Summary
What do they do? The legal recipient is East Bay Bicycle Coalition, doing business as Bike East Bay, EIN 94-2585652. It provides education, events, technical street-safety work and other cycling programs in Alameda and Contra Costa counties. It is distinct from Marin and Silicon Valley bicycle coalitions. More
Why this approach interests us
- An unresolved corridor design can affect many users. Bike East Bay also teaches cycling skills, which may enable additional activity for some people.
Our main reservations
- The public agencies already plan safety work. Historical corridor injuries are not all preventable or donor-attributable, and education attendance is not measured additional health.
What do you get for your dollar?
Signed subjective mixture: $19,089,888 per 10 Bay QALYs. Central scenario: $104,277,379. Whole organization cost is retained while two health pathways are quantified; this is not complete expected value. Inspect the model, assumptions and scenarios.
- whole expense
- $1,908,395 — 2024 functional expense plus separately netted event and inventory costs.
- bay mixture
- $19,089,888 — Per 10 Bay QALYs; signed, subjective and partial-health.
- central
- $104,277,379 — Finite annual-budget cohort; not a measured giving return.
- staff provenance
- Devin Jacob, 2018 — Personal-giving blog entry names Bike East Bay outside GiveWell's criteria.
1. What do they do?
The legal recipient is East Bay Bicycle Coalition, doing business as Bike East Bay, EIN 94-2585652. It provides education, events, technical street-safety work and other cycling programs in Alameda and Contra Costa counties. It is distinct from Marin and Silicon Valley bicycle coalitions.
Use historical injuries as a scale anchor
CCTA's current project page provides a defined corridor and time period. Convert its fatal and severe-injury counts into annual historical rates, then remove risk expected to be addressed by existing projects.
Require actual implementation change
An annual cohort of technical work receives only a share of an additional design change, conditional on implementation. Agencies, other organizations and Bike East Bay's past/future work retain their own contributions.
Model finite, additional activity
Historical education attendance becomes activity only after judgments for additional cycling, alternative exercise, repeated participants, finite utility and overlap.
Charge the whole organization
Keep all programs, overhead, events and netted expenses. Other benefits remain unquantified—not zero. Public construction and volunteered resources are not fully priced.
Scope of this review. No completed lanes, petitions or countywide population totals are treated as donor-created health. One annual budget receives a finite project share, not repeated credit for the same project every year. Physical activity excludes road-injury health already counted. No electoral-persuasion pathway is modeled.
2. Monitoring and information sharing
GiveWell employee Devin Jacob's own donations
2018 personal-giving statement. Names Bike East Bay among local interests, outside organizations meeting GiveWell criteria.
Our assessment. Discovery provenance only; neither organizational GiveWell endorsement nor proof of a 2026 gift.
East Bay Bicycle Coalition DBA Bike East Bay
Original 2024 IRS return. Whole expense $1,908,395 after netted-cost addbacks; actual education output stated.
Our assessment. Historical accounting, not marginal donor capacity or program-only spending.
5.5-mile Contra Costa San Pablo corridor, July2020–June2025
CCTA agency SWITRS summary. 6 deaths and35 severely injured people; future recommendations still under development.
Our assessment. Defined exposure scale, not a causal effect or all preventable burden. Existing projects require residual adjustment.
Several distinct San Pablo segments
Official organization corridor-status inventory. Some improvements completed or funded; others still planned.
Our assessment. Do not merge different segments or claim all22miles as new. Broad campaign establishes relevance, not commissioned work on the precise CCTA task.
Four-lane undivided road diet settings
FHWA official technical synthesis. 19–47% total-crash reduction summarized.
Our assessment. Not a fatality RR, not all protected designs, and not the exact CCTA layout.
12 road diets and25comparison sections in California/Washington
2004 matched before/after primary study. Approximately6% frequency reduction; no significant rate effect.
Our assessment. Observational selection/exposure limitations and heterogeneous effects support null/harm uncertainty.
122 Spanish adolescents
2022 randomized school cycling intervention. Knowledge improved but cycling/active commuting did not change.
Our assessment. Attendance/skills cannot automatically generate sustained activity. Local adult services differ.
43 previously noncycling Danish children
2012 randomized trial, primary abstract. Cardiometabolic composite improved; randomized fitness difference not significant.
Our assessment. Plausible mechanism, not utility calibration or lifelong mortality prevention.
3. Qualitative assessment
An unresolved corridor design can affect many users. Bike East Bay also teaches cycling skills, which may enable additional activity for some people.
Key reservations
- Ordinary gift-to-implementation contribution remains unmeasured.
- Historical risk needs adjustment for existing improvements and changing exposure.
- Education is not observed additional physical activity.
- Future public and partner resource cost is incomplete.
- Favorable tail dominates; both central and favorable cases exceed $1M.
- 2024 finance is not a current funding ledger.
- Staff giving is not an endorsement or causal evidence.
Benefits not included in our estimate
- Already completed or fully funded projects credited automatically to a new gift
- Entire22-mile campaign burden or unrelated Alameda injury counts added to the5.5-mile CCTA cohort
- Training attendance, signatures or event reach treated as health
- Lifetime exercise mortality from a class
- Other corridors, transit access, pollution, dignity and community wellbeing—unquantified, not zero
- Electoral-persuasion effects
4. What do you get for your dollar?
$19,089,888 per 10 Bay QALYs
Weighted finite Bay health per annual-budget equivalent is 0.9996889638491746 QALYs; $10,000 yields 0.005238375513712699 Bay QALYs. Central finite Bay health is 0.18301140860764387 per budget equivalent. These are scenario outputs, not observed outcomes or available packages.
A better life is our comparison unit of 10 additional quality-adjusted life years (QALYs), potentially spread across people. These are uncertain estimates, not measured returns or verified donation offers.
How we calculate the estimate
Finite whole-gift health: gift / whole expense × cash response × Bay share × [(historical annual fatal/severe burden × current-risk scale × residual share × net reduction × discounted advantage-years × annual-work contribution × realization) + additional activity × finite utility × nonoverlap − residual harm]
0.005238375513712699 Bay QALYs per $10,000
Model inputs and assumptions
- Whole accounting resource expense
- $1,908,395 (range: $1,897,776 + $7,478 + $3,141). 2024 original return, including event and inventory costs netted outside Part IX. Noncash resources are not automatically added again. Not a 2026 budget or full public-construction cost. Original filing.
- Five-year corridor injury anchor
- 6 deaths / 35 severe injuries (range: 1.2 deaths and 7 severe injuries/year). CCTA reports July 2020–June 2025 on its 5.5-mile Contra Costa segment. Counts are people, not county totals; raw SWITRS records were not independently extracted. Primary agency summary.
- Current untreated risk versus history
- 1 (range: 0.5–1 declared worlds). Historical traffic and reporting, pandemic conditions and future travel can differ. This is a judgment, not a forecast inferred from one year. Very uncertain / analyst judgment.
- Risk remaining beyond existing improvements
- 75% (range: 50–90%). Separately removes the impact of completed/committed corridor work, including Del Norte and Caltrans baseline. Not a measured segment-weighted estimate. Do not simply remove miles: injuries are not uniformly distributed. Very uncertain / analyst judgment.
- Finite loss per fatal/severe injury
- 10 / 2 QALYs (range: 5–15 / 0.5–3). Integrated finite remaining health/disability priors, disjoint fatal and nonfatal people. No local age/severity utility data. Never multiply by another lifetime horizon. Very uncertain / analyst judgment.
- Net safety-burden reduction
- 20% (range: 10–40%, zero; signed diagnostics allowed). Judgment includes design transfer and local severity effects. FHWA overall-crash results do not establish a mortality reduction here. Main rate covers existing corridor exposure; residual harm separately covers induced/displaced exposure. Very uncertain / analyst judgment.
- Counterfactual advantage and delay
- 3 years after 4-year lag (range: 1–5 years; 2–6-year lag). Integrate annual burden over a finite advantage window discounted at 3%; alternative design eventually catches up. This is not the whole asset life. No selected construction date verified. Very uncertain / analyst judgment.
- Annual work cohort contribution
- 10% (range: 3–25%). Conditional contribution of one annual budget's technical work, diluted by agencies, partners and Bike East Bay's own past/future work. No credit for entire multiyear organizational effort. Very uncertain / analyst judgment.
- Actual implementation probability
- 50% (range: 25–80%). Additional design must be built within modeled timing. Separate from contribution conditional on implementation. Current planning is not construction approval. Very uncertain / analyst judgment.
- Education participant equivalents
- 1,158 (range: 868.5–1,389.6 current modeled scale). Original 2024 report: 1,158 adults/children at 81 classes. Future scale .75–1.2 is judgment. Attendance not a verified unique treatment cohort. Historical output plus judgment.
- Net additional active people
- 10% × 30% = 34.74 (range: 3–30% extra cycling; 20–50% beyond other activity). First fraction must include within-cohort repeat-person/activity-period adjustment and changed cycling; second removes substitute exercise/access. Not probability of health response; not imported trial ITT. Very uncertain / analyst judgment.
- Short-term activity utility
- 0.005 for 0.5 year (range: 0.002–0.02 for 0.25–1 year). Generic finite preference-weighted proxy, not a measured fitness-to-QALY conversion. Additional nonoverlap .8 central (.5–.8) removes duplicated health periods, particularly across road and activity pathways. Very uncertain / analyst judgment.
- Ordinary-gift additional work
- 40% (range: 0–70%). Additional work after public/private funding substitution and capacity limits. Not empirically verified funding room and not a second implementation probability. Very uncertain / analyst judgment.
- Bay share
- 98% (range: 98% across declared worlds). Local exposure but not every traveler is a resident. Same share applied to harms; no headquarters-based 100% assumption. Very uncertain / analyst judgment.
- Other added adverse health
- 0.1 QALY/budget equivalent (range: 0–1). Induced cycling exposure, displaced traffic or other adverse health beyond the modeled treated-corridor net effect. Do not duplicate injuries already reduced in the main pathway. Very uncertain / analyst judgment.
- Joint subjective worlds
- 20% funding-null; 20% health-null; 10% harm (range: 20% cautious; 20% central; 10% favorable). Declared before first run; no empirical calibration or target-price tuning. Very uncertain / analyst judgment.
$10,000 illustration—not a priced additional lane, course or package
- funding-null: No positive finite price. 20% weight; no additional health.
- health-null: No positive finite price. 20% weight; zero modeled clinical gain and harm.
- harm: Net harm: -0.392 Bay QALYs. 10% weight, per annual-budget equivalent.
- cautious: Net harm: -0.00925655334437371 Bay QALYs. 20% weight; no positive finite price.
- central: $104,277,379 per 10 Bay QALYs. 20% weight. Road Q 0.49738583828480576 + activity Q 0.06948 − harm 0.1, then funding .4 and Bay .98.
- favorable: $1,900,531 per 10 Bay QALYs. 10% weight; 10.041379927965206 finite Bay QALYs per budget equivalent.
- without-favorable: No positive finite price. Other worlds renormalized: -0.004943365497051079 Bay QALYs.
- no-road: $163,555,157 per 10 Bay QALYs. No corridor benefit; activity and whole cost/harm remain.
- no-activity: $24,933,077 per 10 Bay QALYs. No physical-activity benefit; safety and whole cost/harm remain.
- zero-harm: $17,080,586 per 10 Bay QALYs. Removing all adverse-burden priors still does not reach $1M.
- half-residual-risk: $34,189,261 per 10 Bay QALYs. Halve the residual-risk fraction in every world; no other inputs changed. This is a stronger existing-project baseline, not a measured revised effect.
- full-cash-response: $13,451,413 per 10 Bay QALYs. Set nonnull-world funding response to1; funding-null remains0.
Uncertainty. Whole-cost partial-health model, not complete portfolio expected value. Favorable share 100.445% exceeds signed total because other worlds subtract health; it is not a probability. No declared world is below $1M/10 Bay QALYs, but this does not establish zero real-world probability. Stronger injury data improve the exposure anchor, not donor attribution. Future technical implementation, education activity and utility remain uncertain. Unquantified other programs could materially change the conclusion.
5. Funding and previous grants
Not a competitive giving recommendation on this evidence. Need a concrete additional technical task, design-specific residual risk and a credible cash response. Historical staff giving is a discovery lead, not GiveWell endorsement.
2024 revenue $1,845,171; deficit $52,605; government contributions $684,673; reported noncash contributions $54,983; net assets $732,835; liabilities $115,390; cash/savings $635,662. These are historical, not verified 2026 unrestricted capacity. Current financials page links a 2025 impact report, but its PDF was not retrieved in this pass; latest original filing found was2024. No current priced extra staff/design task verified. Agency capital, maintenance, partner staff and donated time are not fully monetized; completeResourceCostPer10 is null. No donor receives all corridor benefits or free construction. Official site has donation navigation, but checkout/restriction terms not inspected; donationUrl omitted.
This review does not establish a verified marginal funding offer or a complete history of grants.
We have not verified a suitable donation route for this reviewed activity. Confirm the legal recipient and intended allocation before donating.
6. Sources
- Staff personal donations: Devin Jacob entry. GiveWell Blog. Primary personal-giving statement. Published: 2018-12-10; retrieved: 2026-09-11.
- Financials and legal-name disclosure. Bike East Bay. Primary. Published: Undated current page; retrieved: 2026-09-11.
- East Bay Bicycle Coalition original 2024 Form 990. IRS via ProPublica. Primary. Published: 2024 reporting year; 2025 filing; retrieved: 2026-09-11.
- Organization-hosted 2024 tax return. Bike East Bay. Primary. Published: Prepared November 2025; December upload path; retrieved: 2026-09-11.
- San Pablo Avenue Safety and Access Project. CCTA. Primary. Published: Undated current 2026 project page; injury period July 2020–June 2025; retrieved: 2026-09-11.
- Multimodal San Pablo Avenue campaign. Bike East Bay. Primary. Published: Undated current page, includes July 2026 completion update; retrieved: 2026-09-11.
- Road Diet Desk Reference: evidence and tradeoffs. FHWA. Official evidence synthesis. Published: Undated official technical summary; retrieved: 2026-09-11.
- Evaluation of Lane Reduction Road Diet Measures. FHWA. Primary. Published: 2004; retrieved: 2026-09-11.
- School cycling randomized trial. Aranda-Balboa et al.. Primary. Published: 2022-08-05; retrieved: 2026-09-11.
- Bicycling to school and cardiometabolic risk randomized trial. Østergaard et al.. Primary abstract. Published: 2012-10-31; primary abstract accessed through Europe PMC; retrieved: 2026-09-11.
Annual expenses: years and sources
Average annual expenses (three consecutive fiscal years): $1,722,804. Organization size is separate from the modeled cost-effectiveness of a donation.
East Bay Bicycle Coalition
Form 990/990-EZ reported whole-entity expenses; includes program, administration and fundraising costs, but excludes any costs netted against revenue.
Latest original return and prior-year comparative combined with earlier filing data.