GiveBetter x California

WorkSafe

California worker-safety legal support, enforcement and policy advocacy

Research time: ~14 min on GPT-6 Astra Light
  • Research — reviewed programs, finances and impact evidence.

Updated: 2026-09-14

Donate

Summary

What do they do? WorkSafe supports California workers and legal-aid organizations through legal assistance, training and workplace-safety advocacy. It helps workers press for enforcement and policy changes, including protections for people working in civil immigration detention. An ordinary donation supports this legal and advocacy work, worker education and shared operations.

Why we’re interested in this organization:

  • A current, signed settlement demonstrates a concrete route from legal support to required workplace protections.

  • California inspection research provides a causal bridge from effective enforcement to fewer injuries, although not a WorkSafe-specific return.

  • Original audits and current state implementation records make recipient spending, public funding and counterfactual reforms inspectable.

Our main reservations:

  • No costed marginal plan establishes how many additional inspections, implemented protections or injury reductions an extra donation causes.

  • Training, complaints, legal-service counts and fines are not health outcomes; reported service counts are internally inconsistent.

  • Public agency capacity, co-counsel, established reforms and substantial reserves can limit additionality; legal success does not establish durable compliance.

What do you get for your dollar? $161.3M per better life: ten additional quality-adjusted life years in California. Legal-protection health only; other impacts unestimated.

Net additional completed inspections, durable verified hazard-corrected worker-years, counterfactual injury or infection cases avoided, finite health loss avoided per unique person. Complaints, legal advice, PPE deliveries, standing decisions, penalties and nominal legal coverage are intermediate outputs. For recurrent or overlapping injuries, cap summed losses using each person's joint utility/survival trajectory; the per-claim illustration assumes non-overlapping losses and is not valid automatically for repeated claims.

1. What do they do?

WorkSafe is the California nonprofit, EIN 94-2927954, not a government workplace regulator in another country. An ordinary donation supports its legal assistance, worker education, organizing, agency oversight and shared operations. Its State Bar-supported legal program assists qualified legal-service providers and worker organizations rather than simply selling inspections. About and finances Legal support

Its 2025 report describes a two-attorney legal team, worker outreach, heat-policy advocacy and collaboration with the California Collaborative for Immigrant Justice. The contents page reports 36 legal services while the legal chapter reports 40; neither is a count of additional inspections or completed health improvements, and neither measures marginal productivity. The model uses the lower count only as a historical activity-to-cost proxy. 2025 impact report

2. Monitoring and information sharing

A 2026 GEO Group settlement resolves a four-year enforcement dispute involving WorkSafe, CCIJ and Cal/OSHA. The signed agreement is dated May 26; the public announcement is July 7. It requires facility-specific exposure-control plans and protective equipment, but the original Golden State Annex plan had already been accepted in 2023. Expansion takes effect relative to Cal/OSHA approval of the plans, not automatically on the press-release date. The agreement lists seven Class A and five Class B facilities; the announcement’s broader wording should not replace the operative list. Signed settlement Announcement

Follow-up should verify approval dates, worker coverage, training and equipment actually delivered, exposure measurements, sustained hazard correction, infections and injuries against a credible comparison. Track unique workers and worker-time, including transfers out of facilities. The January 2025 ruling concerns civil detainees’ participation rights in proceedings, not a measured reduction in illness. Board decision

The current State Auditor follow-up reports several recommendations fully implemented in July 2026, including complaint-to-onsite criteria and some inspection practices, while others remain partial. Already-implemented reforms belong in the no-new-gift baseline. The revision therefore does not credit a new donor with the 2025 audit or all subsequent agency changes. Current auditor follow-up

3. Qualitative assessment

A California study compared 409 randomly inspected workplaces with 409 matched controls and estimated 9.4% fewer injury claims. It concerned eligible high-hazard establishments with at least ten employees, not WorkSafe-supported complaints, detention facilities or worker training. Its 26% claim-cost reduction is not a QALY conversion. The inspection year and four subsequent years inform a finite effect window; the reference calculation below uses only four future years. Primary inspection study

There is an implementation problem worth addressing, but the counterfactual is moving. The state auditor found inadequate explanations for non-inspection in nine of 30 sampled complaints, alongside weak hazard-abatement documentation. 2025 audit Current follow-up marks inspection-decision, prevention-program review and worker-interview recommendations fully implemented in July 2026, while letter-investigation and accident-investigation recommendations remain partial. WorkSafe cannot be credited with buying those already-implemented changes anew. Current follow-up

The 2025 joint heat survey covers 338 workers at over 200 locations and found continuing reported heat problems after the indoor rule took effect. It is an outreach survey without a counterfactual intervention group, not evidence that training reduced injuries or a representative statewide prevalence estimate. Survey and methods Existing indoor rule

4. What do you get for your dollar?

A conditional estimate for the legal-support-to-workplace-protection channel is about $161 million per 10 California QALYs. This is a partial health price with all recipient costs charged—not WorkSafe’s comprehensive portfolio return or an empirically measured donation effect. Broad policy, organizing, education and non-health rights benefits remain unquantified. It is not a proven lower bound because displacement and harms are uncertain.

The scale anchor is FY 2025 gross audited expenses of $1,739,390 and the impact report’s legal-support activity. Its contents page says 36 units while page 14 says 40; neither counts inspections or unique beneficiaries. Using 36, the model then assumes half of historical-scale capacity is additional, one in five support units changes an implemented protection, one site per success, and half the gross effect survives displaced enforcement. These are explicit analyst priors, not observed conversion rates. Audit Impact report

For health conversion, the model uses the external California randomized inspection study’s 9.4% injury-claim reduction, its descriptive 3.43 annual claims as a baseline proxy, a 50% transfer factor, 0.1 finite QALY per claim, and four years at 3% discounting. The descriptive mean is not a verified current WorkSafe counterfactual, and severity is judgment. Inspection study

Favorable and low-yield stress tests produce roughly $1.05 million and $7.72 trillion per 10 QALYs; these are not credible bounds. Zero and negative outcomes are retained. The dominant unknown is implemented protection per added legal capacity, not the small 36-versus 40 reporting discrepancy. Government and employer resource costs are additional to the donor price.

Model, assumptions and sensitivity

Full ordinary cash gift to Worksafe, Inc., all programs and support. Historical scale uses gross audited functional expenses including event direct benefits. Public enforcement, employer remediation and partner legal resources are excluded from donor denominator but required for a gross-resource estimate.

Only health gains experienced within the California edition boundary count. g=1 applies to already-filtered California worker-time in the illustrative enforcement branch, not automatically to every nationwide precedent or every worker's future life after leaving California.

Partial legal-to-protection channel only. Legal-support equivalents U=(C/E)*Y*m; net inspection equivalents I=U*p*n*v; Q_CA=I*B*e*k*L*sum(t=1..T,(1+r)^(-t))*g-H. Central E=1739390,Y=36,m=.5,p=.2,n=1,v=.5,B=3.43,e=.094,k=.5,L=.1,T=4,r=.03,g=1,H=0. All ordinary-recipient costs charged, no program-only allocation discount. Price=10*C/Q_CA for positive Q_CA. Legal-support units are scale proxies, not measured inspections. Systemic policy, training and other pathways remain unquantified.

E
1739390 USD per fiscal year (observed). Gross FY2025 audited all-function expense including donor direct benefits. [audit25]
Y
36 reported legal-support service units per report year (observed). The report contents says 36 legal-support units while page 14 says 40. The lower count is a conservative scale anchor, not unique cases, inspections or measured marginal productivity. The report and fiscal year differ. [impact25]
m
0.5 marginal to historical scale multiplier (judgment). Judgment: half the historical-scale service capacity is additional after reserves, grants and replacement funding. There is no observed costed funding gap. Charging the entire recipient budget to this channel includes other programs. [audit25] [legal]
p
0.2 probability per additional support unit (judgment). Judgment: one in five added advice, research or case-support units changes implemented protection beyond existing advocates and agency plans. Successful anecdotes do not identify this probability. [impact25] [legal] [settlement]
n
1 gross inspection-equivalent sites per successful unit (judgment). Prior: one-site equivalent; broad precedent and policy benefits not extrapolated. [impact25]
v
0.5 net fraction after public enforcement displacement (judgment). Judgment: half of the gross protection survives diversion of inspectors from other workplaces. The separate p parameter concerns whether legal support changes action; v concerns displaced public action. [stateaudit] [followup]
B
3.43 claims per establishment-year (observed). The external study reports this descriptive mean. Using it as the current untreated baseline for WorkSafe-supported workplaces is a judgment. [inspection]
e
0.094 relative injury claim reduction (observed). External randomized California inspection estimate, not a WorkSafe effect. [inspection]
k
0.5 effect-transfer multiplier (judgment). Prior discounts study-to-current-site hazard/compliance differences. Existing July2026 reforms remain in counterfactual. [inspection] [followup]
L
0.1 QALY per avoided claim (judgment). Judgment: finite health loss per avoided claim, equivalent to a 0.2 utility loss for half a year. No lifetime mortality credit. Recurrent or overlapping losses require person-level caps.
T
4 years (judgment). Finite four-year prevention window; no effect afterward. Claims can recur across years but QALY losses assumed non-overlapping in reference arithmetic. [inspection]
r
0.03 annual discount (judgment). Year-end discount convention.
g
1 California fraction of modeled benefit (judgment). Channel explicitly filtered to California workplace injury morbidity during the finite window; national precedents uncounted. [about]
H
0 QALYs of harm per analytical normalization (judgment). Central convention, not proof of no retaliation or resource harms; separate negative scenario.

Conditional partial legal-protection channel, not portfolio return: Cost: $100K; California QALYs: 0.006201140516679454; all-population QALYs: 0.006201140516679454. E=1739390, Y=36, m=0.5, p=0.2, n=1, v=0.5, B=3.43, e=0.094, k=0.5, L=0.1, T=4, r=0.03, g=1, H=0. All recipient costs retained; unmodeled pathways are excluded, not assumed valueless.

Favorable partial-channel stress test: Cost: $100K; California QALYs: 0.9524951833619643; all-population QALYs: 0.9524951833619643. E=1739390, Y=40, m=0.8, p=0.6, n=3, v=0.8, B=3.43, e=0.094, k=1, L=0.3, T=4, r=0.03, g=1, H=0. All recipient costs retained; unmodeled pathways are excluded, not assumed valueless.

Low marginal protection stress test: Cost: $100K; California QALYs: 1.2957471922558277e-7; all-population QALYs: 1.2957471922558277e-7. E=1739390, Y=36, m=0.1, p=0.02, n=1, v=0.1, B=3.43, e=0.094, k=0.1, L=0.01, T=1, r=0.03, g=1, H=0. All recipient costs retained; unmodeled pathways are excluded, not assumed valueless.

No additional net protection: Cost: $100K; California QALYs: 0; all-population QALYs: 0. E=1739390, Y=36, m=0, p=0.2, n=1, v=0.5, B=3.43, e=0.094, k=0.5, L=0.1, T=4, r=0.03, g=1, H=0. All recipient costs retained; unmodeled pathways are excluded, not assumed valueless.

Net harm stress test: Cost: $100K; California QALYs: -0.003798859483320546; all-population QALYs: -0.003798859483320546. E=1739390, Y=36, m=0.5, p=0.2, n=1, v=0.5, B=3.43, e=0.094, k=0.5, L=0.1, T=4, r=0.03, g=1, H=0.01. All recipient costs retained; unmodeled pathways are excluded, not assumed valueless.

Counterfactual: Absent an extra donation, existing WorkSafe staff, State Bar grants, reserves, CCIJ and pro bono co-counsel continue planned work; Cal/OSHA continues enforcement and implemented reforms. Historical settlement obligations already exist. Extra complaints may redirect scarce inspectors from other hazardous workplaces rather than add inspections. Count only the net change in timing, coverage, durability or hazard correction.

Attribution: Infer the marginal causal contribution from costed work plans and process evidence, not legal-service totals, a proportion of all California workers, fines or equal division among coalition members. Subtract displaced public enforcement benefits and avoid counting the same injury reduction across legal assistance, training, regulation and inspection branches.

Conditional analyst central, not an empirically identified marginal return or comprehensive portfolio estimate. Low-confidence multiplicative priors dominate. Scenarios are stress tests, not confidence intervals. Zero, net harm, and larger omitted policy effects remain possible. The positive partial channel is not a demonstrated lower bound.

Sensitivity

  • Price varies inversely with Y, m, p, n, v, B, e, k, L and discounted duration. None is empirically established as WorkSafe's marginal productivity.
  • Using 40 rather than 36 legal-support units reduces the central price by 10%, much less than uncertainty about implementation.
  • At the central clinical conversion, $1 million per 10 QALYs requires 16.688 net inspection equivalents per $100,000, versus the central 0.1035. A high-value statewide policy mechanism could do better, but this casework proxy does not establish one.
  • If advice chiefly improves wages or rights rather than workplace hazards, p can approach zero without implying that the organization lacks other value.
  • The scope-limited price cannot rank WorkSafe's full advocacy portfolio. Public inspection and employer remediation resources would increase the gross-resource denominator.

Unresolved inputs

  • Current costed ordinary-gift allocation and work forgone absent additional funding.
  • Net additional completed inspections or implemented protections per funded work increment, including public-capacity displacement and co-counsel counterfactual.
  • Counterfactual worker-time and injury/infection incidence by affected site, with severity and health-utility duration.
  • Verified compliance, approval and abatement dates, persistence, recurrence, worker turnover and California residence/time.
  • Marginal agency, employer and partner costs, harms including retaliation, and overlap across pathways.

5. Funding and previous grants

FY2025 revenue was $1,936,433 against $1,714,887 net-presentation expenses. Revenue included $766,648 cy pres awards and $700,308 State Bar funding. Program, management and fundraising expenses were $1,300,470, $328,172 and $86,245 before $24,503 direct event benefits. Net assets were $3,717,649, including $3,699,941 without donor restrictions; deferred revenue was $839,747. Financial assets should not be mistaken for wholly uncommitted cash, but the audit does not demonstrate an urgent unrestricted gap. 2025 audit

A useful funding case would name additional staff time, litigation or follow-up that would otherwise be forgone, its cost and how grants, reserves, cy pres receipts and partner legal work respond. Government-funded enforcement and employer compliance may remain the bottlenecks even when legal-support funding rises. Fines are neither donation revenue nor health gains; coalition participation is not an automatic equal share of causal credit. A future estimate should include every ordinary-gift use and allow unfunded branches, replacement funding and adverse effects, not allocate the full donation to the best historical case.

Annual expenses

Organization-level spending, including programs, administration and fundraising. The research list averages three comparable, consecutive full fiscal years when available.

  • FY 2023: $1.1M; Worksafe, Inc., 12-month period, Audited accrual, fiscal year ended June 30; full gross functional expenses including direct donor event benefits. Source
  • FY 2024: $1.3M; Worksafe, Inc., 12-month period, Audited accrual, fiscal year ended June 30; full gross functional expenses including direct donor event benefits. Source
  • FY 2025: $1.7M; Worksafe, Inc., 12-month period, Audited accrual, fiscal year ended June 30; full gross functional expenses including direct donor event benefits. Source

6. Sources

  1. About WorkSafe and financial reports. Worksafe, Inc.. Published: not stated; retrieved: 2026-09-13.
  2. Donate to WorkSafe, EIN 94-2927954. Worksafe, Inc.. Published: not stated; retrieved: 2026-09-13.
  3. Agency watchdog. Worksafe, Inc.. Published: not stated; retrieved: 2026-09-13.
  4. 2025 Impact Report. Worksafe, Inc.. Published: not stated; retrieved: 2026-09-13.
  5. Audited financial statements, year ended June 30, 2025. Worksafe, Inc. / independent auditors. Published: not stated; retrieved: 2026-09-13.
  6. Audited financial statements, year ended June 30, 2024, with 2023 comparative totals. Worksafe, Inc. / independent auditors. Published: not stated; retrieved: 2026-09-13.
  7. GEO Group settlement announcement. California Collaborative for Immigrant Justice / Worksafe. Published: 2026-07-07; retrieved: 2026-09-13.
  8. GEO Group / Cal-OSHA settlement agreement and mutual releases. GEO Group / California Department of Industrial Relations. Published: not stated; retrieved: 2026-09-13.
  9. GEO Group decision after reconsideration, inspection 1609228. California Occupational Safety and Health Appeals Board. Published: 2025-01-10; retrieved: 2026-09-13.
  10. Randomized government safety inspections reduce worker injuries with no detectable job loss. Levine, Toffel and Johnson / Science. Published: 2012-05-18; retrieved: 2026-09-13.
  11. Cal/OSHA process deficiencies and staffing shortages, report 2024-115. California State Auditor. Published: 2025-07-17; retrieved: 2026-09-13.
  12. Current responses and auditor assessments for report 2024-115. California State Auditor. Published: not stated; retrieved: 2026-09-13.
  13. Cool It: Hazardous Heat, Fast Food and the Path to a Safer Workplace, August 2025. Worksafe / California Fast Food Workers Union, SEIU. Published: not stated; retrieved: 2026-09-13.
  14. New indoor heat protections approved and effective July 23, 2024. California Department of Industrial Relations. Published: 2024-07-24; retrieved: 2026-09-13.