A Federal Court Just Called HUD's Housing First Reversal "Unreasoned."
Common Ladder — Learn / Evidence. A news-pegged analysis of the June 29, 2026 federal court ruling in National Alliance to End Homelessness v. HUD, and of the August 7, 2026 order that followed it. Written for funders, federal and state policymakers, CoC leadership, and the engaged public. Procedural status verified as of August 14, 2026 — the FY2026 CoC NOFO has since been vacated in its entirety and the August 26 deadline is no longer in force.
On June 29, 2026, a federal court vacated HUD's attempt to steer Continuum of Care funding away from Housing First, ruling that the agency dismantled a two-decade-old evidence-based policy without reasoned analysis — "the hallmark of unreasoned decision making."[1]
That June ruling struck the FY2025 rules and declined to block the FY2026 competition, calling forward relief speculative. Six weeks later the same judge reached FY2026 anyway. On August 7, 2026, Judge McElroy granted summary judgment in the two new FY2026 cases and set aside HUD's issuance of the 2026 NOFO in its entirety. HUD's competition page now states that the NOFO "and its August 26, 2026 application submission deadline are no longer in force as HUD is unable to accept applications at this time."[2]
But the two rulings are not the same kind of win, and the difference is the whole story. June was a reasoned-decision holding — HUD changed an evidence-based policy without reckoning with the record. August was a notice-and-comment holding — HUD skipped a procedural step the McKinney-Vento Act requires before attaching an incentive to the money. In August the court expressly declined to reach the Housing First question, denied a permanent injunction, and wrote that HUD "may attempt to issue yet another NOFO that contains these conditions after undergoing the requisite notice-and-comment process."[2]
So the record now holds one ruling that says the reversal was unreasoned and a second that says this particular vehicle was unlawful — while the policy itself remains entirely available to an agency willing to run a comment docket. What follows is what each court decided, what neither did, and what the people who fund and run this system should do with a pause that has a short shelf life.
A Continuum of Care director in a mid-sized county spent June reading a federal court order that vindicated everything her programs are built on, and July assembling an application that asked her to compete for money by promising to do the opposite. In August she learned she cannot submit it. The notice it was written against no longer exists.
She has spent fifteen years building a system that houses people first and stabilizes them from there. One order calls the attempt to dismantle that approach "unreasoned." A second sets aside the vehicle that would have funded the dismantling — on the ground that HUD did not take public comment first, and with an explicit note that HUD may try again once it does.
That gap — between what the courts have affirmed and what the funding architecture is still free to demand — is the whole story of this moment. It is worth being precise about every part of it.
The argument
Two courts have now ruled against HUD on the same policy for two different reasons — that it was adopted without the reasoned analysis the law requires, and that it was funded through a notice the agency had no authority to issue without public comment. The pivot was never grounded in evidence; the evidence runs the other way. But neither ruling reached the substance, and the second one says so on its face. The policy is procedurally interrupted, not defeated — which means the people who fund and run this system have to make the evidentiary case on a record the courts have declined to build for them.
What the court decided
The case is National Alliance to End Homelessness v. HUD, brought by a coalition that includes major cities and counties — Santa Clara, San Francisco, King County, Boston, and Tucson, Arizona among the named plaintiffs. On June 29, U.S. District Judge Mary S. McElroy granted summary judgment against HUD and vacated the November and December 2025 notices that had reengineered the FY2025 CoC competition — the rules that capped permanent housing at roughly 30% of CoC funds and deemphasized Housing First.[1]
The court did not weigh in on whether Housing First is good policy. It ruled on process, under the Administrative Procedure Act, and its language was unusually blunt. HUD's move "to hastily eliminate its Housing First approach," the opinion held, "serve[s] as the hallmark of unreasoned decision making." It created a "procedural nightmare" for localities and failed to weigh the harm of a "rapid, untimely overhaul." The prior FY24–25 funding framework is restored for the FY2025 competition.[1]
One day later, on June 30, the same court partially granted summary judgment to a separate coalition — nineteen state attorneys general plus the governors of Kentucky and Pennsylvania — holding the November 2025 conditions unlawful under the Administrative Procedure Act. Two rulings, two days, two sets of plaintiffs, one conclusion.[3]
Read the phrase carefully, because it is exact. Unreasoned does not mean unpopular. It means an agency changed a long-standing, evidence-based policy without doing the analysis that would justify the change — without reckoning with the record, the reliance interests, or the harm. That is precisely the objection the evidence has always raised against the reversal: not that it is cruel, though it is, but that it is unsupported.
Why "unreasoned" is the right word
The reversal was sold as a correction — the claim that Housing First had been tried and failed. The record does not support that claim. Housing First produces the strongest evidence in the field for one specific, decisive outcome: keeping people housed. Across multiple randomized trials, it outperforms treatment-first models on housing retention and residential stability. Its results on clinical measures — substance use, psychiatric symptoms — are more mixed, and honest advocates say so. But housing stability is not a side benefit. It is the point, and it is the thing HUD's own scoring grades most consistently: permanent-housing retention runs around 96% nationally, and it is Housing First that produces it.[4] [5]
The economics point the same direction. Permanent supportive housing does not pay for itself, but for the highest-need individuals it offsets a large share of its cost — on the order of $16,000 per unit per year in avoided emergency-room visits, inpatient stays, jail time, and shelter. A funding pivot toward transitional housing and treatment-conditioned assistance is a pivot away from the intervention with the best retention evidence and the clearest cost offset, and toward models whose track record for this population is weaker.[6]
There is a harder edge to this, too. The enforcement-paired posture that travels with "recovery first" — sweeps, camping bans, move-along orders as the front door to services — is not neutral. A peer-reviewed simulation across 23 cities estimates that continual involuntary displacement of unsheltered people who inject drugs would cause between 974 and 2,175 additional overdose deaths per 10,000 over a decade. "Unreasoned" is the court's word. It is also a mercy of understatement.[7]
What the June ruling did not decide — and how August closed part of the gap
Here is the part a celebratory headline missed at the time. The June court vacated the FY2025 rules but declined to issue a forward injunction against the FY2026 competition, calling that relief "speculative." So the FY2026 CoC NOFO — $4.04 billion, carrying forward the same "recovery first" scoring and a renewal guarantee cut from 90% to 60% — remained active through the summer, with applications due August 26.[2] [8]
The coalition tried to fold FY2026 into the existing cases through a supplemental complaint. Judge McElroy rejected it, holding that it would "unduly elongate the court's resolution" of the 2025 matters while noting that plaintiffs were "undoubtedly free to file a separate action." They were, and they did — two new District of Rhode Island actions in early July: the nonprofit-and-local-government coalition on July 2 (newly naming OMB and its director over HUD's untimely obligation of the FY2025 awards), and a states' action on July 7 brought by 21 attorneys general — 20 states plus the District of Columbia — and the governors of Kentucky and Pennsylvania, arguing the $1.3 billion set-aside functions as a de facto 68% cap on permanent housing funds and puts at least 97,000 people at risk.[9]
Both were decided on August 7, and the plaintiffs won. All parties agreed to forgo interim relief and move straight to summary judgment. The court held that HUD's issuance of the 2026 NOFO violated the APA "based on HUD's failure to engage in the notice-and-comment process required by the MVA," and set the NOFO aside in its entirety. The mechanism is specific: the $1.3 billion set-aside is an "incentive" under 42 U.S.C. § 11386b(d), falling in the residual category at § 11386b(d)(2)(C), which expressly requires "notice and comment to the public." HUD's contrary theory was called "specious" and said to lack "even marginal support from the text, structure, or history" of the Act.[2]
And here is the new gap, which is narrower than the old one but sharper. The August court denied a permanent injunction. It declined to reach the departure from Housing First. It declined to reach the DEI, gender-identity, and immigration certifications. It declined to reach the challenge to OMB's two apportionment footnotes conditioning HUD's access to the funds on compliance with executive orders. And it wrote that HUD "may attempt to issue yet another NOFO that contains these conditions after undergoing the requisite notice-and-comment process."[2]
So the policy pivot is not dead. It is procedurally interrupted. Two coalitions have now beaten HUD three times in three months, and not one of those rulings has held that the substance of the reversal is unlawful — only that the agency keeps failing to do it properly. The next round will not be litigated. It will be commented on. That is a different arena, with a different constituency and a much shorter window, and the field has spent a year building capacity for the wrong one.
The coalition estimated the vacated FY2025 conditions would have pushed roughly 170,000 people back into homelessness — an advocacy figure, and worth flagging as one, but a fair sense of the stakes. The figure carried through the FY2026 litigation and cited in the August opinion is 97,000 residents of CoC-funded permanent housing. Those stakes did not disappear on August 7. They moved to a docket that has not opened yet.[10]
Winning is not the same as getting paid
There is a second gap, quieter than the first and arguably more urgent. HUD announced all FY2025 renewals in three tranches between March 31 and May 21 — 6,689 awards. As of May 21, per the National Alliance to End Homelessness, 48 of those 6,689 had been fully executed. Not forty-eight percent. Forty-eight. By late July, NAEH reported that nearly 700 already-expired grants still had no executed grant agreement.[11]
Read that against the rulings. Two coalitions have now won three times, and providers still do not have agreements in hand. Vacatur removed the illegal conditions; it did not obligate the money. Disbursement risk is a distinct failure mode from award risk, and it has been the binding one since spring — which is why the July 2 complaint added the Office of Management and Budget as a defendant, and why the August 7 opinion's refusal to reach the OMB apportionment footnotes leaves the most operationally consequential claim in the case undecided.
The continuity mechanism that actually protects projects is not the competition at all. Section 244 of the Consolidated Appropriations Act, 2026 directs HUD to non-competitively renew expiring CoC projects, with escalating triggers tied to HUD's own delay — including one, reached July 1, 2026, that extends to all remaining projects if awards have not been made. With the FY2026 competition vacated and no FY2026 awards made, that proviso is the provision CoCs and their counsel should be reading this month.[12]
The larger current: courts and Congress vs. the funding architecture
Step back and the ruling is not an isolated event. In the same stretch of weeks, Congress passed the 21st Century ROAD to Housing Act by overwhelming bipartisan margins — 85–5 in the Senate, 358–32 in the House — and it became law on July 11, 2026 without a presidential signature. Its homelessness provisions are structural and non-punitive: a shelter-and-outreach spending-cap waiver, veteran-income protections, landlord-participation reforms, a hundred thousand new units through RAD.[13]
So two branches of government are moving toward evidence-based, structural housing policy — a court striking down an unreasoned reversal, a Congress authorizing structural investment — while the executive's funding architecture moves the other way. That divergence is the real map of this moment. The evidence is not lonely anymore. It has a court opinion and a near-unanimous bill on its side. What it still lacks is the appropriations and the administrative posture to match, and that is where the pressure now has to come from.
What to do now
The rulings change what is arguable. They do not settle what the money will be used for. Here is what follows, by audience.
CoC leaders: confirm your renewal position under § 244 of the appropriations act, in writing, this week — that is now the provision your projects depend on, and it does not run through the competition. Do not demobilize the application work; a re-issued NOFO will draw on most of the same material. And decide now who drafts your comment when the docket opens. The detailed practitioner version of this is in the companion guidance.
Funders and foundations: this is still the moment to backfill, and the case got stronger rather than weaker. Forty-eight of 6,689 FY2025 awards were fully executed as of May 21; a vacated competition pushes any FY2026 award further out, not nearer. Philanthropic capital that bridges evidence-based programs through an interrupted funding cycle is not charity — it is protecting a proven system from a disbursement failure no court has been willing to fix.
Federal and state policymakers: read both opinions, not the summaries. June set a standard — "unreasoned" is a standard, not a slogan, and any future change to homelessness funding has to clear it. August set a procedure: incentives attached to McKinney-Vento money require notice and comment. Neither has been tested against a version of this policy that respects both. Hold the next proposal to both.
Everyone watching: the window has moved, not closed. HUD's lawful route back is a rulemaking, and a comment docket is a far more accessible arena than a federal courtroom — it is open to every provider, every CoC board, every person who has been housed by one of these programs. The evidence now has two court opinions behind it. The task is to get it onto the record where the next decision is actually made.
The argument is won three times; the funding is not
Courts have now said in law what the evidence established long ago: you cannot walk away from what works by pretending it doesn't, and you cannot rewire the money that funds it without telling the public first. Unreasoned was the right word in June. Notice-and-comment was the right holding in August.
But a ruling is not a rehousing, and neither of these rulings reached the substance. Forty-eight contracts out of 6,689 are executed. The policy that produced all of this is intact and available to anyone willing to run a comment period. The people whose housing depends on how this resolves cannot wait for the litigation to catch up, because the litigation has done nearly all it can do.
The evidence won the argument. Whether it wins the docket is still up to the people who read this.
Every factual claim in this piece traces to the Common Ladder evidence base (KB IDs in the notes below) or a named primary source. CANONICAL findings are cited directly; the one advocacy-sourced figure (~170,000) is flagged in the body and note as a magnitude anchor with a conflict-of-interest caveat, not a verified count. Court quotations are from the June 29, 2026 summary judgment opinion.
Companion piece: The FY26 NOFO Is Vacated. That Is Not the Same as Winning. — what CoC leaders should actually do with a suspended competition.
Notes
Notes
- ↩ KB W17-1 (CANONICAL, documentary; SQS 8/10). National Alliance to End Homelessness et al. v. U.S. Department of Housing and Urban Development, U.S. District Court for the District of Rhode Island (Judge Mary S. McElroy), summary judgment issued June 29, 2026 (gov.uscourts.rid.60977, Doc. 99). Quoted language is from the opinion.
- ↩ The August 7, 2026 order. 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.); slip opinion read in full. Verbatim: "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"; HUD's informal-notice theory was called "specious" and said to "lack[] even marginal support from the text, structure, or history of the MVA." Expressly not reached: the DEI / gender-identity / immigration certifications; the departure from Housing First; the challenge to OMB's two apportionment footnotes conditioning HUD's access to funds on compliance with two executive orders. The opinion states that HUD "may attempt to issue yet another NOFO that contains these conditions after undergoing the requisite notice-and-comment process." HUD's response, verbatim from its competition page (fetched August 14, 2026): "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 and still displays the August 26 deadline further down. The deadline is suspended, not extended. Because the order sets the NOFO aside in its entirety, the July 24 modification and its Appendix III — the § 578.13 direct-to-HUD route — fall with it; that is an inference from the scope of the remedy, not a separately stated holding. 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. The vacated NOFO's terms, for reference: $4.04B total, $1.3B set-aside (CRS Insight IN12709, July 15, 2026); Tier 1 renewal guarantee cut from 90% to 60% of ARD; scoring rewarding treatment participation, work/service requirements and law-enforcement partnerships. Sources: homelessness-knowledge-base.md, W22-1; slip opinion via democracyforward.org; hud.gov CoC Program Competition.
- ↩ KB W20-1 (CANONICAL, documentary). Partial summary judgment for a nineteen-state attorney general coalition plus the governors of Kentucky and Pennsylvania, D.R.I., June 30, 2026, holding the November 2025 CoC NOFO conditions unlawful under the Administrative Procedure Act. Sources: Oregon Department of Justice (June 30, 2026); Rhode Island Current (July 9, 2026).
- ↩ KB F-2 (CANONICAL). Housing First produces the strongest evidence for housing retention and residential stability across multiple RCTs and quasi-experimental studies; KB F-3 (CANONICAL) notes clinical outcomes (substance use, mental health symptoms) are more mixed — housing stability is the outcome the model reliably produces.
- ↩ KB W13-2 (CANONICAL). National System Performance Measure benchmarks: permanent-housing retention ~96%; returns to homelessness ~17–18%; PH placement ~34%. Retention is the outcome HUD grades most consistently, and it is the one Housing First produces best.
- ↩ KB F-4 / F-5 (CANONICAL). Permanent supportive housing generates a partial cost offset — on the order of ~$16,281 per unit per year in avoided crisis-system costs (emergency departments, inpatient, jail, shelter) — concentrated among the highest-need, highest-cost individuals. The offset reduces but does not fully eliminate the cost of a PSH unit.
- ↩ KB W16-2 (CANONICAL). Barocas et al., JAMA 2023;329(17):1478–1486, simulation across 23 U.S. cities: continual involuntary displacement of unsheltered people who inject drugs is estimated to cause 974–2,175 additional overdose deaths per 10,000 over 10 years (~15–25% of all deaths in that population), plus more hospitalizations and fewer treatment initiations.
- ↩ KB W22-1, W20-1. The FY2026 competition schedule as it stood before vacatur: CoC Builds applications due July 23, 2026; full applications due August 26, 2026 at 8:00 PM ET. Project applications and detailed instructions did not reach e-snaps until the evening of July 21, 2026, and the CoC Application itself was never loaded — HUD's page still states CoC Applications "are not yet available but will be loaded into e-snaps soon," while separately posting a CoC Application Detailed Instructions PDF for it. HUD amended the NOFO on July 24, 2026 (Modified NOFO plus Appendix III). All of the above is superseded by the August 7 vacatur (note 2). ⚠️ A "30-day extension for direct applicants" previously recorded in this project's knowledge base does not appear anywhere in Appendix III, which was fetched and read in full, and has been struck. Source: homelessness-knowledge-base.md, W22-1; hud.gov.
- ↩ KB W22-1, W20-1 [DOCUMENTARY]. Judge McElroy rejected the coalition's proposed supplemental complaint (filed ~June 22, 2026) as one that would "unduly elongate the court's resolution" of the 2025 cases, noting plaintiffs "are undoubtedly free to file a separate action." Two new D.R.I. actions followed: NAEH, NLIHC et al. v. HUD, Turner, OMB and Vought, No. 1:26-cv-00436, filed July 2, 2026, and Washington et al. v. HUD, 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. Verbatim from RI AG Neronha's August 7 release: "On July 7, 2026, Attorney General Neronha co-led a coalition of 21 attorneys general and two governors in filing the lawsuit against the U.S. Department of Housing and Urban Development (HUD)…" Co-leads Washington (Brown), New York (James), Rhode Island (Neronha); joining: Arizona, California, Colorado, Connecticut, Delaware, D.C., Illinois, Maine, Maryland, Massachusetts, Michigan, Minnesota, New Jersey, New Mexico, Oregon, Vermont, Virginia, Wisconsin. The RI and WA releases agree exactly. (This corrects a "more than twenty attorneys general" formulation carried in earlier versions of this piece, and a "22 AGs" figure carried in this project's knowledge base — 22 is defensible as a state count and wrong as an AG count. The separate FY2025-era coalition count remains unreconciled across sources and stays unverified.) The complaint argues the $1.3B set-aside is "a de facto 68% cap on permanent housing funds," with at least 97,000 people at risk — the figure cited in the August 7 opinion at ECF 30-1 at 9. 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. Both were decided August 7, 2026 — see note 2. Appeals: no notice of appeal of the June 29 or June 30 judgments has been found, and the FRAP 4(a)(1)(B) sixty-day federal-defendant window does not close until approximately August 28–29, 2026, so silence is not yet informative. ⚠️ 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; First Circuit denied HUD's stay ~April 1, 2026). ⚠️ Do not confuse the D.R.I. proceedings with the separate § 705 stay Judge Carter granted LAHSA in the Central District of California on July 2, 2026, or with his August 13, 2026 partial preliminary injunction in that case — a different court, a different case, and an order whose scope is not yet public. Sources: homelessness-knowledge-base.md, W22-1; RI and WA Attorney General releases (August 7, 2026); Democracy Forward.
- ↩ Coalition harm estimate: ~170,000 people would have been pushed back into homelessness under the vacated FY2025 conditions (Democracy Forward / NLIHC press release, June 29, 2026). Advocacy-sourced; methodology not published — treat as a magnitude anchor with a conflict-of-interest flag, not a verified figure.
- ↩ KB W22-1. HUD announced FY2025 CoC renewals under the Consolidated Appropriations Act, 2026 (Pub. L. 119-75) in three tranches between March 31 and May 21, 2026, totalling 6,689 awards. 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." ⚠️ Two figures carried in earlier versions of this piece are struck. The per-tranche dollar breakdown and the aggregate "$3.847 billion" could not be located in any source and have no traceable provenance — do not reintroduce them. And 6,689 is the count of FY2025 awards announced, not FY2026 projects at risk; the at-risk figure in this litigation is 97,000 residents of CoC-funded permanent housing. 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. The FY2026 Grant Inventory Worksheets were posted per-CoC in early August; they are now artifacts of a vacated competition. Sources: homelessness-knowledge-base.md, W22-1; NAEH; HUD.gov.
- ↩ 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 escalating triggers tied to HUD's own 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. With the FY2026 competition vacated and no FY2026 awards made, this proviso — not the NOFO — is the operative continuity mechanism. 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). ⚠️ Read here through NACo's summary, not from the enacted statutory text. Confirm the section language and your own project expiration dates directly before relying on it.
- ↩ 21st Century ROAD to Housing Act (H.R. 6644, 119th Congress), passed Senate 85–5 (June 22, 2026) and House 358–32 (June 23, 2026); became law July 11, 2026 without a presidential signature under the Article I, Section 7 ten-day clock — Public Law 119-101. Homelessness-relevant provisions are structural, not punitive: ESG 60% shelter/outreach cap waiver, USICH progress reporting, HUD-VASH veteran income exclusion, HCV landlord-participation reforms, RAD +100k units. Authorized is not appropriated — the Act sets policy and authorizes programs; the money still runs through appropriations. Sources: Congress.gov H.R.6644; House Financial Services Committee, "21st Century ROAD to Housing Act Becomes Law"; National League of Cities, "Ten Things for Local Leaders" (July 2, 2026).
Read next: The FY26 NOFO Is Vacated — what CoC leaders should do with a suspended competition →
← Back to Evidence & policy