Skip to content
ContentLora

    Tip: press / anywhere to search.

    Analysis

    Does Housing First work? The evidence and the 2026 policy fight

    Randomized trials consistently find that Housing First, which offers permanent housing without requiring sobriety or treatment first, keeps more people housed.[1][2] Evidence that it improves health, addiction or quality of life is weaker.[3][4] The Trump administration says Housing First has failed and is redirecting funds toward treatment and transitional housing, a shift now being contested in court.[5][6][7]

    Editor reviewedStrict sourcingUpdated AnalysisSocial welfare, Mental health, Tech policy
    Show:

    “Housing First” was the organising idea of US homelessness policy for about a decade. In 2025 and 2026 the federal government moved to dismantle it.[5][8] This page separates what the research shows from how each side reads it. It is not medical or legal advice. Anyone dealing with mental illness or addiction should talk to a clinician, and US readers in emotional distress can call, text or chat the 988 Suicide & Crisis Lifeline at any hour.[9]

    What the trials show

    Housing First provides permanent housing without first requiring sobriety or treatment; services are offered but not required.[10] Much of the evidence comes from randomized controlled trials (RCTs), which assign people to programmes by chance so that outcomes can be compared fairly.[1]

    • New York, 2004. 225 homeless adults with mental illness were randomly assigned to housing conditional on treatment and sobriety or to immediate housing. The Housing First group got housed sooner and stayed housed. The control group used more addiction treatment, but substance use and psychiatric symptoms did not differ.[11]
    • Canada, At Home/Chez Soi. 1,198 homeless adults with mental illness in four cities were randomized. Scattered-site housing with intensive case management raised housing stability over 24 months but did not improve generic quality of life.[2] After one year, 13.5% of Housing First participants met the study’s definition of housing instability.[12]
    • Meta-analysis, 2019. Pooling four RCTs, participants spent more days housed and were about 2.5 times as likely to be housed at 18-24 months. All four trials were rated at high risk of bias.[1] Emergency department visits and hospitalisations were lower.[13] There was no clear difference in substance use, and short-term health effects were imprecise.[3]
    • National Academies, 2018. A review of permanent supportive housing found no substantial published evidence that it improves health outcomes or cuts health care costs, except some evidence for people with HIV/AIDS. It noted that such housing helps some people become and remain housed.[4] Pre/post studies often showed big cost savings, but the few careful RCTs did not.[14]

    What the population data show

    HUD dates its Housing First policy changes from 2013.[5] It reports that between 2013 and 2025 homelessness rose 27%, unsheltered homelessness 36% and chronic homelessness 81%, while Continuum of Care spending rose 111%.[15] Housing costs also climbed: the median asking rent in 2025 was nearly 18% above 2019 after inflation, and about half of renter households were cost burdened in 2024.[16][17] Veteran homelessness fell 56% from 2009 to 2025, a decline HUD has attributed to targeted and sustained funding.[18][19]

    Two kinds of evidence point in different directions, and they answer different questions. The RCTs ask whether a given person offered Housing First is more likely to be housed than one offered treatment-first services. They consistently say yes. The national trend asks whether a system built around Housing First lowered total homelessness. That is much harder to judge, because rents, migration and local policy moved at the same time. HUD’s 2013-2025 comparison does not control for those factors. The trials are also mostly of people with serious mental illness in a few cities, so they say less about families or people whose main problem is cost.

    The policy shift

    The July 24, 2025 executive order directs HHS and HUD to end support for Housing First “to the extent permitted by law”. It also allows HUD to require treatment as a condition of participation for people with substance use disorder or serious mental illness.[8] HUD’s FY2025 and FY2026 Continuum of Care notices put less money into renewing permanent housing and prioritised transitional housing and services.[6] Up to 20 of 200 points now reward treatment and recovery services and up to 8 reward requiring participation in services.[20] HUD had not funded new transitional housing projects since FY2012.[21] A district court found both FY2025 notices unlawful under the Administrative Procedure Act and vacated the FY2026 notice. A First Circuit stay on September 16, 2026 let the FY2026 competition go ahead pending appeal.[22][7]

    Our assessment, not a prediction from any official source: if the FY2026 awards go forward as designed, some permanent supportive housing grants may not be renewed in full. People moved into transitional housing would be counted as homeless, while people in permanent supportive housing are not.[23][24] That could raise measured homelessness in future counts even if the same people are served. The outcome depends on the First Circuit appeal and on Congress, which must act on the FY2027 request to eliminate the CoC program.[25]

    Where the views differ

    The disagreement is mostly about values and goals, not about the trial results. Supporters stress housing stability and point to the RCTs.[26] Critics say the RCT results are real but narrow. They argue that housing alone does not treat addiction or mental illness, and that the system-wide numbers worsened.[27][15] The administration goes further. It calls for civil commitment and enforcement of camping bans, which the Supreme Court’s 2024 Grants Pass ruling made easier for cities.[28][29][30] For background, see homelessness explained. For dated developments, see the homelessness policy tracker.

    Competing views

    Housing First works and should stay

    Advocates such as the National Alliance to End Homelessness point to randomized trials showing faster, more stable housing and fewer emergency visits, without more substance use.[26][1][13][11]

    Housing First has been oversold

    The Manhattan Institute and the administration argue it stabilises housing but does little for addiction or mental illness, and that homelessness rose while it was federal policy.[27][5][15][4]

    Treatment and enforcement first

    The July 2025 executive order calls for civil commitment, treatment as a condition of assistance and enforcement of camping bans, saying most people on the streets have addiction or mental illness.[31][28][8][29]

    Next in the crash course

    Developing in World & Policy

    Questions readers ask

    What is Housing First?

    An approach that provides rapid access to permanent housing that is not conditional on sobriety or treatment. Services are offered but not required.[10]

    Does Housing First reduce homelessness for the people it serves?

    In randomized trials, yes. A 2019 meta-analysis of four trials found participants spent more days housed and were about 2.5 times as likely to be housed at 18-24 months, though all four trials had a high risk of bias.[1]

    Does Housing First improve mental health or addiction?

    The evidence is weak. The 2019 meta-analysis found no clear difference in substance use and imprecise effects on short-term health, and a 2018 National Academies review found no substantial evidence that permanent supportive housing improves health outcomes, except some evidence for people with HIV/AIDS.[3][4]

    Does Housing First save money?

    The National Academies committee said pre/post studies often showed large cost reductions, but the few careful randomized trials did not show significant cost savings.[14]

    Why did the federal government turn against Housing First?

    HUD Secretary Scott Turner said in May 2026 that Housing First had failed to meaningfully reduce homelessness, and HUD reported that homelessness rose 27% between 2013 and 2025. The July 2025 executive order directs agencies to end support for Housing First.[5][15][8]

    Sources

    · 31 claims from 14 publishers

    Each numbered claim is a statement we checked against the sources listed with it. Tap a number in the text to see its sources in place.

    1. [1]

      A 2019 meta-analysis of four randomized trials found Housing First participants spent more days housed and were about 2.5 times as likely to be housed at 18-24 months; all four trials were rated at high risk of bias. confirmedas of 2019-02-18

    2. [2]

      Canada's At Home/Chez Soi randomized trial enrolled 1,198 homeless adults with mental illness in four cities; scattered-site housing with intensive case management increased housing stability over 24 months but did not improve generic quality of life compared with usual care. confirmedas of 2015-03-03

    3. [3]

      The same meta-analysis found no clear difference in substance use, imprecise effects on most short-term health outcomes, and said long-term health effects need further study. confirmedas of 2019-02-18

    4. [4]

      A 2018 National Academies review found no substantial published evidence that permanent supportive housing improves health outcomes or reduces health care costs, except some evidence for people with HIV/AIDS, while noting that it helps some people become and remain housed. confirmedas of 2026-10-11

    5. [5]

      Releasing the 2025 count, HUD Secretary Scott Turner said the "housing first" status quo had failed to meaningfully reduce homelessness; HUD measured trends from 2013, which it described as the start of Housing First policy changes in its programs. confirmedas of 2026-05-29

    6. [6]

      The FY2025 and FY2026 CoC funding notices shifted priorities toward treatment and recovery and required participation in services, made less money available to renew existing grants (mostly permanent housing), and prioritized new transitional housing and supportive services projects. confirmedas of 2026-09-25

    7. [7]

      A lawsuit challenging the FY2026 notice was filed on July 2, 2026; on August 7 the district court vacated the notice under the Administrative Procedure Act, and on September 16 a First Circuit panel stayed that order pending appeal, letting the competition proceed. confirmedas of 2026-09-18

    8. [8]

      The order directs HHS and HUD, to the extent permitted by law, to end support for "housing first" policies and lets HUD require program participants with substance use disorder or serious mental illness to use treatment as a condition of participation. confirmedas of 2025-07-24

    9. [9]

      The 988 Suicide & Crisis Lifeline provides free, confidential emotional support by call, text or chat to people in suicidal crisis or emotional distress 24 hours a day across the US and its territories, through a network of over 200 local crisis centers. confirmedas of 2026-10-11

    10. [10]

      Housing First provides rapid access to permanent housing that is not conditional on sobriety or on taking part in treatment; services are offered but not required. confirmedas of 2026-10-11

    11. [11]

      In a 2004 New York trial, 225 homeless adults with mental illness were randomly assigned to housing conditional on treatment and sobriety or to immediate housing without prerequisites; the Housing First group got housed sooner and stayed housed, with no difference in substance use or psychiatric symptoms. confirmedas of 2004-04-01

    12. [12]

      In At Home/Chez Soi, 13.5% of Housing First participants met the study's criteria for housing instability after one year. confirmedas of 2026-10-11

    13. [13]

      In the 2019 meta-analysis, Housing First groups had fewer emergency department visits and hospitalizations than control groups. confirmedas of 2019-02-18

    14. [14]

      The National Academies committee said pre/post studies often showed large cost reductions, but the few careful randomized trials failed to show significant cost reductions or health improvements. confirmedas of 2026-10-11

    15. [16]

      In 2024, 22.7 million renter households (49.4%) and 20.7 million owner households (23.9%) were cost burdened, and 26.2% of renter households were severely cost burdened. confirmedas of 2026-05-13

    16. [17]

      The median asking rent was $1,492 in 2025, down 2% from 2024 after inflation but nearly 18% above 2019 in inflation-adjusted terms; asking rents cover new leases, not rents paid by current tenants. confirmedas of 2026-09-25

    17. [18]

      In 2025, 32,495 veterans were counted as homeless, the lowest number since veteran data collection began in 2009 and 56% below that year. confirmedas of 2026-05-29

    18. [19]

      HUD's 2024 report attributed the long decline in veteran homelessness to targeted and sustained funding. confirmedas of 2024-12-31

    19. [20]

      Under the FY2026 notice, up to 20 of 200 points reward providing treatment and recovery services and up to 8 reward requiring participation in supportive services; in FY2024, participation was not required under the Housing First approach. confirmedas of 2026-09-18

    20. [21]

      HUD had not funded new transitional housing projects through the CoC program since FY2012, before the FY2026 notice prioritized them. confirmedas of 2026-09-18

    21. [22]

      On June 29, 2026, the district court ruled that HUD violated the Administrative Procedure Act in issuing both the November and December 2025 CoC notices. confirmedas of 2026-07-15

    22. [23]

      People living in rapid rehousing and permanent supportive housing are not counted as homeless in the point-in-time count, while people in transitional housing are. confirmedas of 2026-09-18

    23. [24]

      HUD released the FY2026 CoC notice on June 1, 2026 and modified it on July 24, 2026. It offers about $4.04 billion, sets aside $1.3 billion for new grants prioritized for transitional housing and supportive services-only projects, and protects 60% of renewal demand in tier 1. confirmedas of 2026-09-18

    24. [25]

      The President's FY2027 budget, submitted April 3, 2026, proposes cutting Homeless Assistance Grants by $393 million (9%), eliminating the CoC program and consolidating homeless assistance under the Emergency Solutions Grants program. confirmedas of 2026-04-30

    25. [26]

      Supporters such as the National Alliance to End Homelessness say Housing First gets people housed faster and keeps them housed, a finding supported by randomized trials and a 2019 meta-analysis. confirmedas of 2026-10-11· interpretation

    26. [27]

      The Manhattan Institute, a conservative think tank, argues that Housing First proponents overstate its ability to end homelessness, its cost-effectiveness and its effects on well-being, and that its record on mental illness and addiction is weaker than on residential stability, a gap the National Academies review also found on health outcomes. confirmedas of 2026-10-11· interpretation

    27. [28]

      The order directs the Attorney General to seek reversal of precedents and consent decrees that impede civil commitment, and to help states adopt "maximally flexible" civil commitment and treatment standards. confirmedas of 2025-07-24

    28. [29]

      The order tells agencies to consider prioritizing discretionary grants for states and cities that enforce bans on open drug use and on urban camping and loitering. confirmedas of 2025-07-24

    29. [30]

      In City of Grants Pass v. Johnson (decided June 28, 2024), the Supreme Court held that enforcing generally applicable laws regulating camping on public property is not cruel and unusual punishment under the Eighth Amendment. confirmedas of 2024-06-28

    30. [31]

      The executive order asserts that the overwhelming majority of people living on the streets are addicted to drugs, have a mental health condition, or both, and cites 274,224 unsheltered people in the last year of the previous administration as a record. confirmedas of 2025-07-24

    Revision history (1)
    1. Page created.

    Created Oct 11, 2026. Last reviewed by an editor on Oct 11, 2026. Next scheduled review: Dec 15, 2026.

    Cite this page

    "Does Housing First work? The evidence and the 2026 policy fight." ContentLora, updated Oct 11, 2026. https://contentlora.com/analysis/housing-first-evidence-debate

    Spotted an error? Suggest a correction or emailcorrections@contentlora.com.

    ☰ Contents