separately reported "nearly 700 already expired grants still without a grant agreement," which forms the basis of the OMB/Vought claim in No. 1:26-cv-00436. Source: homelessness-knowledge-base.md, W22-1; NAEH.
6. ↩ KB W22-1, W20-1, W19-1, W17-1. FY2025 predicate: summary judgment for the nonprofit-and-municipal coalition on June 29, 2026 (D.R.I.), vacating the November and December 2025 notices and describing HUD's move "to hastily eliminate its Housing First approach" as "the hallmark of unreasoned decision making"; partial summary judgment for a state-AG coalition plus two governors on June 30, 2026, holding the November 2025 conditions unlawful under the APA. FY2026 track: Judge McElroy rejected the coalition's proposed supplemental complaint as one that would "unduly elongate the court's resolution" of the 2025 cases, noting plaintiffs were "undoubtedly free to file a separate action." Two new D.R.I. actions followed — the nonprofit/local-government suit No. 1:26-cv-00436, filed July 2, 2026 (adding OMB and Russell Vought over HUD's untimely obligation of FY2025 awards), and the states' suit No. 1:26-cv-00439, filed July 7, 2026. The states' coalition is 21 attorneys general (20 states plus the District of Columbia) and the governors of Kentucky and Pennsylvania — 22 states and D.C. in total. Co-leads: Washington (Brown), New York (James), Rhode Island (Neronha); the RI and WA AG releases of August 7 agree exactly on the roster. (The earlier FY2025-era coalition count remains unreconciled across sources and stays unverified.) Source: homelessness-knowledge-base.md, W22-1.
7. ↩ KB W22-1. CA-600 / LAHSA. HUD's published status is still the pre-deadline holding language, verbatim: "HUD has not made a final determination as to whether CoC CA-600 meets the requirements of 42 U.S.C. 11360a… HUD will withhold a final determination… until August 10, 2026, or until the relevant court decides whether to grant LAHSA emergency relief, whichever is earlier." The determination was never made. HUD's page characterizes LAHSA as the "formerly designated Collaborative Applicant," but that status was never formally stripped — it was noticed on June 11 and June 18 and overtaken by events. Judge Carter's order: LAHSA v. HUD (C.D. Cal.), Judge David O. Carter, August 13, 2026, a partial preliminary injunction against HUD restoring LAHSA's access to FY2026 federal homelessness funding; reported scope ~$241 million and 140 projects serving 89 veterans, 1,030 families, 1,923 children, 823 transition-age youth, 1,627 seniors, and 901 individuals affected by intimate partner violence. ⚠️ Provenance is City News Service wire copy, not the order. The order PDF could not be obtained; cacd.uscourts.gov did not surface it and CourtListener was unreachable. Quotes reproduced in that coverage: HUD's June 11 action was "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law" and "flying in the face of Congressional mandates"; HUD has been "complicit"; "For decades, LAHSA and HUD have been in a joint partnership in failure"; "This has, at times, enabled a cycle of blame-shifting… allowing both entities to evade accountability, with no single party willing to take responsibility." ⚠️ "Partial" is load-bearing and unresolved — no source establishes which claims were enjoined, or whether the injunction reaches the June 18 § 578.13 letter as distinct from the June 11 suspension. Do not characterize the order's scope until it is read. Separately: the § 578.13 direct-to-HUD route and Appendix III were created by the modified NOFO and fall with it; HUD's underlying remedial-action authority under 42 U.S.C. § 11360a and 24 CFR § 578.13 is statutory and regulatory and was not struck. Also struck from the KB this wave: a previously recorded "30-day extension for direct applicants," which does not appear anywhere in Appendix III (read in full) and should not be repeated. Source: homelessness-knowledge-base.md, W22-1; mynewsla.com (August 13, 2026); hud.gov.
8. ↩ KB W22-1. Appeals. No notice of appeal of the June 29 or June 30, 2026 summary judgments has been found; the FRAP 4(a)(1)(B) sixty-day federal-defendant window closes around August 28–29, 2026 and is genuinely open. HUD's "considering all legal options" language on its live page is forward-looking and refers to the August 7 order. ⚠️ Do not confuse either with the earlier, completed appeal of the December 19, 2025 preliminary injunction (notice of appeal March 2, 2026; USCA No. 26-1217; the First Circuit denied HUD's stay around April 1, 2026 — the last two search-snippet only). Appropriations. The House passed a clean continuing resolution through December 4 on July 21, 2026 and adjourned. The Senate passed its own version through December 11 in the early hours of August 8, 2026 by 90–6, adding anomalies including a provision delaying OMB's proposed Uniform Guidance revisions and language that would extend the availability of some homeless assistance and fair housing grants that would otherwise expire, plus authority for PHAs to repurpose unobligated TBRA for expiring vouchers through calendar 2026. Because the Senate amended the text, the bill returns to the House when members reconvene August 31. Nothing is enacted. Sources: homelessness-knowledge-base.md, W22-1; NLIHC, "Senate Passes Funding Extension Through December 11 and Adjourns for August Recess" (August 10, 2026) and "House Adjourns for August Recess After Sending Continuing Resolution Funding Federal Programs Through December 4 to Senate."
9. ↩ KB F-2 [CANONICAL, SQS 8–9/10]. At Home/Chez Soi five-city RCT: Housing First participants spent 73% of follow-up time in stable housing vs. 32% for treatment-as-usual — an improvement of roughly 41 percentage points. Source: homelessness-knowledge-base.md, F-2.
10. ↩ KB F-3 [CANONICAL]. Rees et al., systematic review and meta-analysis of RCTs (BMJ Open, 2019, SQS 9/10) and eClinicalMedicine (2022): Housing First produces large, consistent housing-stability gains but no clear advantage over treatment-as-usual on mental health, substance use, or quality of life — while reducing non-routine healthcare use. Source: homelessness-knowledge-base.md, F-3.
11. ↩ KB F-28 [CANONICAL, SQS 9/10]. Lachaud, Nisenbaum, Mejia-Lancheros, Latimer, Aubry, et al., Housing and Support Intervention and Mortality Among Homeless Adults With Mental Illnesses: A Secondary Analysis of a Randomized Clinical Trial, JAMA Network Open 2025 Jul 1;8(7):e2524302, doi:10.1001/jamanetworkopen.2025.24302. N = 2,255 randomized, 2,108 (93.5%) linked to health administrative data across five Canadian cities; recruitment 2009–2011, mortality followed to March 30, 2019. Pooled adjusted hazard ratio 0.83 (95% CI 0.43–1.22). Canonized as a well-evidenced null. Bounding limitation: Canadian universal-healthcare setting with a different mortality baseline and treatment-as-usual service floor than the U.S.; secondary analysis not powered for mortality. Source: homelessness-knowledge-base.md, F-28.
12. ↩ KB W13-2 [CANONICAL]. National HUD System Performance Measure benchmarks, 2020–2024: returns to homelessness ≈17–18%; permanent-housing placement from shelter/TH/RRH declined to ≈34%; permanent-housing retention (SPM 7b2) steady at ≈96%. Source: homelessness-knowledge-base.md, W13-2.
13. ↩ KB W20-3. AHAR Part 2 has not been published for the 2023, 2024 or 2025 cycles — the most recent edition covers 2022 and was published August 2024 — so annualized national prevalence figures in circulation are roughly four years stale. FY2025 System Performance Measures are not confirmed posted; FY2024 is the latest verified vintage. 2025 CoC-level PIT/HIC tables are posted. Source: homelessness-knowledge-base.md, W20-3.
14. ↩ KB F-4 and F-5 [CANONICAL]. PSH/Housing First cost offsets outweigh program costs in the majority of U.S. studies measuring both. The ≈$16,281 per-occupied-unit annual savings figure is Culhane, Metraux & Hadley's service-by-service analysis of New York/New York placements (1989–1997), stated in 1999 dollars, quasi-experimental with matched controls; the authors caution it should be treated as an estimate, not a causal effect size. The same analysis found Medicaid outpatient spending increased by ≈$1,982 per unit as tenants connected to routine care — the net figure is net of that increase. One-year housing retention ≈75%. Savings accrue to health, corrections, and crisis budgets — not to the CoC ledger that pays for housing. Source: homelessness-knowledge-base.md, F-4/F-5.
15. ↩ KB F-24 [CANONICAL, SQS 8/10], carried with the F-28 null attached. Impact of Housing Support Services for Medicaid Enrollees with Serious Mental Illness, Substance Use Disorder, Health Affairs (2026), doi:10.1377/hlthaff.2025.00581. Retrospective cohort of 1,300 Florida Section 1115 pilot enrollees, December 2017–June 2024, linking Medicaid claims to vital-statistics records, with multivariable regression controlling for age, sex, race/ethnicity, county, SMI, SUD, prior housing status, enrollment continuity and pre-enrollment health. Peer support: ~20% reduction in ED use. Transitional-housing support: ~15% increase in ED visits. The study also reports 51% lower odds of all-cause mortality for tenancy-sustaining support; that figure is deliberately not used as a headline here because a well-powered randomized null (F-28, note 11) failed to reproduce a survival effect, and the most parsimonious reconciliation is that F-24's mortality association is partly selection. Limitations carried: observational, service receipt not randomized; the transitional-housing ED signal may reflect higher acuity rather than an effect of the service; single state, n=1,300. Access caveat: the publisher full text is paywalled/JS-rendered; methodological detail is abstract- and secondary-confirmed. Source: homelessness-knowledge-base.md, F-24 and F-28.
16. ↩ KB F-26 [CANONICAL, SQS 8/10], updated 2026-08-13. Blackwell, B., & Gross, M., "Does Rapid Re-Housing Reduce Homelessness?", Journal of Urban Economics, August 2026; ScienceDirect PII S0094119026000604. ⚠️ The registered DOI could not be retrieved and must not be constructed from the JUE pattern — cite the PII. (The openICPSR replication package DOI, https://doi.org/10.3886/E250137V1, is a data DOI and must not be conflated with the article DOI.) Design: selection-on-observables, exact matching on project-by-month cells plus regression adjustment on 30+ pre-program characteristics — not an RCT, lottery, RD, or DiD. Analytic n = 3,620 heads of household enrolling in RRH in Los Angeles County in calendar 2019 (individuals 2,116; families 1,504; 2,266 leased up). Subsidy $1,463/month for ~6.9 months. Four-year effects: individuals −0.118 (SE 0.025), −30% off a comparison mean of 39.6%; families −0.121 (SE 0.026), −25% off 49.2%. Authors' own bounds: "we can bound the effect of lease up on the use of homeless services between a 3- and 20-percentage reduction." Fade: by year three individuals −0.017 (0.014), not significant; families −0.045 (0.017), still significant. Days: −77 individuals, −116 families. Two nulls: no increase in PSH enrollment within four years (RRH is not a PSH pipeline), and no effect on employment or quarterly earnings. Lease-up rates range 17%–100% across providers. ⚠️ The "28% reduction over four years" headline does not appear in the paper — the only "28%" is an unrelated footnote on outcome misclassification — and should be struck wherever it appears. ⚠️ The paper reports standard errors and significance stars only; any confidence interval attributed to it is a derivation. Study period overlaps COVID-19, which the authors flag for generalizability. Source: homelessness-knowledge-base.md, F-26.
17. ↩ KB F-37 [PROVISIONAL — capped by draft status; design would otherwise support CANONICAL. SQS 9/10]. Ortuzar, Grace, David C. Phillips, and James X. Sullivan, The Impact of Temporary Rental Subsidies on Homelessness: A Randomized Controlled Trial, working paper, October 2025. AEA RCT Registry AEARCTR-0002533; Notre Dame IRB 17-04-3851. RCT, n = 733 single homeless adults, Santa Clara County CA, enrolled July 2018–November 2023. 12 percentage-point reduction in incidence of homelessness between 6 and 18 months after random assignment — a 32% decline — and roughly a two-thirds reduction in shelter days while the subsidy is active; preliminary sub-sample results indicate most of the effect persists post-subsidy (~9 pp, 28%). First stage strong: treatment 86 pp more likely to enroll in a subsidy program, 46 pp more likely to lease with rental assistance within six months. Cost: $262 per homeless day averted in-window, $103 under a 15-year persistence assumption; author benchmarks HUD-VASH $265/day, At Home/Chez Soi $344/day. ⚠️ Capped at PROVISIONAL: the document is marked "This draft includes preliminary results"; randomization was suspended during part of the COVID period with all eligible participants assigned to treatment, resuming August 31, 2020, with strata and ratio changes on October 1, 2021; the persistence estimate is explicitly a sub-sample. Promote on publication or peer review, not before. Source: homelessness-knowledge-base.md, F-37.
18. ↩ KB F-38 [CANONICAL, SQS 9/10]. Reina, Fowle, Jaffee, Tanski, Mulbry & Fortenberry, PHLHousing+ Housing Security Outcomes After Two Years, The Housing Initiative at Penn. Three-arm RCT: 301 cash households, 170 voucher households, 725 controls; n = 1,196; program began Fall 2022. Cash arm reported homelessness at 3.0 per 100 households at two years vs. 6.9 per 100 for controls — a 57% reduction — with a 63–75% reduction in forced moves and a 22% reduction in serious housing-quality concerns (31% for the voucher arm). Verbatim: "100% of households that were offered the cash rental subsidy were able to use it," against "25% of households offered a voucher were still not able to lease a unit" — in a city with an above-average voucher use rate. The trial ran entirely outside the COVID emergency rental assistance era, which answers the generalizability challenge to F-30. ⚠️ COI disclosed: the Housing Initiative at Penn is both evaluator and design partner on PHLHousing+ — this is a program-team evaluation. Not peer-reviewed. Source: homelessness-knowledge-base.md, F-38.
19. ↩ KB F-21 [PROVISIONAL, Canonical Confidence Score 6/10; constituent source SQS 6–7]. Peer-reviewed multivariate analyses of 300+ CoCs (Kim & Sullivan 2023, Public Administration Review, n=380; Jenisa & Jang 2025, Systems, n=343; Nisar et al. 2019, HUD PD&R) find SPM performance varies systematically with governance composition, board structure, federal funding intensity, and local housing-market conditions — not primarily with individual program quality. PROVISIONAL: associational, not causal; no study tests whether changing these attributes causes SPM improvement. Source: homelessness-knowledge-base.md, F-21.
20. ↩ Filing volumes: Eviction Lab Eviction Tracking System (10 states / 42 cities, updated July 1, 2026) — landlords filed more than 1.2 million eviction cases in 2025, essentially back to pre-pandemic levels; Nashville ran ~46% above its 2023–24 average, with Austin and Greenville also elevated. The elasticity now has a primary source. KB F-36 [PROVISIONAL]: Treglia, D., Byrne, T., & Rai, S. (2023), "Quantifying the Impact of Evictions and Eviction Filings on Homelessness Rates in the United States," Housing Policy Debate, doi:10.1080/10511482.2023.2186749. A one-percentage-point increase in the county eviction filing rate is associated with 0.205 additional sheltered homeless people per 10,000 the following year; no association for eviction judgments, and none for unsheltered homelessness; 67% of the increase is attributable to individuals rather than families. Method: HUD PIT estimates at CoC level matched to county-level Eviction Lab filing and judgment rates, 2007–2017, two-way fixed effects, restricted to CoCs with complete filing data (64% of CoCs in the 2017 PIT), controlling for median household income, temperature, vacancy rate, and cost-burdened renter share. Sample means: filing rate 7.3%, judgment rate 3.1%, homelessness rate 22.3 per 10,000. PROVISIONAL and associational — no causal identification strategy, one study, no replication. Against a mean of 22.3 per 10,000, 0.205 is a 0.9% increase per percentage point — materially smaller than the rhetorical use of this finding implies. Always carry the association framing and the judgments null. Sources: homelessness-knowledge-base.md, F-36; NLIHC Memo to Members, April 10, 2023; Field-Signal Briefs 2026-06-19 item 4 and 2026-07-29 item 4.
21. ↩ KB F-22, F-23 and F-30 [all CANONICAL]. Targeting: Shinn & Greer 2013 (AJPH, Cox regression, n=11,105) and Von Wachter et al. 2021 (LA County Prediction Tool) show statistical models substantially outperform worker judgment in identifying imminent risk. Intervention: Evans et al. 2016 (Science, n=4,448; 76% shelter-entry reduction) and Phillips & Sullivan 2025 (Review of Economics and Statistics, doi:10.1162/rest_a_01344; 81%/73% reduction at 6/12 months; MVPF 2.47). F-30 adds the randomized Santa Clara prevention estimate: homelessness cut from a 4.1% base to 0.9% at six months (−3.8 pp, growing to −5.1 pp at 12 months), n=1,263, MVPF 2.5. The contrast that matters is with F-31: cash delivered after housing loss (Denver Basic Income Project) produced no detectable public-cost offset against an active control. Standing caveat carried from the KB: prevention currently operates at marginal scale relative to the size of the homeless population, so strong per-case effects should not be read as a system-scale solution. Source: homelessness-knowledge-base.md, F-22/F-23/F-30/F-31.

Read next: Intervention Matching — Which Interventions Work for Whom →