FactHub

FactHub curated dataset

Homeless system performance: Seattle/King County vs. 66 large U.S. Continuums of Care, FY2015-FY2024

HUD System Performance Measures (time homeless, returns, first-time homelessness, exits to housing) for 67 large Continuums of Care, with Seattle/King County's trend and rank.

Coverage: Federal fiscal years 2015-2024 (HUD); FY2025 KCRHA-reported figures

Node: dataset:homeless-system-performance:1

Send feedback on this dataset

What this dataset measures

Tests whether King County's homeless system moves people into housing as well as other large regions, using the measures HUD requires every CoC to report. Same 67 CoCs as homelessness-rate-vs-rent.

Geography
United States, King County, WA, Seattle, WA
Unit of observation
CoC by federal fiscal year
Update cadence
Annually, after HUD updates its SPM data file
Curator
FactHub

Curated data tables

These related tables form this versioned dataset release. Each table has its own downloadable file and field definitions.

System Performance Measures, 67 large CoCs, FY2015-FY2024

HUD SPM counts and percentages for each CoC and fiscal year, with population, first-time homelessness per 10,000 and data-quality flags.

Unit of observation: CoC by fiscal year (Oct 1 to Sep 30)

View this table

spm_panel.csv text/csv · 164.7 KB (168,681 bytes)
45 field definitions
FieldTypeMeaningUnitsValue type
coc_code string HUD CoC code — observed
short_name string Short display name — derived
fy integer Federal fiscal year of the report — observed
spm_code_used string CoC code as it appears in HUD's sheet (MO-604M before FY2023) — observed
category string HUD CoC category in the SPM file — observed
hic_nondv_beds_es_th integer Non-DV ES+TH beds in the HIC beds observed
hmis_nondv_beds_es_th integer Non-DV ES+TH beds participating in HMIS beds observed
bed_coverage_pct_reported number HMIS bed coverage as HUD reports it percent observed
m1_es_sh_avg number M1: average days homeless, ES+SH days observed
m1_es_sh_median number M1: median days homeless, ES+SH days observed
m1_es_sh_th_avg number M1: average days homeless, ES+SH+TH days observed
m1_es_sh_th_median number M1: median days homeless, ES+SH+TH days observed
m2_exits_to_ph integer M2 (All): people who exited to permanent housing in the cohort period people observed
m2_ret6 integer M2: of those, returned within 6 months people observed
m2_ret12 integer M2: returned within 12 months (cumulative) people observed
m2_ret24 integer M2: returned within 24 months (cumulative) people observed
m5_first_time_es_sh_th integer M5.1: people entering ES/SH/TH with no HMIS record in the prior 24 months people observed
m5_first_time_es_sh_th_ph integer M5.2: same, including PH entries people observed
m7a_so_exits integer M7a: street outreach exits people observed
m7a_so_temp integer M7a: outreach exits to temporary destinations people observed
m7a_so_perm integer M7a: outreach exits to permanent destinations people observed
m7a_pct_reported number M7a % as HUD reports it percent observed
m7b1_exits integer M7b.1: exits from ES/SH/TH/RRH people observed
m7b1_to_ph integer M7b.1: of those, exits to permanent housing people observed
m7b1_pct_reported number M7b.1 % as reported percent observed
m7b2_universe integer M7b.2: people in PSH/OPH people observed
m7b2_success integer M7b.2: stayed or exited to permanent housing people observed
m7b2_pct_reported number M7b.2 % as reported percent observed
bed_coverage_pct number Recomputed HMIS bed coverage = hmis / hic percent derived
m2_pct6 number Returns within 6 months, % percent derived
m2_pct12 number Returns within 12 months, % percent derived
m2_pct24 number Returns within 24 months, % percent derived
m7a_pct number Recomputed M7a % percent derived
m7b1_pct number Recomputed M7b.1 % percent derived
m7b2_pct number Recomputed M7b.2 % percent derived
population integer ACS population (from dataset homelessness-rate-vs-rent); FY2020 is the mean of 2019 and 2021 people derived
m5_per10k_es_sh_th number First-time homeless (M5.1) per 10,000 residents per 10,000 residents derived
m5_per10k_es_sh_th_ph number First-time homeless (M5.2) per 10,000 residents per 10,000 residents derived
dq_flags string Data-quality flags (e.g. bed_cov<85%, MO-604 split, non-cumulative TH returns) — derived
dq_unknown_dest_pct_SO number Unknown or missing exit destinations, street outreach (FY2019+) percent observed
dq_unknown_dest_pct_ES number Unknown destinations, emergency shelter percent observed
dq_unknown_dest_pct_TH number Unknown destinations, transitional housing percent observed
dq_unknown_dest_pct_RRH number Unknown destinations, rapid rehousing percent observed
dq_unknown_dest_pct_PSH number Unknown destinations, PSH percent observed
pop_note string Notes on the population value — derived

Seattle/King County System Performance Measures, FY2015-FY2024

King County rows from spm-panel, with readable column names.

Unit of observation: Fiscal year

View this table

spm_king.csv text/csv · 590 bytes
9 field definitions
FieldTypeMeaningUnitsValue type
fy integer Federal fiscal year — observed
bedCoveragePct number HMIS bed coverage percent derived
daysHomelessAvg number M1 average days homeless, ES+SH days observed
daysHomelessMedian number M1 median days homeless, ES+SH days observed
returnsWithin24MonthsPct number M2 returns within 24 months percent derived
firstTimePer10k number M5.1 first-time homeless per 10,000 residents per 10,000 residents derived
outreachSuccessfulExitPct number M7a outreach exits to temporary or permanent destinations percent derived
exitsToPermanentHousingPct number M7b.1 exits to permanent housing percent derived
pshRetentionPct number M7b.2 PSH retention or exit to permanent housing percent derived

Supporting files

Source snapshots, manifests, quality reports, documentation, and other materials used to produce or review this release.

  • build_spm.js Transformation manifest · text/javascript · 6.2 KB (6,301 bytes)

    Node.js script that builds spm-panel from the HUD workbook sheets (exported to CSV).

  • rank.js Transformation manifest · text/javascript · 2.7 KB (2,754 bytes)

    Node.js script that ranks WA-500 among the large CoCs, with and without data-quality exclusions.

  • csvlib.js Transformation manifest · text/javascript · 518 bytes

    CSV helper used by the SPM scripts.

  • build_spm_dataset.js Transformation manifest · text/javascript · 23.2 KB (23,731 bytes)

    Node.js script that builds spm-king and the saved analyses in this dataset.

  • findings-spm.md Documentation · text/markdown · 16.5 KB (16,875 bytes)

    Method notes, results, KCRHA cross-check and data-quality caveats for the System Performance Measures.

Saved analyses

Median days homeless in shelter: King County vs. the large-CoC median and Houston, FY2015-FY2024

Question: How long do people stay homeless in King County's shelter system, and how has that compared with other large CoCs over time?

Result: King County's median days homeless in shelter rose from 24 (FY2015) to 119 (FY2024); the large-CoC median went from 30 to 56; Houston from 14 to 36.

Operation: M1 median days homeless (ES+SH) by fiscal year for WA-500 and TX-700, and the median across the 67 CoCs

Scope: 67 large CoCs, FY2015-FY2024

View this analysis

  • SPMs are self-reported from each CoC's HMIS and not audited by HUD; CoCs may revise prior years.
  • HMIS bed coverage differs by CoC and year (King County 77-87%); domestic-violence providers are excluded everywhere.
  • Descriptive comparisons; differences may reflect data practice, coverage or system design as well as performance.
  • FY2020-FY2021 cover the COVID-19 period, when shelters reduced capacity and lengthened stays.

Seattle/King County System Performance Measures, FY2015-FY2024

Question: How did King County's homeless system perform on HUD's main measures from FY2015 to FY2024?

Result: FY2015 to FY2024: average days homeless in shelter 76 to 206.8; exits to permanent housing 30.5% to 49.2%; returns within 24 months 11.7% to 15.3%; first-time homeless per 10,000 31.1 to 32.3.

Operation: WA-500 rows from spm-panel

Scope: King County

View this analysis

  • SPMs are self-reported from each CoC's HMIS and not audited by HUD; CoCs may revise prior years.
  • HMIS bed coverage differs by CoC and year (King County 77-87%); domestic-violence providers are excluded everywhere.
  • Descriptive comparisons; differences may reflect data practice, coverage or system design as well as performance.
  • Destination-based measures (M2, M7) can move tens of points on record cleanup alone; KCRHA attributes part of its FY2025 changes to this.

King County rank among large CoCs on HUD System Performance Measures, FY2019 and FY2024

Question: Where does King County rank among large CoCs on each System Performance Measure?

Result: FY2019 Average days homeless in emergency shelter and safe havens (M1): 104, rank 51 of 67 (median 78); FY2019 Median days homeless in emergency shelter and safe havens (M1): 47, rank 40 of 67 (median 44); FY2019 Returns to homelessness within 24 months of exit to permanent housing, % (M2): 18.8, rank 28 of 67 (median 20.1); FY2019 People homeless for the first time per 10,000 residents (M5): 36.8, rank 45 of 66 (median 19.8); FY2019 Street outreach exits to temporary or permanent destinations, % (M7a): 62.7, rank 15 of 67 (median 37); FY2019 Exits from shelter, transitional housing and rapid rehousing to permanent housing, % (M7b.1): 36.6, rank 42 of 67 (median 42); FY2019 Permanent supportive housing retention or exit to permanent housing, % (M7b.2): 95.4, rank 42 of 67 (median 96.3); FY2024 Average days homeless in emergency shelter and safe havens (M1): 206.8, rank 61 of 67 (median 119.6); FY2024 Median days homeless in emergency shelter and safe havens (M1): 119, rank 57 of 67 (median 56); FY2024 Returns to homelessness within 24 months of exit to permanent housing, % (M2): 15.3, rank 16 of 67 (median 18.8); FY2024 People homeless for the first time per 10,000 residents (M5): 32.3, rank 42 of 67 (median 28.7); FY2024 Street outreach exits to temporary or permanent destinations, % (M7a): 76.1, rank 4 of 67 (median 31.1); FY2024 Exits from shelter, transitional housing and rapid rehousing to permanent housing, % (M7b.1): 49.2, rank 7 of 67 (median 35.6); FY2024 Permanent supportive housing retention or exit to permanent housing, % (M7b.2): 94.4, rank 57 of 67 (median 96.7)

Operation: Rank WA-500 among CoCs reporting each measure; rank 1 = most favorable in HUD's direction

Scope: 66-67 large CoCs per measure

View this analysis

  • SPMs are self-reported from each CoC's HMIS and not audited by HUD; CoCs may revise prior years.
  • HMIS bed coverage differs by CoC and year (King County 77-87%); domestic-violence providers are excluded everywhere.
  • Descriptive comparisons; differences may reflect data practice, coverage or system design as well as performance.
  • Most large CoCs have 30% or more unknown exit destinations for shelter or outreach, which weakens M2 and M7 comparisons.

System Performance Measures for selected large CoCs, FY2024

Question: How does King County compare with other West Coast and large CoCs on HUD's measures in FY2024?

Result: Seattle/King County: median days 119, exits to PH 49.2%, returns 15.3%; Tacoma/Pierce: median days 52, exits to PH 32.9%, returns 15.9%; Portland/Multnomah: median days 82, exits to PH 33.8%, returns 16.2%; San Francisco: median days 126, exits to PH 17.4%, returns 15.2%; Oakland/Alameda: median days 138, exits to PH 39.5%, returns 16%; Los Angeles (excl. Long Beach/Pasadena/Glendale): median days 142, exits to PH 29.6%, returns 16.7%; San Diego: median days 78, exits to PH 31.3%, returns 20.8%; San Jose/Santa Clara: median days 124, exits to PH 39%, returns 23.4%; Metro Denver: median days 49, exits to PH 29.9%, returns 19%; Phoenix/Maricopa: median days 77, exits to PH 42.7%, returns 25.2%; New York City: median days 228, exits to PH 37.7%, returns 13.4%; Boston: median days 141, exits to PH 35.1%, returns 20.6%; Houston/Harris/Fort Bend/Montgomery: median days 36, exits to PH 47.6%, returns 21.2%

Operation: FY2024 values from spm-panel

Scope: Selected large CoCs

View this analysis

  • SPMs are self-reported from each CoC's HMIS and not audited by HUD; CoCs may revise prior years.
  • HMIS bed coverage differs by CoC and year (King County 77-87%); domestic-violence providers are excluded everywhere.
  • Descriptive comparisons; differences may reflect data practice, coverage or system design as well as performance.
  • Comparison set includes the regions named in the accompanying article; the full panel is in spm-panel.

KCRHA-reported figures, FY2024 vs. FY2025

Question: What did KCRHA report for FY2025, the year after HUD's latest published file?

Result: Average time in shelter 248 to 263 days; returns 10% to 15%; first-time homeless 10,332 to 9,592; outreach exits to permanent housing 75% to 50%; annual sheltered 14,470 to 13,884.

Operation: Transcribed from KCRHA's April 1, 2026 CoC Board deck

Scope: King County

View this analysis

  • FY2025 is not yet in HUD's published file.
  • Self-reported by KCRHA.

Derived from

This dataset uses data from these FactHub dataset versions.

Sources

  1. HUD Exchange, CoC System Performance Measures Data Since FY 2015 CoC-level System Performance Measures, one sheet per fiscal year FY2015-FY2024, as submitted by each CoC.

    Retrieved October 10, 2026. U.S. government work (public domain). Public domain; not rehosted here.

    Archived copy captured October 10, 2026

  2. KCRHA, "Understanding System Performance Measures: A HUD Federal Report Overview," CoC Board, April 1, 2026 KCRHA's own FY2024 and FY2025 figures and its explanation that record cleanup drove part of the FY2025 changes.

    Retrieved October 10, 2026. Public agency document. Linked, not rehosted.

    Archived copy captured October 10, 2026

Limitations and reproducibility

Reproducible from public sources: Export each sheet of the HUD SPM workbook to CSV, run spm-build-script with the CoC list and panel from dataset homelessness-rate-vs-rent, then rank-script and spm-dataset-script (Node.js).

  • SPMs are self-reported from each CoC's HMIS and not audited by HUD; CoCs may revise prior years.
  • HMIS bed coverage differs by CoC and year (King County 77-87%); domestic-violence providers are excluded everywhere.
  • Descriptive comparisons; differences may reflect data practice, coverage or system design as well as performance.
  • HUD's FY2024 file has zeroed transitional-housing return counts for most CoCs; only the combined "All" columns are used.
  • HUD revised exit-destination categories in its FY2024 HMIS Data Standards.

Publications using this dataset

Version details

The exact citing versions, and the version of this dataset each cites.

  • Article seattle-homelessness-spending-results — version 5 (published) cites version 1
  • Article seattle-homelessness-spending-results — version 4 (published) cites version 1
  • Article seattle-homelessness-spending-results — version 3 (published) cites version 1
  • Article seattle-homelessness-spending-results — version 2 (published) cites version 1
  • Chart king-county-shelter-days-chart — version 1 (published) cites version 1