JAMA Netw Open. 2026 Apr 1;9(4):e265187. doi: 10.1001/jamanetworkopen.2026.5187.
ABSTRACT
IMPORTANCE: Homelessness has risen sharply in the US in recent years, posing a major threat to population health. Although researchers have identified factors contributing to differences in homelessness prevalence between jurisdictions, factors contributing to year-over-year change in homelessness are poorly understood.
OBJECTIVE: To examine competing explanations for rising levels of homelessness within US states between 2019 and 2024.
DESIGN, SETTING, AND PARTICIPANTS: This longitudinal cohort study was conducted to measure the association between changing state-level factors and relative change in counts of people experiencing homelessness from 2019 to 2024. The sample comprised adults (aged ≥18 years) and children (aged <18 years) experiencing homelessness in all 50 states and Washington, DC, as measured in the US Department of Housing and Urban Development point-in-time (PIT) homeless count data. Due to PIT count limitations owing to the pandemic, outcome data for 2021 were excluded.
EXPOSURES: State-year average rents, unemployment, pandemic-era housing supports (including emergency rental assistance and eviction moratoria), substance use (proxied by overdoses), immigration, and climate-related property damage were measured via a variety of publicly available datasets.
MAIN OUTCOMES AND MEASURES: The primary outcome was year-over-year change in homelessness within states overall, by homelessness type (sheltered or unsheltered), and by age (children or adults).
RESULTS: From 2019 to 2024 (excluding 2021; N = 255 state-years), eviction moratoria and climate-related damages were consistently associated with change in homelessness within the 50 states and Washington, DC. The mean (SD) year-over-year increase in homelessness was 7.0% (22.5%), and there were mean (SD) changes of 6.1% (29.1%) for sheltered homelessness, 16.0% (36.5%) for unsheltered homelessness, 7.2% (20.6%) for adult homelessness, and 8.4% (44.0%) for child homelessness. Each additional 1.0% of person-time covered by a moratorium was associated with a 0.36-percentage point (pp) lower increase in homelessness (95% CI, 0.10-0.61 pp; P = .01). Each home-equivalent lost to climate-related events per 10 000 population was associated with a 1.00-pp greater increase in overall homelessness (95% CI, 0.78-1.22 pp; P < .001). Results were largely consistent by homelessness type and age. Whereas the mean (SD) year-over-year increase in homelessness was 11.0% (36.4%) for a typical state in 2022, homelessness would have risen by an estimated 19.9% (95% CI, 13.5%-26.3%) absent eviction moratoria, while counts would have risen by only 7.8% (95% CI, 7.1%-8.5%) absent property loss due to climate-related disasters.
CONCLUSIONS AND RELEVANCE: In this cohort study of 50 states and Washington, DC, measured over 5 years, eviction moratoria and climate-related damages were associated with within-state, relative change in homelessness. These findings suggest that to attenuate surges in homelessness, policymakers should take action to prevent evictions and address climate change.
PMID:41941184 | DOI:10.1001/jamanetworkopen.2026.5187
