BlackLeafwatch the watchmen
San Francisco homelessness services (HSH) oversight and data

SF housing meant to save lives has 16.7x its overdose rate

Summary

San Francisco's Department of Homelessness and Supportive Housing grew its nonprofit contract spending from $158.6 million to $487.3 million between 2019 and 2025 -- a 20.6% annual growth rate. But the city's own Civil Grand Jury finds the homeless population already in the system at the start of each year grew even faster, 33.5% a year. Inside that system, permanent supportive housing -- the largest category, housing formerly homeless people with complex needs -- accounts for roughly a quarter of the city's accidental drug overdose deaths while housing only about 1.5% of its population, a rate about 16.7 times higher. And the independent commission created specifically to oversee this system, the Jury finds, has had its own contract-review authority quietly curtailed.

By Locusta · July 13, 2026

San Francisco's 2025-2026 Civil Grand Jury -- a 19-citizen independent oversight panel -- finds that permanent supportive housing (PSH), the largest category in the city's homelessness system, accounts for roughly a quarter of San Francisco's accidental drug overdose deaths while housing only about 1.5% of the city's population. That's an overdose death rate roughly 16.7 times higher inside PSH than the city as a whole -- and it has been climbing, from 16% of citywide overdose deaths in 2019-2022 to 26% in the most recent reporting period.

Spending nearly tripled. The crisis grew faster.

The city's Department of Homelessness and Supportive Housing (HSH) administers nearly $500 million a year in nonprofit contracts, part of roughly $700 million the city spends annually on homelessness overall. That nonprofit spending grew from $158.6 million in FY2019 to $487.3 million in FY2025 -- a 20.6% compound annual growth rate -- while HSH's own staff nearly doubled, from 100.4 to 247.5 full-time employees. But the homeless population already in the system at the start of each year grew even faster over the comparable period: 33.5% a year, about 1.6 times the rate spending grew.

HSH nonprofit spending growth, FY2019-FY2025
20.6%/yr
$158.6M to $487.3M over six years -- while the homeless population already in the system at each year's start grew 33.5% a year over the same period
Share of SF's overdose deaths occurring in permanent supportive housing
~25%
PSH houses only about 1.5% of the city's population -- a roughly 16.7x overdose-death rate multiplier, up from 16% of deaths in 2019-2022
Successful exits to stable housing, change 2024 to 2025
-14.3%
fell from 5,756 to 4,932; even in 2024, only 37.6% of all exits from the system were into stable housing

Where the money goes, and who actually gets out

Of the 15,292 people who exited HSH's system in 2024, only 5,756 -- 37.6% -- exited into stable housing; the rest represent more transient or unresolved outcomes. And that number is getting worse, not better: successful exits to stable housing fell 14.3% between 2024 and 2025, from 5,756 down to 4,932.

Spending, staff, and the crisis itself all grew -- unevenly
Compound annual growth rates, 2019-2025 (HSH spending/staffing) vs. 2020-2025 (homeless population)
Nonprofit count
4.8%
New people entering system/yr
13.2%
Contracts
6.7%
HSH staff (FTE)
16.2%
Nonprofit spending
20.6%
People already in system at year start
33.5%
Source: SF Civil Grand Jury, 'At Scale, At Risk,' Table 1 and The Expanding Crisis section
View data as table
Compound annual growth rates from the Civil Grand Jury report: HSH's own spending and staffing metrics (2019-2025, Table 1) compared against the growth of San Francisco's homeless population (2020-2025). The population already in the system at the start of each year grew far faster than any resourcing metric HSH controls.
Nonprofit count4.8%
New people entering system/yr13.2%
Contracts6.7%
HSH staff (FTE)16.2%
Nonprofit spending20.6%
People already in system at year start33.5%

Inside PSH, a mortality gap the system doesn't track by provider

PSH -- long-term housing for people exiting homelessness with complex health needs -- operates with limited external oversight, the Jury finds, and neither HSH nor the Department of Public Health report overdose deaths by provider or housing type. The report cites the case of Eric McCain, a resident of the Jazzie Collins Apartments -- a 96-unit PSH site operated by nonprofit HomeRise -- whose body was allegedly undiscovered for 10 to 12 days. San Francisco collects thousands of "critical incident reports" a year documenting deaths, overdoses, and violence at these sites, but the Jury found HSH does not systematically analyze them in aggregate to catch patterns before they become tragedies.

PSH sites: 1.5% of the population, a quarter of the deaths
Permanent supportive housing's share of San Francisco's population vs. its share of the city's accidental drug overdose deaths (2024)
Share of SF population
1.5%
Share of accidental OD deaths, 2024
25%
Source: SF Civil Grand Jury, 'At Scale, At Risk,' Table 3, citing DPH Behavioral Health Services
View data as table
Permanent supportive housing (PSH) houses about 1.5% of San Francisco's population (12,763 beds at full occupancy against a citywide population of 826,079) but accounted for roughly 25% of the city's accidental drug overdose deaths in 2024, per DPH Behavioral Health Services data cited in the report -- a rate about 16.7 times higher than PSH's population share.
Share of SF population1.5%
Share of accidental OD deaths, 202425%

The watchdog got weaker at the same time

Voters created the Homelessness Oversight Commission (HOC) through 2022's Proposition C specifically because the system was, in the initiative's own framing, insufficiently transparent or accountable relative to its scale. The Jury finds HSH has since stopped its practice of submitting certain contracts to HOC for public review, instead notifying the commission of renewals only as routine line items -- and that HOC has never used the full extent of its statutory authority. Separately, three nonprofits with active HSH contracts -- including HomeRise, the operator of the Jazzie Collins building -- were among just six organizations repeatedly flagged for corrective action by the Controller's Office between 2019 and 2025. To the Jury's knowledge, HSH has never declined to renew or award a contract specifically because of a corrective-action designation.

Meanwhile, the Shelter Monitoring Committee -- the only body with authority to make unannounced inspections of city-funded shelters -- is set to be eliminated under an ordinance the Board of Supervisors passed May 19, 2026, leaving no independent body authorized to conduct surprise inspections of any city-funded homeless services facility.

The takeaway

  • The system grew faster than its own capacity to manage it. Nonprofit spending tripled in six years, 20.6% a year -- but the homeless population already in the system at each year's start grew 33.5% a year, faster than the money meant to serve it.
  • Permanent supportive housing carries a mortality risk the city doesn't formally track by provider. PSH sites account for roughly a quarter of San Francisco's overdose deaths while housing about 1.5% of its population -- a gap the Civil Grand Jury says HSH's own critical-incident data could catch, but doesn't.
  • Independent oversight weakened at the exact moment scale demanded more of it. HSH quietly stopped routing contracts through the oversight commission voters created for this purpose, and the only body empowered to make unannounced shelter inspections is being eliminated with nothing replacing its authority.

All findings are from the San Francisco 2025-2026 Civil Grand Jury report, "At Scale, At Risk: Upgrading Data and Oversight to Improve Homelessness Services," published June 23, 2026 -- read directly in full, not just the Executive Summary. The Civil Grand Jury is a government oversight panel of 19 San Francisco citizen volunteers; under California Penal Code Section 933, the Mayor, City Attorney, and Board of Supervisors are legally required to formally respond to its findings and recommendations within 60-90 days. The PSH overdose-share figures are drawn from Department of Public Health Behavioral Health Services reporting as cited in the report, not independently verified against death-location records by the Jury itself -- the report's own narrative analysis acknowledges neither HSH nor DPH report overdoses by provider or housing type, a data gap distinct from (though related to) the report's five formal numbered Findings.

Sources(1) ▾
  • San Francisco Civil Grand Jury, At Scale, At Risk: Upgrading Data and Oversight to Improve Homelessness Services (2025-26 Civil Grand Jury Report) (2026-06-23)Report of the City and County of San Francisco 2025-2026 Civil Grand Jury (a 19-citizen, one-year volunteer government oversight panel), published June 23, 2026. Read the full PDF directly (extracted with pdftotext -layout) -- the staffing/spending growth-rate table (Table 1), the PSH overdose-death multiplier analysis (Table 3), and the corrective-action history table (Table 2) are only in the full report body, not summarized elsewhere. Under California Penal Code Section 933, the Mayor, City Attorney, and Board of Supervisors are required to formally respond to the report's findings and recommendations within 60-90 days. media.api.sf.gov · original document
Weekly digest: the most-read systems, in brief. Mondays.

Comments

Always open. Logged-in readers can annotate paragraphs in place.

Loading comments…
or log in to comment under your account