Records · Evidence review

Impact review: Inequality — the actions on /inequality/do/ and /inequality/practice/

The working record behind the pages, published as written. Source file source/impact/inequality-2026-11-05.md. All records.

Reviewed 2026-11-05 (research run 20:55–21:35 EST). Scope: the evidence behind every action the two pages recommend, the best-evidenced things one person can do that the pages leave out, who benefits and who might be harmed, what the evidence says about doing these things with other people, and the standing of every organisation the pages name. The content pages were not edited; this file is the review. Isaac decides what changes.

Decision being made: which actions the inequality Do and Practice pages recommend, in what order, with what wording. Population: one US person (the pages lean NYC), no special position, a few hours a month, a modest budget. What is being estimated: the effect of each action on (a) the poverty or wellbeing of the people it touches and (b) the distribution itself (inequality), per hour or dollar the reader puts in. The two are graded separately because the evidence separates them: the actions with the best evidence move poverty; the actions that could move the distribution have the thinnest evidence for one person.

Grades, fixed before reading:

• strong — at least one well-powered randomised trial, or several quasi-experimental studies, covering the action's whole causal chain for the affected people.
• moderate — causal evidence for part of the chain, or consistent administrative/quasi-experimental evidence, with a known gap.
• weak — descriptive, correlational, self-reported, or a mechanism with a plausible but unmeasured link to the outcome.
• none-but-plausible — no outcome evidence at all; the mechanism is direct or definitional.
• contradicted — well-identified evidence that the claimed effect is absent or negative.

Source grades, fixed list: primary (the agency, journal, dataset or court that produced the number), research (peer-reviewed or working paper), charity's own (an organisation's figures about itself, including city press releases about city programmes where noted), news, industry (trade bodies, vendors, projections), unread (a search-result snippet; the page was not opened). Every figure below carries one. Nothing here is from memory; where a figure could not be fetched today it is marked and not used.

**Mapping to the site's <span class="grade"> vocabulary** (CLAUDE.md): primary → primary; research → research (NBER/IZA working papers → preprint); GiveWell and Charity Navigator → evaluator; charity's own → "GiveDirectly's own figures", "NYC's own figures", etc.; news → news; industry → industry research or industry projection as fits; unread → unread; this file's own sums (fees kept per volunteer-season, refund per return) → our arithmetic. Action grades map one-to-one onto <span class="evidence"> (strong, moderate, weak, none but plausible, contradicted).

How this was done: four research streams. Three ran as parallel agents (EITC/VITA/CTC; cash transfers; civic/unions/housing/group formats) and reported sources they opened; I re-opened the keystone sources the pages already lean on (IRS participation and EITC statistics, GiveWell's GiveDirectly page, Egger et al., Vivalt et al., Baby's First Years, Census 2023, NYC DCWP) and ran the organisation checks myself. The session's shared web-search budget ran out partway through; everything after that was fetched by known URL. The gaps that caused are listed in section E, not papered over.

---

0. The answer first

Top five for one person, ranked by evidence that the action changes outcomes for the people it touches

Reduce inequality vs reduce poverty. Items 1, 3, 4 and 5 move poverty or wellbeing for identifiable people, with the best evidence. Only item 2 (and the union clause of 5) touches the distribution, and there the evidence for one person's marginal effect is the weakest on the list. The pages already say this honestly ("You can't change that from your kitchen"); the ranking should say it too.

Drop, keep, re-rank

Group formats: what the evidence supports

• Works: a match on giving (RCT: +22% response, +19% revenue; 2:1 and 3:1 no better than 1:1 — Karlan & List 2007); after-the-fact recognition of a completed gift (RCT: +2.7 points any gift, +14% gift size — Karlan & McConnell 2013); a scheduled, trained, supervised shift (VITA: every return gets a quality review; volunteers 76,314 in FY2025 → 84,500+ in FS2026, IRS).
• Contested: matching's long-run effect (Meier 2007: net negative participation afterwards; Huck & Rasul 2011: announcing a lead gift beat matching on donations given).
• Self-report only: giving circles (63% of members say they give more since joining; ~4,000 groups, 370,000 people, $3.1B 2017–23, imputed — In Abundance 2024); congregations (highly religious adults volunteer 45% vs 28%, 23% vs 4% through a congregation — Pew 2016, correlational, "the causal arrow could point in the other direction").
• Backfires or fades: advance pledges (<1% of blood-donation pledges fulfilled — Meyer & Tripodi 2021; "pledges are often reneged later, resulting in no overall change" — Capra et al. 2022); Giving What We Can's 10% pledge: ~60% recording in year one, ~25% of direct pledgers by year five, ~7% over the pledge lifetime (GWWC's own 2025 evaluation); peer endorsement reduced click-throughs to volunteering in a 100,000-student RCT (John et al. 2019).
• No comparison-group evidence found: mutual aid groups; tenant unions as such (one panel study: ten more community nonprofits per 100,000 residents ↔ 10% fewer eviction filings — Messamore 2023); faith-hosted VITA sites specifically.
• Reading of the pattern: structures that make the act automatic or witnessed after it happens (a match, a recognised gift, a rota) have evidence; promises made in advance do not. For the Practice page: set the monthly gift once, book the shifts, skip the pledge.

Organisation flags

• IRS VITA/TCE — operating and growing; grants $53M for 2026 (315 VITA, 48 TCE), 2027 notice $46M estimated. Flag: no independent (TIGTA/GAO) accuracy audit since 2016–17; the 98.5% is the IRS's own announced reviews. IRS lost 27% of its workforce in 2025; no VITA-specific cut found.
• NYC Free Tax Prep (DCWP) — operating; 2027 season recruiting; 2026 season 100,216 returns, $188.8M refunds, $37.8M fees saved (city's own). Flag: returns fell from 110,000 (2025); the city gives no reason. The nonprofit site operators are not named on the pages; not checked.
• GiveDirectly — operating; FY2024 revenue $206.8M, expenses $130.4M (990); Charity Navigator 4-star 98%. Flags: DRC 2022 fraud, $1.2M (0.8% of $144M delivered), disclosed June 2023 — outside the 24-month window; 2025 USAID cuts ended its US-government pure-cash programmes (news). Own "87%" efficiency vs GiveWell's 71–85%: both kept. Its research summaries overstate the papers in at least three places (harvest value "+50%/+20%" vs the article's +35%/+16%; "−29% inequality" for the Kenya UBI trial, whose text reports no significant Gini change; the 48% infant-mortality figure attributed to two different studies on one page). Cite the papers, not the summaries.
• GiveWell — FY2024 revenue $269.5M; Charity Navigator 4-star 100% on one of four beacons only.
• Oxfam (data source only) — Oxfam America FY to Mar 2026 filed Aug 2026, CN 4-star 96%. Flag (Oxfam GB, in window): CEO Halima Begum left Dec 2025 and filed a tribunal claim 3 Feb 2026 (Devex — snippet only, article blocked); board-commissioned review by Rebecca Tuck KC found trustees "acted in good faith and within their powers" (Oxfam GB, 7 Jul 2026); new chair 9 Oct 2026; ethnicity pay-gap reporting 26 Oct 2026. No Charity Commission inquiry found (register page blocked).
• Hand in Hand (templates link) — site live. Flag: legal entity not stated on its site and not found on ProPublica under either name; fine for a templates link, verify before any donation link.
• Western Queens CLT — 501(c)(3) since Nov 2024, 990-N (gross receipts under $50,000) for 2024 and 2025. Flag: the page's "strong for the people housed" grade does not describe WQCLT, which has not yet housed anyone.
• NDWA (not on the page; Hand in Hand's core partner) — FY2024 revenue $48.6M; CN 4-star 97%. Flag (in window): 28 union staff laid off Sept 2025 (about a third of the bargaining unit); 93% strike authorisation 17 Nov 2025 (CWA, union's account); outcome not found.
• Certified EO is an LLC selling certification; USFWC operating (revenue $1.96M, 2024). Government lookups (NCUA, DOJ, FTC, NY Assembly) were fetched in today's research run.

---

A. The evidence behind each action on the pages

A1. Volunteer at a free tax-preparation site (VITA / NYC Free Tax Prep) — grade: moderate (families served) / weak (take-up)

What the pages claim. Do: "About one in five eligible workers never claim the Earned Income Tax Credit… Volunteers file it for them, and you can count every return." Practice: "Each return you finish for a family that qualifies for the EITC is, on average, a few thousand dollars they wouldn't otherwise get" (source note: "$2,700, recalled rather than re-read").

The participation gap is real. IRS, EITC participation rate by state (reviewed 19 Aug 2026, primary): 80.8% nationally for TY2022, New York 83.7%; "Approximately 19.2% of all eligible earned income taxpayers in TY2022 didn't claim" — modelled from ACS linked to IRS records. The CPS-based estimate is 78% (IRS–Census CES-24-75, Dec 2024, primary: "eligible taxpayers participated in the EITC program at a rate of 78 percent" for TY2021; dollar participation 81%). Two methods, two numbers: keep both. Average credit: $2,894 nationally, $2,753 in New York, TY2024, 24M claims, $69.6B (IRS EITC statistics, reviewed 28 Jul 2026, primary). This replaces the Practice page's unverified "$2,700".

Who the missing fifth are. Bhargava & Manoli (AER 2015, research, RCT, 35,050 filers who had not claimed): "of the 25 percent who do not take-up, 16 percent do not file taxes." Linos et al. (AEJ: Economic Policy 2022, research): "Twenty-five percent of nonfilers are eligible for credits of $300 or less, as compared with 9 percent of filers." Most of the gap is people who file no return at all, and their credits skew small.

Does steering people to free tax prep raise claiming? This is where the page's chain is weakest.

• Linos, Prohofsky, Ramesh, Rothstein & Unrath 2022 (research): "six preregistered, large-scale field experiments with 1 million observations" in California, arms including "Local in-person free tax preparation information" and the "Address of local in-person free tax preparation assistance": "we find no evidence that they affected households' likelihood of filing a tax return or claiming the credit… In Studies 1, 3, and 4, we can rule out effects of 0.5 percentage point or larger." Their own reading: "actually receiving that help would have involved seeking out that assistance, going to a VITA site at specified times, compiling all of the required documents… Expanding the availability of this type of free tax support by extending the hours and availability of VITA sites… may be crucial."
• Goldin, Homonoff, Javaid & Schafer (NBER w28398, 2021, research; JPubE 2022 version not opened): IRS letters to 56,015 non-filers near VITA sites. Among delivered letters: filing +0.74 points (+3.5%), EITC claiming +0.32 points (+7%); "approximately 43%" of induced filers claimed the EITC, averaging $861; no significant shift away from paid preparers. Effects did not persist into the next year (consistent with Guyton et al. 2017, research: reminders +1 point per J-PAL's summary, "about 0.5 percentage points" as Linos et al. describe it; either way "did not persist").
• Bhargava & Manoli 2015: simplifying the notice and displaying the benefit raised claiming (0.23 → 0.31); messages aimed at stigma, audit fear or application cost did not — one stigma message reduced claiming 18%.
• Code for America's own RCTs (charity's own, embedded randomisation): among prior clients, simplified self-filing completed at 2.38% vs 0.30% sent to virtual VITA (2022, n=23,778); adding a help-hotline number "depresses submission rates by about 20%".

So: information and referral to VITA have no measured effect on claiming; what moves take-up is simplification and hands-on assistance. The assistance evidence comes from other programmes: Finkelstein & Notowidigdo (QJE 2019, research, RCT, ~30,000 elderly Pennsylvanians): SNAP enrolment 6% control, 11% information only, 18% information plus phone application assistance, ~$60 per enrollee. Bettinger et al. (QJE 2012, research, RCT, 26,162 in H&R Block offices): completing the FAFSA with the family raised college enrolment 8.1 points (from 34.2%); information alone did nothing. Against this, Bird et al. (NBER w26158, 2021, research, two RCTs, 800,000+ students): one-on-one advising "no impacts on financial aid receipt or college enrollment." Keep the contest: assistance works when it does the form with the person; it is not guaranteed.

What VITA demonstrably does. IRS Publication 5347, Sept 2026 (primary): "During the 2026 filing season… over 9,000 VITA/TCE sites, with more than 84,500 volunteers… over 2.9 million federal tax returns." Accuracy 98.47% weighted (293 of 298 returns reviewed), 90% CI 96.81–99.40% — but "representative only of VITA/TCE sites… that prepared at least 50 tax returns during Filing Season 2025," and TIGTA's 2016 audit (primary) notes "For most of the site visits and return reviews, volunteers are aware that a QSS review is being conducted." TIGTA's anonymous shopper audits found 41% of returns inaccurate in 2009, 61% in 2011 (14 of 36 correct), 51% in 2012 and 49% in 2013 (20 of 39 correct) — small judgemental samples (TIGTA 2013-40-110, primary; AccountingToday 2009–2013, news). No TIGTA or GAO VITA accuracy report after 2017 was found. Keep both numbers on the page: the IRS's 98% is announced-review accuracy; the last independent anonymous tests were far lower and are a decade old. VITA/TCE prepared 1.8% of domestic individual returns in TY2024 (National Taxpayer Advocate 2025 report, primary). NYC: "In just one tax season, you'll complete about 40 tax returns" (DCWP volunteer page, updated 10/2026, primary).

The fee the family keeps. NYC DCWP: average New York preparation fee "$346" (FAQ, updated 03/2026, primary); 2026 season $37.8M saved on 100,216 returns (Mayor's office, 30 Sep 2026, charity's own for a city programme). NACo deck (Nov 2023, industry): 51% of EITC claimants use a paid preparer; average fee $220. PPI 2016 (research, think-tank mystery shop, two cities): chain fees $309–$509 for EITC filers, "between 13 and 22 percent of the average EITC refund." Forty returns × $346 ≈ $13,800 in fees kept per volunteer-season, plus refunds delivered on time (NYC 2026: $188.8M / 100,216 = ~$1,884 per return, our arithmetic).

Verdict. Keep the action; it is the most countable thing on the page and the benefit per return is certain. Change the claim. The honest sentence: "Each return you finish is a family's refund filed free and right; in New York the fee they keep averages $346, and the average EITC is about $2,900. The one-in-five who miss the credit are mostly people who file nothing at all; sites that bring them in need greeters, interpreters and evening hours as much as preparers."

A2. Pay the person who works in your home as the law requires — grade: strong by construction (that worker); none (inequality)

No study is needed for the direct effect and none exists for the indirect ones: paying the legal rate, overtime, paid leave and Social Security is a transfer to one worker plus future benefit eligibility. The legal facts on the page (NY DOL, NYS Labor Law §170, IRS Pub 926, DCWP leave FAQ) were fetched in today's research run and are not re-graded here.

Context for "who benefits": BLS Occupational Outlook Handbook (May 2025, primary): childcare workers median $16.82/hour (1,010,300 jobs); home health and personal care aides $17.21 (4,677,100 jobs; 760,500 openings a year); janitors and cleaners $17.71. Caveat the page should carry if it uses these: BLS pay data "does not include… workers in private households because these data are not collected by the Occupational Employment and Wage Statistics (OEWS) survey" — the very workers this action is about are missing from the official wage data. The EPI domestic-workers chartbook (poverty rates, benefits) could not be opened (403, unread). Share of household employers who pay off the books, and any evaluation of the Domestic Workers' Bill of Rights's effect on wages: not researched (search budget).

Possible harm, unmeasured: formalising adds about 11% (page's own arithmetic) and may lead some employers to cut hours; some workers prefer cash for immigration or benefit reasons. These are reasoning, not evidence; the page already frames the action as a floor the employer controls, which is right.

A3. Give cash directly (GiveDirectly) — grade: strong for recipients; contested on spillovers and the multiplier

The trial the page cites. Egger, Haushofer, Miguel, Niehaus & Walker, NBER w26600 (Dec 2019, rev. Oct 2021; Econometrica 90(6), Nov 2022, research): "one-time cash transfers of about USD 1000 to over 10,500 poor households across 653 randomized villages in rural Kenya. The implied fiscal shock was over 15 percent of local GDP… large impacts on consumption and assets for recipients… large positive spillovers on non-recipient households and firms, and minimal price inflation. We estimate a local transfer multiplier of 2.4." The page's "$2.40 per dollar, little inflation" matches the abstract. GiveWell's summary of the same paper (research-evaluator): "statistically significant, but economically minimal, local price inflation"; non-recipient consumption 13% higher; 5–7-year preliminary data show non-recipients still ~8% higher but not significant (GiveWell gives it little weight).

The contest the page does not carry. GiveWell (primary for its own assessment, Nov 2024, updated Oct 2025): "Four of six" studies measuring spillovers "find null or slightly negative effects"; Haushofer & Shapiro (2016, 2018) found eligible non-recipients' consumption 4% lower at nine months and 16% lower at three years; McIntosh & Zeitlin (2022) no significant spillovers; Banerjee et al. (2023) could not reject zero; Egger et al.'s results "stand in stark contrast." GiveWell's reasons to weight Egger more: Haushofer & Shapiro collected no data on ineligible households (Berk Özler's point), so their spillover estimate is partial. Elsewhere, Filmer et al. (2023, government cash transfers in the Philippines): perishable food prices up 6–8% in saturated remote villages and stunting "among nonbeneficiary children" up 11 percentage points (34%) — a different programme, context and saturation level, but it is the clearest evidence that cash can hurt neighbours.

Mortality. Walker et al. (2025, via GiveWell): under-5 mortality "45% decline relative to control" during the transfer years, none for children born after transfers ended; GiveWell applies a 50% internal-validity haircut to 23% and says it "cannot identify a mechanism that rationalizes an effect this large."

Cost-effectiveness and efficiency. GiveWell: "~30-40% as cost-effective as our marginal funding opportunity"; "~3-4x more cost-effective than we had previously estimated"; transfers as a share of total cost: Kenya 84%, Malawi 84%, Mozambique 75%, Rwanda 85%, Uganda 71%; transfer ~$1,000 per household (Kenya currently ~$865). GiveDirectly's own page (charity's own, undated): "$87 of every $100 you give goes to someone in extreme poverty as cash" for its Poverty Relief programmes; its own footnote says the 990-based figure is lower. Both numbers are right about different denominators; keep both as the page does.

Fraud. GiveDirectly DRC case page (charity's own, 5 Jun 2023, updates to 24 Sep 2024): "just under $1.2M was illegally diverted from our D.R.C. operations in 2022", "0.8% of the $144M we delivered globally last year"; 1,900 families lost five months of payments and "are in the process of being paid back"; $11.5K recovered. GiveWell applies a −2% fraud/wastage adjustment. Outside the 24-month window, but the page should say "fraud losses run about 1% a year, publicly disclosed."

US cash: the sobering half. Vivalt, Rhodes, Bartik, Broockman, Krause & Miller, NBER w32719 (rev. Aug 2026; QJE 141(4) 2026, research): 1,000 low-income adults given $1,000 a month for three years vs 2,000 controls at $50: labour-market participation −4.2 points, 1–2 fewer hours a week, individual income excluding transfers −$1,900 a year, "no impact on quality of employment", wellbeing up in year one then back to control. Baby's First Years, NBER w33844 (May 2025, research): $333 vs $20 a month from birth, n=891 at age four, "no statistically significant impacts" on four preregistered primary outcomes (language, executive function, social-emotional problems, brain activity) or three secondary ones. The other US pilots (Stockton, Denver, Chelsea, Compton), the later Baby's First Years papers, and the cash-vs-in-kind and labour-supply literature are in A3-bis.

Verdict. Keep; promote to step 1 on evidence grade; add one sentence of contest ("one very large trial found neighbours gained too; four smaller ones found they did not, or lost a little") and one on fraud. Keep the GiveWell pointer for readers who want lives per dollar: Malaria Consortium ~$4,000 per life saved, Helen Keller ~$3,500, New Incentives ~$4,500, Against Malaria ~$5,500 (GiveWell top charities, updated Oct 2026, research-evaluator; GiveWell's general range "$3,000 and $5,500", Apr 2024 page).

A3-bis. Cash transfers: the wider literature and the US pilots

**GiveDirectly's other randomised trials (all research unless noted).**

• Haushofer & Shapiro, QJE 2016 (author-hosted copy): Rarieda, Kenya, 2011–13; mean transfer $709 PPP ("almost two years of per capita expenditure"); at nine months monthly consumption "$158 PPP to $193 PPP"; assets +$302 PPP on a base of $495; happiness +0.18 SD, stress −0.23 SD; "no effects on health and educational outcomes"; monthly instalments improved food security more, lump sums bought durables.
• Haushofer & Shapiro, three-year follow-up (Jan 2018 working paper, 503 treatment households): within villages, "a 40 percent increase in asset holdings (USD 416 PPP)… a 25 percent increase in consumption (USD 47 PPP)" that "do not significantly decrease over time"; across villages (recipients vs distant controls) only assets persist ("USD 422 PPP… equivalent to 60% of the initial transfer"), and "spillover households spend USD 30 PPP less than pure control households, or about 16%… ~0.25 SD less on an index of food security… ~0.18 SD less on… psychological well-being"; the authors "speculate it could be due to the sale of productive assets by spillover households to treatment households." The consumption and food-security spillovers survive attrition bounds. This is the main counterweight to Egger et al.; keep both.
• Banerjee, Faye, Krueger, Niehaus & Suri, Universal Basic Income: Short-Term Results from a Long-Term Experiment in Kenya (15 Sep 2023 working paper; two authors are unpaid GiveDirectly directors, disclosed): 295 villages; long-term UBI ($0.75 a day for 12 years, 44 villages), short-term UBI (two years, 80), lump sum (~$500, 71), control (100). "Communities receiving UBI experienced substantial economic expansion—more enterprises, higher revenues… Labor supply did not change overall, but shifted out of wage employment and towards self-employment… The lump sums had similar, if not larger, economic impacts, while the short-term transfers had noticeably smaller effects." Lump-sum income effect "50% of the control group income"; consumption +6–11% of the control median in all arms; prices: "we do not reject the null that consumer prices were unaffected"; alcohol self-reports "approximately zero"; all arms reduced depression; "no significant changes in either the 75-25 ratio or the Gini coefficient for either wealth or income from any of the interventions." GiveDirectly's summary page (charity's own) headlines "−29% inequality" for this study; the paper's text does not support it. Keep the disagreement; it is a direct inequality claim.
• Aggarwal et al., AEJ: Applied 2026 (GiveDirectly-hosted copy of the article): 600 villages in Liberia and Malawi, ~6,000 households, $250–$750; harvest value "+35 percent in Liberia and 16 percent in Malawi"; farm profits +33%/+15%; food security +0.25 SD persistent in Liberia, fading to 0.11 SD in Malawi; casual labour hours −45% of the control mean in Liberia; intimate-partner violence −8 points in Liberia, none in Malawi. GiveDirectly's research page says "+50% in Liberia and +20% in Malawi" — the article says +35%/+16%.
• Walker, Shankar, Miguel, Egger & Killeen, NBER w34152 (Aug 2025, rev. Jan 2026): the Egger villages; "48% fewer infant deaths before age one", "45% fewer child deaths before age five", via a "53% reduction in labor supply around childbirth" and "45% increase in hospital deliveries"; no effect for children born after transfers ended. (GiveWell's adjusted figure: 23%.)
• Haushofer, Ringdal, Shapiro & Wang, NBER w25627 (2019): transfers to women cut physical and sexual violence (−0.26, −0.22 SD); to men, physical violence −0.18 SD; non-recipient women in treatment villages −0.16 SD physical violence.
• US, one-time: Pilkauskas, Jacob, Rhodes, Richard & Shaefer, COVID Cash Transfer Study (May 2022 working paper): $1,000 from GiveDirectly to SNAP-connected families in May 2020, powered for 0.09 SD: "no statistically significant effects… on any of our five pre-specified primary outcomes" (hardship, mental health, partner conflict, child behaviour, parenting); hardship fell among those with under $500 of earnings the prior month. (A second US one-time trial, Jaroszewicz et al., exists; SSRN blocked — unread.)

The critical literature.

• Labour supply, developing countries — Banerjee, Hanna, Kreindler & Olken, World Bank Research Observer 2017: seven government-programme RCTs in six countries re-analysed; pooled effect of a transfer worth 13.6% of consumption: employment "a decrease of 0.4 percentage points from a base of 56 percent", hours "a decrease of 5 minutes of work per week, from a base of 21 hours"; "with 95% probability, a new program has an impact no lower than a 2.3 percentage points reduction." Bastagli et al., ODI 2016 (165 studies): "for just over half of studies on adult work… the cash transfer does not have a statistically significant impact. Among those… reporting a significant effect… the majority find an increase."
• Temptation goods — Evans & Popova, EDCC 2017: 19 quantitative studies; "cash transfers have a significant negative effect on total expenditures on temptation goods, equal to -0.18 standard deviations."
• Prices and neighbours — Filmer, Friedman, Kandpal & Onishi, REStat 2023 (Philippines Pantawid, government CCT): higher perishable-food prices "in some markets"; stunting among non-beneficiary children "by 11 percentage points (34%)"; "Impacts increase in the size of the village income shock and remoteness"; median village income shock 15%. Cunha, De Giorgi & Jayachandran, REStud 2019 (Mexico PAL, 208 villages): in-kind food "a 4 percent fall in prices while cash transfers lead to a positive but negligible increase in prices", larger in remote villages; "The pecuniary benefit to consumers of in-kind transfers, relative to cash transfers, equals 11% of the direct transfer." Egger et al. (Kenya): minimal inflation at a 15%-of-GDP shock. Banerjee et al. (Kenya UBI): cannot reject no price effect. The contest is about remoteness and saturation, not about whether cash "causes inflation" in general.
• Fade-out — Haushofer & Shapiro (assets persist, consumption contested); Aggarwal et al. (Malawi food security fades to 0.11 SD); Egger et al. 5–7-year non-recipient gains ~8%, not significant (via GiveWell).

Cash versus in-kind, and paternalism.

• Cunha, AEJ: Applied 2014 (Mexico PAL, 208 villages; food box vs equal-value cash vs nothing): "households do not indulge in the consumption of vices when handed cash… the majority of the transfer is spent on nutritious food"; the box was "infra-marginal and nondistorting" for total food; "Small differences in the nutritional intake of women and children under in-kind transfers did not lead to meaningful differential improvements in health outcomes compared to cash."
• Hidrobo et al., JDE 2014 (Ecuador; cash vs food vs voucher): "all three modalities significantly improve the quantity and quality of food consumed", food raised calories most, vouchers dietary diversity most; magnitudes and costs not on the pages opened.
• McIntosh & Zeitlin, Gikuriro benchmarking (arXiv 2106.00213, 2021): a $142 child-nutrition programme "did not lead to improvements in consumption, dietary diversity, wealth, child anthropometrics, or anemia within the thirteen-month period"; cost-equivalent cash raised consumption, paid down debt and bought assets; a $567 transfer produced "a drop in child mortality, and modest improvements of about 0.1 standard deviation" in child growth; "the benefits of flow or in-kind transfers are not larger for households that report self- or other-control problems." Cost figures differ across versions ($124/$517 on the arXiv abstract; $142/$567 in the body; "about $110"/"about $532" on GiveDirectly's page) — kept as a discrepancy. Huguka Dukore (arXiv 2209.08574, 2022, 3.5-year results): training improved time use and assets, cash improved assets, livestock, savings and wellbeing; "Both interventions have powerful effects on entrepreneurship… the implied wage rate in these businesses is low." The one-year "cash proves superior" line is GiveDirectly's wording (charity's own), not the paper's.
• Gentilini, World Bank Research Observer 2016 (ten-country review): "relative effectiveness cannot be generalized"; "Costs for cash transfers and vouchers tend to be significantly lower relative to in-kind food."
• Reading: the paternalist worry (cash gets wasted) is contradicted across settings; the in-kind case rests on specific nutrition goals and on local prices in remote markets, not on recipients' choices.

US guaranteed-income pilots. Grades matter here: several headline findings are programme-published (charity's own) and the peer-reviewed versions are quieter.

• Stockton SEED. Programme report, March 2021 (charity's own, written by its evaluators): full-time employment "28% of recipients… One year later, 40%" vs control "32%… 37%" — reported descriptively, with the caveat "these findings cannot point to larger labor market trends"; income volatility 46.4% vs 67.5%; Kessler-10 p=.056; "Less than 1% of tracked purchases were for tobacco and alcohol." Peer-reviewed version, West & Castro, Journal of Urban Health 2023 (research): 110 treatment / 198 control in the intention-to-treat sample; the year-one employment shift "was not statistically significant" (full-time not reported separately); income volatility year one 19% vs 26% (p=0.039), year two not significant; mental-health gains significant before the pandemic, none after withdrawal; differential retention (~55% vs ~35%). The 28→40 headline the US debate still quotes is the descriptive one; the peer-reviewed paper does not support it. No later final report exists (CGIR publications page).
• Denver Basic Income Project, Year One report, June 2024 (charity's own, University of Denver evaluators): 807 enrolled; $1,000/month vs $6,500 then $500/month vs $50/month "active comparison group"; "housing outcomes improved at a similar rate for participants in all three groups"; own house or apartment 6%→44%, 6%→48%, 12%→43%; significance marked within groups only. Groups A and B did better on paying bills and financial wellbeing. A Year Two report could not be found on the project's research page; the earlier claim of one is unverified.
• Chelsea Eats (Massachusetts; up to $400/month for nine months, lottery, ~2,000 households). Working paper, Dec 2022 (research): higher food spending and satisfaction, lower financial distress; "the decline in food insecurity… did not persist"; "no statistically significant treatment-control differences in… self-reported physical/mental health and child school attendance"; "did not reduce employment or hours of work." Agarwal, Cook & Liebman, JAMA 2024 (research, 1,746 cash / 1,134 control): emergency-department visits 217.1 vs 317.5 per 1,000, adjusted −87.0 (95% CI −160.2 to −13.8; P=.02), a 27% relative decrease; no effects on blood pressure, weight, HbA1c, cholesterol or vaccination. The clearest positive US health result.
• Compton Pledge, Balakrishnan, Chan, Constantino, Haushofer & Morduch, NBER w33209 (Nov 2024, research, independent): 695 treated, ~$500/month for two years; "no impact on the labor supply of full-time workers"; baseline part-timers −13 points participation; non-transfer income −$333/month; "a significant improvement in housing security" but "no overall effects on indices of psychological and financial well-being."
• OpenResearch (Vivalt et al., A3): participation −4.2 points, income excluding transfers −$1,900, wellbeing gain faded, "no impact on quality of employment."
• Baby's First Years, beyond age four: EEG at age one showed "more power in high-frequency bands" (beta ES 0.26, P=.02) that did not survive multiple-testing adjustment (Troller-Renfree et al., PNAS 2022, research); spending on child-specific goods rose ~$68 a month with no offset in earnings (Gennetian et al., NBER w30379; J-PAL summary); income-to-needs +11% and learning activities +0.16 SD, no change in maternal stress, anxiety or depression at age three (Magnuson et al., Nature Communications 2025); at age four no effects on maternal mental health, maternal or child BMI, or child health (Duncan et al., JAMA Pediatrics 2025); a slower epigenetic pace of ageing in children, −0.17 SD (P=0.037), with no effect on the preregistered epigenetic-g outcome (Raffington et al., Nature Human Behaviour, Sept 2026). Age-eight data collection began July 2026; no age-six results published.
• Syntheses. Corinth & Mayhew, AEI, Feb 2026 (research, think-tank working paper): 122 pilots, 2017–25, $481M to 40,000+ recipients; "Among the 30 pilots that are randomized controlled trials with published employment outcomes, the mean effect of a GBI is an increase of 0.8 percentage points… Among the four such pilots with a treatment group of at least 500 participants… the mean effect on employment is -3.2 percentage points"; 20 of 30 pilots had 200 or fewer treated. University of Chicago Inclusive Economy Lab, Policymakers' Report, Oct 2025 (research, synthesis of three RCTs): "Guaranteed income leads to moderate declines in the labor force"; OpenResearch financial-security effects "peaked in year one, fell in year two, and faded by year three"; temptation-goods spending +$13/month (OpenResearch), none (BFY), tobacco slightly down (Compton). Peng et al., Indiana University systematic review, draft Jan 2025 (research): health effects "either temporary positive effects or no significant effects." Economic Security Project, Cash Works. So Do Its Recipients. (Sept 2026, charity's own, advocacy; ESP funded SEED): "Full-time work… increased substantially in nearly 60 percent of programs", mostly pre/post comparisons. Keep the three readings side by side: small pilots average a small employment gain; the four large RCTs average a 3-point loss; advocates count pre/post rises.

Alaska. Jones & Marinescu, AEJ: Economic Policy 2022 (research, synthetic control): "the dividend had no effect on employment and increased part-time work by 1.8 percentage points (17 percent)." Guettabi, ISER 2019 (via the university's write-up, news; PDF blocked): "substantial poverty reductions for rural Alaska Natives", shrinking over time; the dividend increases inequality "in both the short and long run"; obesity down among three-year-olds; "substance abuse related incidents increase while property crime related events decrease" after payout. The page's treatment is consistent with this.

What this adds to the page. The Learn and Do pages' line "Cash works in Kenya; in the US the trials are sobering" holds, with three refinements: (1) the Kenyan evidence is strong on recipients and contested on neighbours; (2) the US evidence is not uniformly null — Chelsea Eats cut emergency visits 27% and Compton improved housing security — but the large RCTs show modest work reductions and fading wellbeing gains, and the two most-quoted positive headlines (Stockton employment, Denver housing) do not survive their own peer-reviewed or between-group tests; (3) lump sums did at least as well as monthly payments in the Kenya UBI trial, so a reader's one-off gift is not a lesser form of the action.

A4. Write a personal message to a state legislator on a named bill — grade: moderate (state); weak (federal)

The page's source is the Congressional Management Foundation's 2017 Citizen-Centric Advocacy (primary for its own survey): 2015 survey of congressional staff (n≈190–233; 233 responses Aug–Oct 2015). Share saying a contact has "some" or "a lot" of positive influence on an undecided member: in-person constituent visit 94%, individualised email 92%, individualised letter 88%, phone call 84%, form email 56%. This is federal staff's perception, not measured behaviour; 47% of staff "believe that 'most advocacy campaigns of identical form messages are sent without the constituent's knowledge or approval.'"

Field experiments, all at state level:

• Bergan & Cole, Political Behavior 2015 (research, abstract read; full text paywalled): "Michigan state legislators are randomly assigned to be contacted by their constituents about a specific bill or to a control group… increases the probability of supporting the relevant legislation by about 12 percentage points." Authors' write-up (LSE blog, 2015): legislators assigned 22, 33 or 65 calls; "the number of calls did not matter beyond the effect of simply being contacted at all"; "Anti-bullying is a relatively low salience issue… The effects… would presumably be smaller for more highly publicized and polarizing issues."
• Butler & Nickerson, QJPS 2011 (research, abstract read): 10,690 New Mexicans surveyed; district results shared with a random half of the legislature; "legislators receiving their district-specific survey results were much more likely to vote in line with constituent opinion." Treatment is information about opinion, not contact; magnitude not read.
• Bergan 2009 (New Hampshire email campaign): **unread** at source (SAGE blocked); a secondary post reports +20 points. Not used for the grade.
• Federal: Kalla & Broockman, AJPS 2016 (research, RCT, 191 House offices): donors got a meeting with the Member or chief of staff 12.5% of the time vs 2.4% for "constituents" — access, not votes; the one cosponsorship outcome was null. No field experiment on calls or emails changing a federal roll-call vote was found.

Verdict. The page under-sells this. It is cheap, it is the only action with experimental evidence of changing a rule, and the effect (≈12 points on one legislator's vote on a low-salience bill) is large relative to the cost. Target state legislators and the City Council, where the experiments were run; name the bill; write or call in your own words (form messages are the one format staff discount). Honest ceiling: the NY antitrust and champerty bills are low-salience, which is where the effect was measured; a millionaire-tax fight is not.

A5. Join or fund a community land trust; back COPA — grade: moderate for households housed; weak as a lever for one reader

Wang, Cahen, Acolin & Walter, Lincoln Institute WP19RW1 (Apr 2019, research, administrative data, 58 programmes, 4,108 properties): "Ninety-five percent of shared equity homes are priced affordably for families earning 80 percent of area median income (AMI) or below"; "only 3 percent of resales since 2007… unaffordable to households below 80 percent of AMI"; median household "accumulates approximately $14,000" on a median $1,875 equity investment; annual move rate 2.6% vs 6.9% for all homeowners; 58% buy again. Foreclosure: this paper reports none of its own; the page's 2011 Thaden figures (1.30% vs 8.57% seriously delinquent) were fetched in the research run and are a self-reported survey of 62 CLTs. About 44,000 units nationally (Lincoln census, 2023).

For one reader: becoming a "monthly sustainer" of Western Queens CLT (501(c)(3) since Nov 2024; 990-N, gross receipts under $50,000, FY2024 and FY2025, ProPublica primary; three projects "pursuing", none completed, wqclt.org charity's own) is a small bet on an early-stage group, not a contribution to the measured outcomes above. The measured outcomes belong to mature CLTs with housing stock. Keep COPA/Intro 650 as the legislator ask (A4 evidence); grade the WQCLT line honestly.

A6. Move your money to a credit union — grade: none-but-plausible (inequality); the page already says "weak as a lever"

No outcome evidence was sought or found beyond the research run's NCUA rate comparison (primary), which is about the reader's own costs. Recommend moving to the Q&A.

A7. Buy from co-ops and employee-owned firms — grade: weak

Research run: meta-analysis r≈0.04 on performance; layoffs 1.9% vs 5.1%; NCEO's own caveat "correlation is not causation." No evidence a buyer's choice changes a worker's income. Recommend moving to Learn or the Q&A.

A8. Report anticompetitive conduct — grade: none-but-plausible

No evidence was found that individual complaints lead to cases; none was sought beyond the agencies' intake pages. Recommend dropping.

A9. "Tax" as an action

The page's tax actions resolve to A1 (claim what the law already gives) and A4 (ask for a different law). Both graded above. For completeness: refundable credits lifted 6.1 million people above the SPM line in 2025, 3.3 million of them children, cutting the child SPM rate 4.5 points; the EITC alone lifted 4.0 million (Census P60-290, Sept 2026, primary). The 2021 expansion: SPM child poverty 9.7% (2020) → 5.2% (2021) → 12.4% (2022) → 13.7% (2023) → 13.5% (2024, revised) → 13.4% (2025) (Census releases 2022–2026, primary; the 2019–2024 figures are pre-revision after BLS re-released SPM thresholds in July 2026 — re-check when Census publishes revised rates). Labour-supply contest on a permanent child allowance: Corinth, Meyer, Stadnicki & Wu (NBER w29366, simulation) 1.5 million exits; Ananat, Glasner, Hamilton & Parolin (NBER w29823, difference-in-differences on 2021) "very small, inconsistently signed, and statistically insignificant impacts." Different questions (a permanent programme modelled vs six temporary months observed); keep both.

---

B. The highest-impact things one person can do — the reasoning behind the top five

The ranking in section 0 weighs three things: grade of evidence that the action changes an outcome for the affected people; size of that change per hour or dollar for someone with a few hours a month and a modest budget; and whether the reader can actually do it (conditional actions rank below unconditional ones of equal grade).

1. Cash to the global poor has the only "strong" grade among unconditional actions and the clearest unit economics (71–87 cents of each dollar becomes a household's cash). It reduces poverty; its effect on inequality is confined to being the most progressive transfer available. The contests (spillovers, fade-out, fraud) are quantified and small relative to the recipient effect.
2. A personal contact to a state legislator ranks second not because the per-person effect is large in absolute terms but because it is the one action with experimental evidence of changing a rule, it costs fifteen minutes, and nothing else on the page touches the distribution at all.
3. VITA volunteering has a certain, countable, direct benefit (≈40 families, fee kept, refund delivered) and a credible but unproven take-up story. It is the best unconditional time action on the page.
4. Tenant legal representation has an RCT and a quasi-experiment with large effects, a measured per-household cost, and a measured funding gap in NYC; it belongs on the page and is not there.
5. Conditional actions (your household employee; your workplace union) are strong where they apply, which is why they are on the list, and conditional, which is why they are last.

Considered and left out of the five: volunteer tutoring and mentoring (RCT evidence — Reading Partners effect sizes 0.09–0.11 SD; Big Brothers Big Sisters 46% less likely to start drugs — but the outcome is education and youth development, not poverty or inequality as the page defines them); GiveWell top charities (stronger per-dollar evidence than cash, but on mortality — already pointed to from the cash step); voting and party work (no one-person evidence sought). Not researched: political donations to advocacy organisations.

---

C. Communal impact

C1. Who benefits, and how much, per kept action

C2. Harm and backfire

• Stigma in means-tested programmes. The best RCT finds complexity, not stigma, is the binding barrier: simplifying the notice and showing the benefit raised EITC claiming from 23% to 31%; messages trying to reduce personal or social stigma did not help and one cut claiming 18% (Bhargava & Manoli 2015). Lesson for a volunteer: do the form, do not reassure.
• A wrong return hurts the family (repayment, audit, lost credit). The IRS's own figure is 98.5% accurate; the last anonymous audits (2009–13) found 39–59% on small samples; TIGTA 2017 found 10% of grantee returns carried schedules requiring advanced certification the IRS could not verify. Mitigations on the page: certify properly, let the site's quality review happen, or take a greeter/interpreter role.
• Cash to neighbours' detriment. Haushofer & Shapiro's negative within-village spillovers and Filmer et al.'s price and stunting effects are the two documented harms; Egger et al. found the opposite; "only six non-recipient villages out of 653 lost real income" (Walker et al., via GiveWell). Fraud ~1% a year.
• Cash in the US. Recipients worked less (−4.2 points participation) and wellbeing gains faded (Vivalt et al.); no child-development effects at four (Baby's First Years); across the four large RCTs employment −3.2 points on average, across 30 mostly small ones +0.8 (Corinth & Mayhew 2026). Against: Chelsea Eats cut emergency-department visits 27% (JAMA 2024); Compton improved housing security; the Kenya trials cut intimate-partner violence (−0.22 to −0.26 SD) and the Liberia trial by 8 points.
• Paternalism. The worry that cash is wasted is contradicted where tested: temptation-goods spending falls (−0.18 SD, Evans & Popova 2017); "households do not indulge in the consumption of vices when handed cash" (Cunha 2014, Mexico); OpenResearch +$13 a month, Stockton under 1% of card spend. In-kind beat cash only on narrow nutrition targets (calories in Ecuador; nothing in Rwanda's Gikuriro programme at equal cost) and on local prices in remote markets (Mexico: in-kind −4%, cash negligible). The one documented harm to third parties from cash is through prices in saturated, remote villages (Philippines).
• Form letters. 56% vs 92% in staff perceptions; "47 percent believe that most advocacy campaigns of identical form messages are sent without the constituent's knowledge."

C3. Doing it with others

Summarised in section 0. The grades, fully sourced:

• Matching gifts — Karlan & List, AER 2007 (research, natural field RCT, 50,083 prior donors): match "increases the revenue per solicitation—by 19 percent… probability that an individual donates—by 22 percent"; "larger match ratios… have no additional impact." Meier, JEEA 2007 (research, field RCT, University of Zurich funds): after the match ended, non-contribution rose 4.1 points (p<0.10); mean donations CHF 7.85 vs 8.20 (p<0.05): "a negative net effect on the participation rate." Huck & Rasul, JPubE 2011 (research, 14,000 opera attendees): lead-gift announcement €132 average vs 50% match €101 vs 100% match €92 vs control €74; "straight linear matching schemes… partially crowd out the actual donations given." Workplace scale: 26 million employees at matching companies, ~$2.86B a year matched, $4–7B unclaimed (Double the Donation, industry, unsourced on page); 21% average employee participation in matching, 25% in volunteering, "changed little in recent years" (CECP Giving in Numbers 2026, industry, 650+ large companies).
• Recognition vs pledge — Karlan & McConnell 2013 (research, RCT, 4,168 donors): newsletter recognition +2.7 points any gift, +14% size. Meyer & Tripodi 2021 (research, 614 participants): visible pledging no effect on pledge rate; "less than 1 percent of pledges were fulfilled." Capra, Jiang & Su, Economic Inquiry 2022 (research): "pledges are often reneged later, resulting in no overall change in volunteering rates." Giving What We Can 2025 Impact Evaluation (charity's own): recording "from around 60% in year one to roughly 7% by the end of the 35-year pledge lifetime"; direct 10% pledgers ~25% recording at year five vs ~50% for those who started with a trial pledge; 2023–24 report: surveys "didn't find a meaningful signal that most of these pledgers are, in fact, donating." John et al., JNPSM 2019 (research, RCT, 100,000+ UK students): "Peer endorsements reduced click-throughs to volunteering unit websites."
• Giving circles — CGRG 2017 (research report): ~1,600 groups, ~150,000 donors, $1.29B extrapolated from 37% of groups; "More than half of GC members report that participation… has increased the amount of money." In Abundance 2024 (Johnson Center with Philanthropy Together, research/charity's own): "nearly 4,000 collective giving groups mobilized approximately 370,000 philanthropists to donate more than $3.1 billion" 2017–23 — totals imputed ("assumed an equal giving amount per year"); 63% self-report giving more. No comparison group in either.
• Congregations — Pew, Religion in Everyday Life, 2016 (primary, survey): volunteered last week 45% vs 28%; through a congregation 23% vs 4%; helped the poor last week 65% vs 41%; holds with controls; "The causal arrow could point in the other direction"; surveys "may overstate… volunteering." Religion received $151.58B of $617.2B US giving in 2025 (Giving USA via NonProfit Times, news). No evidence specific to congregation-hosted VITA sites found.
• VITA as a volunteer community — no retention study specific to VITA found. IRS counts rose from 76,314 volunteers (FY2025, Data Book) to 84,500+ (FS2026, Pub 5347). General: volunteers with multiple ties to an organisation 27% more likely to stay (Yu, Tan & Tan 2024, research, observational, Beijing, 827,260 residents). Formal volunteering 28.3% of Americans in 2023 (AmeriCorps/Census via a news copy, news; agency pages blocked).
• Mutual aid — one qualitative study (32 UK organiser interviews, 2021, research); no outcome evaluation with a comparison group found. Grade: none.
• Tenant unions — Messamore, Social Problems 2023 (research, 75 cities, 2000–16): ten more community nonprofits per 100,000 residents ↔ 10% fewer eviction filings; not tenant unions specifically. Organiser claims (unread) not used. Grade: thin.
• Unions as a group format — Farber, Herbst, Kuziemko & Naidu (NBER w24587, rev. 2021; QJE 2021, research): union households "10-20% higher family income", larger for non-white and less-educated workers; union growth "explains over one-fourth in the 1936-1968 decline in the Gini coefficient"; Card, Lemieux & Riddell (IZA 2018, research): "unions reduce economy-wide wage inequality by less than 10%" today. BLS 2025 (primary): 10.0% membership, union $1,404 vs $1,174 weekly, uncontrolled. Individual-level, close-election designs disagree (DiNardo & Lee 2004 ≈0 on wages; Frandsen 2010/2021 compression within the firm, negative establishment effects via composition; Knepper 2020 benefits and labour share up).

---

D. Organisations the pages link to or name

Alternatives checked for status: Benefits Data Trust — site unreachable today (certificate error), closure not verified; the SNAP RCT above was run with it. California Policy Lab CTC-gap project: results "expected in early 2025", not published as of today.

---

E. What would overturn this, what is stale, what was not verified

Would overturn the ranking:

• Cash #1: a failed replication of Egger et al.'s spillovers plus evidence of recipient fade-out would pull the grade to "moderate"; GiveWell's next update (annual) is the trigger to re-read.
• Legislator contact #2: a well-powered null on constituent contact at state level, or evidence the effect vanishes on salient bills, would drop it below VITA.
• VITA #3: an RCT showing in-person VITA availability (hours, sites) raises EITC claiming among non-filers would lift it to "strong" and back the page's current wording; a new TIGTA anonymous audit at 2009–13 accuracy levels would cut it.
• Right to counsel #4: published null results from a later NYC evaluation.

Time-sensitive, with dates:

• NYC season figures: September each year (Mayor's office). Volunteer info sessions end 3 Dec 2026 (already in stale.tsv).
• IRS grants: late January each year; VITA-2027 awards due by Jan 2027.
• Census SPM 2019–2024 revision: Census said it will publish revised rates; the Learn page's 5.2%/12.4% may move. Re-check at the next Census poverty release (Sept 2027) or sooner if Census posts the revision.
• GiveWell's GiveDirectly page: last updated Oct 2025.
• Oxfam GB tribunal: "ongoing wider legal proceedings" as of 7 Jul 2026.
• NDWA dispute: last dated item 18 Nov 2025.
• BLS OOH wages: May 2025 data.

Not verified today (do not use as read):

• Bergan 2009 effect size (+20 points) — secondary only.
• Butler & Nickerson 2011 magnitude.
• Devex account of the Oxfam GB CEO departure — snippet only.
• Benefits Data Trust closure.
• EPI domestic-workers chartbook figures (403).
• Jaroszewicz et al., US one-time cash RCT (SSRN 403) — exists, not read, no figures used.
• Denver Basic Income Project Year Two report — not found on the project's site; do not cite.
• Filmer et al.'s exact food-price percentages; Evans & Popova's estimate count; Hidrobo et al.'s magnitudes; Guettabi's PDF (figures via the university's write-up only); Nature Food 2026 Chelsea nutrition magnitudes; Gennetian 2024 dollar figure (J-PAL's "$68").
• Any TIGTA/GAO VITA report after 2017 (none found).
• Share of household employers paying off the books; evaluations of the Domestic Workers' Bill of Rights.
• Nonprofit operators of NYC Free Tax Prep sites (not named on the pages fetched).
• Charity Commission status of Oxfam GB (register blocked).
• Whether NDWA's strike happened or a settlement was reached.

Searches that could not be run (budget exhausted): Oxfam GB regulatory status; NDWA resolution; Jaroszewicz et al.; Benefits Data Trust; domestic-worker compliance; DWBR evaluations; GetYourRefund 2026 outcomes; CMF 2020s surveys; federal roll-call experiments; tenant-union causal studies; Benevity; Independent Sector recruitment data; workplace peer-effect RCTs; Urban Institute VITA work; SPEC staffing.

---

Sources (fetched 2026-11-05 unless marked; grade in brackets)

Keystones re-opened by the reviewer

• IRS, EITC participation rate by state (reviewed 19 Aug 2026) — https://www.irs.gov/tax-professionals/eitc-central/eitc-participation-rate-by-state [primary]
• IRS, Statistics for tax returns with the EITC (reviewed 28 Jul 2026; TY2024: 24M, $69.6B, $2,894; NY 1.5M, $4.2B, $2,753) — https://www.irs.gov/credits-deductions/individuals/earned-income-tax-credit/statistics-for-tax-returns-with-the-earned-income-tax-credit-eitc [primary]
• GiveWell, GiveDirectly cash for poverty relief program (Nov 2024, updated Oct 2025) — https://www.givewell.org/international/technical/programs/givedirectly-cash-for-poverty-relief-program [research/evaluator]
• GiveWell, Top charities (updated Oct 2026) — https://www.givewell.org/charities/top-charities [research/evaluator]
• GiveWell, How much does it cost to save a life (Apr 2024) — https://www.givewell.org/how-much-does-it-cost-to-save-a-life [research/evaluator]
• Egger, Haushofer, Miguel, Niehaus & Walker, NBER w26600; Econometrica 90(6) 2022 — https://www.nber.org/papers/w26600 [research]
• Vivalt, Rhodes, Bartik, Broockman, Krause & Miller, NBER w32719 (rev. Aug 2026); QJE 141(4) 2026 — https://www.nber.org/papers/w32719 [research]
• Noble et al., Baby's First Years at age four, NBER w33844 (May 2025) — https://www.nber.org/papers/w33844 [research]
• US Census, Income, Poverty and Health Insurance Coverage in the United States: 2022 (12 Sep 2023; SPM child 5.2% → 12.4%; BLS threshold re-release note) — https://www.census.gov/newsroom/press-releases/2023/income-poverty-health-insurance-coverage.html [primary]
• NYC DCWP, Become an NYC Free Tax Prep volunteer (updated 10/2026) — https://www.nyc.gov/site/dca/talk-money/become-an-nyc-free-tax-prep-volunteer.page [primary]
• NYC DCWP, File your taxes (2025 tax year; limits $97,000/$68,000) — https://www.nyc.gov/site/dca/talk-money/file-your-taxes.page [primary]
• NYC DCWP, FAQ (updated 03/2026; $346 average fee) — https://www.nyc.gov/site/dca/talk-money/file-your-taxes-faqs.page [primary]
• NYC Mayor's Office, "$226.6 million back in New Yorkers' pockets" (30 Sep 2026; 100,216 returns, $188.8M, $37.8M) — https://www.nyc.gov/mayors-office/news/2026/09/mayor-mamdani-puts--226-6-million-back-in-new-yorkers--pockets-w [charity's own/city]
• NYC Mayor's Office, Free tax prep launch (3 Feb 2026; 110,000 returns 2025, $346, $38.2M; 140+ sites) — https://www.nyc.gov/mayors-office/news/2026/02/mayor-mamdani-joins-dcwp-and-nyc-health---hospitals-to-promote-f [charity's own/city]
• BLS OOH, Childcare workers (May 2025, $16.82) — https://www.bls.gov/ooh/personal-care-and-service/childcare-workers.htm [primary]
• BLS OOH, Home health and personal care aides ($17.21) — https://www.bls.gov/ooh/healthcare/home-health-aides-and-personal-care-aides.htm [primary]
• BLS OOH, Janitors and building cleaners ($17.71; private households excluded) — https://www.bls.gov/ooh/building-and-grounds-cleaning/janitors-and-building-cleaners.htm [primary]
• EPI, Domestic workers chartbook — https://www.epi.org/publication/domestic-workers-chartbook-a-comprehensive-look-at-the-demographics-wages-benefits-and-poverty-rates-of-the-professionals-who-care-for-our-family-members-and-clean-our-homes/ [403 — unread]
• Jaroszewicz et al., SSRN 4154000 — https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4154000 [403 — unread]

Organisations (reviewer)

• ProPublica Nonprofit Explorer: GiveDirectly — https://projects.propublica.org/nonprofits/organizations/271661997 ; The Clear Fund (GiveWell) — https://projects.propublica.org/nonprofits/organizations/208625442 ; Oxfam America — https://projects.propublica.org/nonprofits/organizations/237069110 ; NDWA — https://projects.propublica.org/nonprofits/organizations/352420942 ; Western Queens CLT — https://projects.propublica.org/nonprofits/organizations/853774663 ; searches "National Domestic Workers Alliance", "Hand in Hand Domestic Employers", "Domestic Employers Network", "United States Federation of Worker Cooperatives" [primary]
• Charity Navigator: GiveDirectly — https://www.charitynavigator.org/ein/271661997 ; GiveWell — https://www.charitynavigator.org/ein/208625442 ; Oxfam America — https://www.charitynavigator.org/ein/237069110 ; NDWA — https://www.charitynavigator.org/ein/352420942 [research/evaluator]
• GiveDirectly, Financials ("$87 of every $100") — https://www.givedirectly.org/financials/ ; Financials by year (2024 990 and audit posted) — https://www.givedirectly.org/financials-by-year/ ; DRC case 2023 (5 Jun 2023, updates to 24 Sep 2024) — https://www.givedirectly.org/drc-case-2023/ [charity's own]
• Fortune/AP, GiveDirectly SNAP shutdown payments (24 Dec 2025) — https://fortune.com/2025/12/24/government-shutdown-snap-payments-givedirectly-nonprofit-food-stamps/ [news]
• NPR via Connecticut Public, aid cuts and cash (13 Aug 2025) — https://www.ctpublic.org/2025-08-13/trump-administrations-cuts-to-foreign-aid-threaten-trend-toward-giving-cash-directly [news]
• Oxfam GB press releases (listing; pages 4, 10, 11) — https://www.oxfam.org.uk/media/press-releases/ ; Statement from the Board of Trustees (7 Jul 2026) — https://www.oxfam.org.uk/media/press-releases/statement-from-the-board-of-trustees/ ; chair appointment (9 Oct 2026) — https://www.oxfam.org.uk/media/press-releases/oxfam-appoints-stevie-spring-cbe-as-chair-of-trustees/ ; ethnicity pay gap (26 Oct 2026) — https://www.oxfam.org.uk/media/press-releases/oxfam-responds-to-reporting-on-its-ethnicity-pay-gap/ [charity's own]
• Oxfam America press page — https://www.oxfamamerica.org/press/ [charity's own]
• Devex, ousted Oxfam CEO (403) — https://www.devex.com/news/ousted-oxfam-ceo-lambastes-charity-for-brutal-dismissal-111893 [unread]
• Charity Commission register, Oxfam 202918 (403) — https://register-of-charities.charitycommission.gov.uk/charity-search/-/charity-details/202918 [unread]; Charity Commission news page — https://www.gov.uk/government/organisations/charity-commission [primary]
• CWA, NDWA strike authorisation (17 Nov 2025) — https://cwa-union.org/news/releases/national-domestic-workers-alliance-union-members-vote-93-authorize-strike-amid [news/union]; CWA Local 1180 (18 Nov 2025) — https://cwad1.org/news/cwa-local-1180-members-authorize-strike-if-management-cannot-come-fair [news/union]; CWA Local 1180 news list — https://cwad1.org/news ; CWA news — https://cwa-union.org/news ; GoFundMe, NDWA staff union — https://www.gofundme.com/f/fund-the-fight-support-striking-ndwa-staff-union [news/union]; NDWA home — https://www.domesticworkers.org/ [charity's own]
• Hand in Hand: home — https://domesticemployers.org/ ; templates — https://domesticemployers.org/resources-and-faqs/templates/ ; donate — https://domesticemployers.org/donate/ ; privacy policy — https://domesticemployers.org/privacy-policy/ [charity's own]
• Western Queens CLT — https://wqclt.org/ [charity's own]
• USFWC — https://www.usworker.coop/ ; Certified EO — https://www.certifiedeo.com/ [own]
• IRS, Tax volunteers (reviewed 20 Aug 2026) — https://www.irs.gov/individuals/irs-tax-volunteers ; VITA grant program (reviewed 3 Nov 2026) — https://www.irs.gov/individuals/irs-vita-grant-program ; IR-2026-19 (30 Jan 2026) — https://www.irs.gov/node/155196 ; IR-2026-60 (1 May 2026) — https://www.irs.gov/node/159571 [primary]
• Code for America, Tax benefits — https://codeforamerica.org/programs/tax-benefits/ ; GetYourRefund — https://www.getyourrefund.org/ [charity's own]
• Benefits Data Trust — https://bdtrust.org/ [certificate error — unread]

EITC / VITA / CTC stream (agent-opened)

• Bhargava & Manoli, AER 2015 — https://www.cmu.edu/dietrich/sds/docs/bhargava/bhargava-aer-2015.pdf [research]
• Manoli & Turner, NBER w20718 — https://www.nber.org/papers/w20718 [research]; venue check — https://www.federalreserve.gov/econres/nick-turner.htm [primary]
• Linos, Prohofsky, Ramesh, Rothstein & Unrath, AEJ:EP 2022 — https://www.aeaweb.org/doi/10.1257/pol.20200603 and https://par.nsf.gov/servlets/purl/10430894 [research]
• Goldin, Homonoff, Javaid & Schafer, NBER w28398 — https://nber.org/system/files/working_papers/w28398/w28398.pdf [research]
• Guyton et al., AER P&P 2017 — https://econpapers.repec.org/RePEc:aea:aecrev:v:107:y:2017:i:5:p:471-75 ; J-PAL summary — https://www.povertyactionlab.org/evaluation/sending-reminders-improve-tax-filing-decisions-among-low-income-individuals-united [research]
• IRS Publication 5347, QSS results (Sept 2026) — https://www.irs.gov/pub/irs-access/p5347_accessible.pdf [primary]
• IRS Data Book FY2025, Service to taxpayers — https://www.irs.gov/statistics/service-to-taxpayers [primary]
• National Taxpayer Advocate, 2025 Annual Report (Jan 2026) — https://www.taxpayeradvocate.irs.gov/wp-content/uploads/2026/01/ARC_Publication-2104_2025_Web.pdf [primary]
• TIGTA 2016-40-045 — https://www.tigta.gov/sites/default/files/reports/2025-08/201640045fr.pdf ; TIGTA 2013-40-110 — https://www.tigta.gov/sites/default/files/reports/2024-11/201340110fr.pdf ; TIGTA 2017-40-088 — https://www.tigta.gov/sites/default/files/reports/2025-08/201740088fr.pdf ; TIGTA readiness memo (26 Jan 2026) — https://www.tigta.gov/sites/default/files/reports/2026-01/2026400002-Readiness-Memo.pdf [primary]
• AccountingToday on TIGTA audits: 2013 — https://accountingtoday.com/news/irs-volunteer-tax-prep-sites-have-accuracy-problems ; 2011 — https://www.accountingtoday.com/news/accuracy-rates-declining-at-volunteer-tax-prep-sites ; 2009 — https://www.accountingtoday.com/news/volunteer-tax-preparers-have-high-error-rates ; 2016 — https://www.accountingtoday.com/news/irs-review-of-volunteer-tax-prep-program-falls-short ; Direct File ends (6 Nov 2025) — https://www.accountingtoday.com/news/irs-ends-direct-file-program [news]
• Route Fifty, Direct File (5 Nov 2025) — https://www.route-fifty.com/digital-government/2025/11/direct-file-wont-happen-2026-irs-tells-states/409322/ [news]
• Russell Sage Foundation, Alba project (Jun 2026) — https://www.russellsage.org/node/16275 [charity's own]
• NACo webinar deck (Nov 2023) — https://naco.org/sites/default/files/2023-11/11.13%20Webinar%20Slide%20Deck.pdf [industry]
• New America, IRS data on VITA — https://www.newamerica.org/insights/improving-public-assistance-for-low-income-tax-filers/iii-irs-data-on-vita-programs/ [charity's own]
• PPI, The Price of Paying Taxes II (Apr 2016) — https://www.progressivepolicy.org/wp-content/uploads/2016/04/2016.04-Weinstein_Patten_The-Price-of-Paying-Takes-II.pdf [research, think tank]
• Brookings, The Price of Paying Taxes (2002) — https://www.brookings.edu/articles/the-price-of-paying-taxes-how-tax-preparation-and-refund-loan-fees-erode-the-benefits-of-the-eitc/ [research, think tank]
• Code for America research: success story — https://codeforamerica.org/success-stories/making-it-easier-for-people-with-low-income-to-claim-their-tax-credits/ ; research index — https://codeforamerica.org/news/tax-benefits-research-findings-and-analysis/ ; non-filer learnings (Apr 2021) — https://digitalgovernmenthub.org/wp-content/uploads/2022/09/filer-learnings-and-recommendations-april-2021.pdf ; GetCTC 2021 — https://codeforamerica.org/news/getctc-simplified-filing-report-2021/ ; Importance of Simplified Filing (Feb 2023) — https://files.codeforamerica.org/2023/02/13122111/importance-of-simplified-filing-getctc-2022-learnings-report.pdf ; Comparing Levels of Service II — https://files.codeforamerica.org/2023/02/08121452/comparing-levels-of-service-in-tax-filing-part-II-getctc-2022-research.pdf ; Offers of Assistance — https://files.codeforamerica.org/2023/03/06150903/offers-of-assistance-in-tax-benefits-outreach-getctc-2022-research.pdf ; Outreach to Traditional Non-Filers (Jan 2024) — https://files.codeforamerica.org/2024/01/26101934/outreach-traditional-non-filers-getctc-2022-learnings-report.pdf ; Zucker NTA slides (2023) — https://ntanet.org/wp-content/uploads/2023/05/Zucker-2023-NTA-Spring-Presentation.pdf [charity's own, with embedded RCTs]
• Prosperity Now VITA Support Fund — https://www.prosperitynow.org/vitasupport [charity's own]
• Finkelstein & Notowidigdo, QJE 2019 — https://www.nber.org/papers/w24652 ; https://www.law.nyu.edu/sites/default/files/Notowidigdo%20Take-up%20and%20Targeting%20Experimental%20Evidence%20from%20SNAP.pdf [research]
• Bettinger, Long, Oreopoulos & Sanbonmatsu, QJE 2012 — https://www.nber.org/papers/w15361 ; J-PAL — https://www.povertyactionlab.org/print/pdf/node/2133?lang=en [research]
• Bird et al., NBER w26158 — https://www.nber.org/papers/w26158 [research]
• Census releases: 2022 — https://www.census.gov/newsroom/press-releases/2022/income-poverty-health-insurance-coverage.html ; 2024 — https://www.census.gov/newsroom/press-releases/2024/income-poverty-health-insurance-coverage.html ; 2025 — https://www.census.gov/newsroom/press-releases/2025/income-poverty-health-insurance-coverage.html ; 2026 — https://www.census.gov/newsroom/press-releases/2026/income-poverty-health-insurance-coverage.html ; P60-290 (Sept 2026) — https://www.census.gov/library/publications/2026/demo/p60-290.html and https://www2.census.gov/library/publications/2026/demo/p60-290.pdf [primary]
• NBC Chicago/AP on Columbia CPSP (23 Feb 2022) — https://www.nbcchicago.com/news/national-international/3-7-million-more-children-in-poverty-a-month-after-tax-credit-expired-study-says/2767558/ [news]
• Ananat, Glasner, Hamilton & Parolin, NBER w29823 — https://www.nber.org/papers/w29823 [research]
• Corinth, Meyer, Stadnicki & Wu, NBER w29366 — https://www.nber.org/papers/w29366 ; slides (Jan 2023) — https://cehd.uchicago.edu/wp-content/uploads/2023/01/CTC-slides-for-SGE-ASSA-1.0.pdf ; Harris summary — https://harris.uchicago.edu/node/39376 [research]
• California Policy Lab, CTC gap — https://capolicylab.org/topics/social-safety-net/addressing-the-child-tax-credit-gap/ ; EITC nudges WP page — https://www.capolicylab.org/wp-content/uploads/2020/07/EITC-nudges-WP-202007.pdf [research]
• Grants.gov VITA-2027 — https://simpler.grants.gov/opportunity/361972 [primary]; Don't Mess With Taxes (2 Dec 2024) — https://dontmesswithtaxes.com/?p=3673 [news]; CPA Practice Advisor (11 May 2026) — https://www.cpapracticeadvisor.com/2026/05/11/irs-applications-for-tce-and-vita-program-grants-now-accepted/183161/ [news]; IRS newsroom IR-2026-60 — https://www.irs.gov/newsroom/irs-seeking-applications-for-tax-counseling-for-the-elderly-and-volunteer-income-tax-assistance-program-grants-0 [primary]
• The Tax Adviser (28 Jan 2026) — https://www.thetaxadviser.com/news/2026/jan/new-law-irs-workforce-cuts-raise-red-flags-for-tax-season-reports-say.html ; AP via Bangor Daily News (28 Jan 2026) — https://www.bangordailynews.com/2026/01/28/nation/irs-faces-stiff-challenges-in-2026-tax-season-due-to-workforce-cuts-and-new-laws-a-watchdog-says/ [news]
• IRS, Free tax return preparation for qualifying taxpayers — https://www.irs.gov/individuals/free-tax-return-preparation-for-qualifying-taxpayers [primary]
• IRS–Census CES-24-75 (Dec 2024) — https://www.census.gov/library/working-papers/2024/adrm/CES-WP-24-75.html ; Jones CARRA 2014-04 — https://www.census.gov/content/dam/Census/library/working-papers/2014/adrm/carra-wp-2014-04.pdf [primary]

Civic / unions / housing / group formats stream (agent-opened)

• Bergan & Cole 2015 abstract — https://www.springerprofessional.de/en/call-your-legislator-a-field-experimental-study-of-the-impact-of/5682222 ; authors' LSE post (29 Apr 2015) — https://blogs.lse.ac.uk/usappblog/2015/04/29/constituent-contacts-can-influence-how-legislators-vote [research; authors' summary]
• Bergan 2009 (403) — https://journals.sagepub.com/doi/10.1177/1532673X08326967 [unread]; Dickens posts (8–9 Nov 2025) — https://ea.greaterwrong.com/posts/PvJL4Rnz2Dq9J2omd/writing-your-representatives-a-worthwhile-and-neglected ; https://buttondown.com/mdickens/archive/writing-your-representatives-a-cost-effective-and-neglected-intervention/ [news/blog]
• Butler & Nickerson 2011 — https://isps.yale.edu/research/publications/isps11-016 ; https://ippsr.msu.edu/node/713 [research]
• Kalla & Broockman, AJPS 2016 — https://escholarship.org/content/qt8wm3k5g4/qt8wm3k5g4_noSplash_9cb11c3ab404d8434ad6016ad9a51da6.pdf [research]
• Congressional Management Foundation, Citizen-Centric Advocacy (2017) — https://www.congressfoundation.org/s/cmf-citizen-centric-advocacy.pdf (redirects to https://static1.squarespace.com/static/67ead1d67cfe8944d45170dd/t/68d303702e3b2b315776617b/1771948878481/cmf-citizen-centric-advocacy.pdf) [primary, survey]
• Farber, Herbst, Kuziemko & Naidu, NBER w24587 (rev. Apr 2021) — https://www.nber.org/papers/w24587.pdf [research]
• Card, Lemieux & Riddell, NBER w9473 — https://www.nber.org/papers/w9473.pdf ; IZA DP 11964 (2018) — https://docs.iza.org/dp11964.pdf [research]
• BLS, Union Members 2025 (18 Feb 2026) — https://www.bls.gov/news.release/union2.nr0.htm [primary]
• DiNardo & Lee, QJE 2004 — https://public.websites.umich.edu/~jdinardo/Pubs/dinardo_lee2004.pdf [research]; Frandsen 2010 — http://www.aeaweb.org/conference/2011/retrieve.php?pdfid=42 ; Frandsen, JOLE 2021 abstract — https://econpapers.repec.org/RePEc:ucp:jlabec:doi:10.1086/711852 [research]; Mackinac summary — https://www.mackinac.org/23120 [news/advocacy]; Knepper, REStat 2020 abstract — https://econpapers.repec.org/RePEc:tpr:restat:v:102:y:2020:i:1:p:98-112 [research]
• Cassidy & Currie, NBER w29836 (rev. Jul 2022) — https://www.nber.org/system/files/working_papers/w29836/w29836.pdf [research]
• Seron, Van Ryzin, Frankel & Kovath, Law & Society Review 2001 — https://legalaidresearch.org/wp-content/uploads/2020/02/861809a07c335b6c9705bc3b849c5d4813e7.pdf [research, RCT]
• Stout, Right to counsel cost–benefit (16 Mar 2016) — https://www2.nycbar.org/pdf/report/uploads/SRR_Report_Financial_Cost_and_Benefits_of_Establishing_a_Right_to_Counsel_in_Eviction_Proceedings.pdf ; https://www.stout.com/en/experience/cost-benefit-analysis-for-nyc-right-to-counsel-legislation [primary, consulting model]
• NYC Office of Civil Justice, Universal Access Annual Report 2022 — https://www.nyc.gov/assets/hra/downloads/pdf/services/civiljustice/OCJ_UA_Annual_Report_2022.pdf ; NYC Opportunity brief (Dec 2021) — https://www.nyc.gov/assets/opportunity/pdf/policybriefs/right-to-counsel.pdf [primary, descriptive]
• Messamore, Social Problems 2023 — https://academic.oup.com/socpro/article/70/3/809/6969383 [research]
• Wang, Cahen, Acolin & Walter, Lincoln Institute WP19RW1 (Apr 2019) — https://lincolninst.edu/sites/default/files/pubfiles/wang_wp19rw1_rev_0.pdf [research]
• CGRG, State of Giving Circles (2017) — https://philanthropy.indianapolis.iu.edu/doc/institutes/giving-circles2017-executive-summary.pdf ; press release — https://www.prnewswire.com/news-releases/raising-nearly-13-billion-since-inception-giving-circles-effectively-engage-all-types-of-donors-new-study-finds-300555469.html [research report]
• Loson-Ceballos & Layton, In Abundance (2024) — https://johnsoncenter.org/wp-content/uploads/2024/04/in-abundance-an-analysis-of-the-thriving-landscape-of-collective-giving-in-the-u-s.pdf ; blog (15 Jan 2025) — https://johnsoncenter.org/blog/collective-giving-projected-to-double-in-the-next-five-years/ [research report/charity's own]
• Karlan & List, AER 2007 — https://econweb.ucsd.edu/~jandreon/PhilanthropyAndFundraising/Volume%202/14%20KarlanList.pdf ; registry — https://www.socialscienceregistry.org/trials/1135 [research]
• Meier, JEEA 2007 — https://econweb.ucsd.edu/~jandreon/PhilanthropyAndFundraising/Volume%202/13%20Meier.pdf [research]
• Huck & Rasul, IZA DP 5267 (2010; JPubE 2011) — https://docs.iza.org/dp5267.pdf [research]
• Double the Donation, matching gift statistics — https://doublethedonation.com/matching-gift-statistics/ [industry]; CECP Giving in Numbers (6 Aug 2026) — https://www.3blmedia.com/news/new-cecp-research-corporate-employee-engagement-programs-hit-participation-ceiling-return [industry]
• Pew, Religion in Everyday Life (12 Apr 2016) — https://www.pewresearch.org/religion/2016/04/12/religion-in-everyday-life/ ; PDF — https://www.pewresearch.org/wp-content/uploads/sites/7/2016/04/Religion-in-Everyday-Life-FINAL.pdf [primary, survey]
• Giving USA 2026 via NonProfit Times (23 Jun 2026) — https://thenonprofittimes.com/npt_articles/giving-usa-617-2-b-was-given-to-nonprofits-during-2025/ [news]
• Fernandes-Jesus et al., Frontiers in Psychology 2021 — https://www.frontiersin.org/articles/10.3389/fpsyg.2021.716202/full [research, qualitative]
• Giving What We Can, 2025 Impact Evaluation — https://www.givingwhatwecan.org/impact/2025-impact-evaluation (full report Google Doc 1GPFnzgSL_HROQYoDDjvYyT_WiPtI1skTfc3NGA0l0Bk) ; 2023–2024 evaluation — https://www.givingwhatwecan.org/impact/2023-2024-impact-evaluation [charity's own]
• AmeriCorps/Census, Volunteering and Civic Life in America (2023 data) via — https://stagingcbi.nonprofitpro.com/?p=21183 [news copy]; Census 2021 story — https://www.census.gov/library/stories/2023/01/volunteering-and-civic-life-in-america.html [primary]
• Yu, Tan & Tan 2024 via UT McCombs — https://news.mccombs.utexas.edu/?p=17 [news/university]
• John et al., JNPSM 2019 — https://kclpure.kcl.ac.uk/portal/en/publications/3835790f-fe5a-48db-b769-0fc2af7e66f2 [research]
• Jacob, Armstrong, Bowden & Pan, JREE 2016 (Reading Partners) — https://static1.squarespace.com/static/583b86882e69cfc61c6c26dc/t/595a36d24402439a88194a47/1499084499015/RP_JREE.pdf ; MDRC — https://www.mdrc.org/work/publications/mobilizing-volunteer-tutors-improve-student-literacy [research]
• Tierney, Grossman & Resch 1995 (BBBS) — https://www.ojp.gov/ncjrs/virtual-library/abstracts/making-difference-impact-study-big-brothers-big-sisters ; Grossman & Tierney 1998 — https://econpapers.repec.org/RePEc:sae:evarev:v:22:y:1998:i:3:p:403-426 [research]
• Karlan & McConnell 2013 — https://www.povertyactionlab.org/sites/default/files/research-paper/336_187%20hey%20look%20at%20me%20May2013.pdf [research]
• Meyer & Tripodi 2021 — https://ora.ox.ac.uk/objects/uuid:45fff1a5-80ce-4984-a455-a9cd3e76e886/files/s3r074v86d [research]
• Capra, Jiang & Su, Economic Inquiry 2022 — https://ideas.repec.org/a/bla/ecinqu/v60y2022i1p87-100.html [research]
• GivingTuesday blog (5 Nov 2019) — https://www.givingtuesday.org/blog/does-givingtuesday-more-giving-or-different-giving/ [charity's own]; Vance-McMullen registry — https://www.socialscienceregistry.org/trials/3264 [primary, no results]

Cash-transfer stream (agent-opened)

• Haushofer & Shapiro, QJE 2016 (author copy) — https://haushofer.ne.su.se/publications/Haushofer_Shapiro_UCT_QJE_2016.pdf [research]
• Haushofer & Shapiro, long-term impact (Jan 2018) — https://www.povertyactionlab.org/sites/default/files/research-paper/The-long-term-impact-of-conditional-cash-tranfer_Kenya_Haushofer_Shapiro_January2018.pdf [research, working paper]
• Banerjee, Faye, Krueger, Niehaus & Suri, UBI Kenya (15 Sep 2023) — https://econweb.ucsd.edu/~pniehaus/papers/UBI_main_paper.pdf [research, working paper]; GiveDirectly summary — https://www.givedirectly.org/2023-ubi-results/ ; research index — https://www.givedirectly.org/research-at-give-directly/ [charity's own]
• Aggarwal et al., AEJ: Applied 2026 (GiveDirectly-hosted) — https://www.givedirectly.org/wp-content/uploads/2026/10/dynamic_effects_AEJ.pdf [research]
• Walker, Shankar, Miguel, Egger & Killeen, NBER w34152 — https://www.nber.org/papers/w34152 [research]
• Haushofer, Ringdal, Shapiro & Wang, NBER w25627 — https://www.nber.org/papers/w25627 [research]
• Pilkauskas et al., COVID Cash Transfer Study (May 2022) — https://sites.fordschool.umich.edu/poverty2021/files/2022/06/Pilkauskas_et_al_RCT1_May2022.pdf [research, working paper]
• Filmer, Friedman, Kandpal & Onishi, REStat 2023 — https://econpapers.repec.org/RePEc:tpr:restat:v:105:y:2023:i:2:p:327-343 ; World Bank PRWP 8377 — https://documents1.worldbank.org/curated/en/989031522077749796/pdf/Cash-Transfers-Food-Prices-and-Nutrition-Impacts-on-Nonbeneficiary-Children.pdf [research]
• Banerjee, Hanna, Kreindler & Olken, WBRO 2017 — http://academic.oup.com/wbro/article/32/2/155/4098285/Debunking-the-Stereotype-of-the-Lazy-Welfare ; Sept 2016 version — https://www.povertyactionlab.org/sites/default/files/research-paper/Debunking-the-Streotype-of-the-Lazy-Welfare-Recipient_Banerjee-Hanna-Kreindler-Olken_Sept2016.pdf [research]
• Evans & Popova, EDCC 2017 — https://econpapers.repec.org/RePEc:ucp:ecdecc:doi:10.1086/689575 [research]
• Bastagli et al., ODI (Jul 2016) — https://cdn.odi.org/media/documents/11316.pdf [research, review]
• Cunha, De Giorgi & Jayachandran, REStud 2019 — https://www.nber.org/papers/w17456 ; FRBNY SR 735 — https://www.newyorkfed.org/medialibrary/media/research/staff_reports/sr735.pdf [research]
• Cunha, AEJ: Applied 2014 — https://www.aeaweb.org/articles?id=10.1257/app.6.2.195 ; 2010 draft — https://www.frbsf.org/TestingPaternalism_JesseCunha.pdf [research]
• Hidrobo et al., JDE 2014 — https://econpapers.repec.org/RePEc:eee:deveco:v:107:y:2014:i:c:p:144-156 ; IFPRI record — https://cgspace.cgiar.org/items/94a9e5e0-f16a-4370-bbfa-f96e337b0caf [research]
• McIntosh & Zeitlin, Gikuriro — https://arxiv.org/abs/2106.00213 ; Huguka Dukore — https://arxiv.org/abs/2209.08574 [research, working papers]
• Gentilini, WBRO 2016 — https://econpapers.repec.org/RePEc:oup:wbrobs:v:31:y:2016:i:1:p:135-167 [research]
• Stockton SEED, preliminary analysis (Mar 2021) — https://static1.squarespace.com/static/6039d612b17d055cac14070f/t/6050294a1212aa40fdaf773a/1615866187890/SEED_Preliminary+Analysis-SEEDs+First+Year_Final+Report_Individual+Pages+-2.pdf [charity's own]; West & Castro, J. Urban Health 2023 — https://link.springer.com/article/10.1007/s11524-023-00723-0 [research]; CGIR publications — http://www.penncgir.org/publications ; https://www.stocktondemonstration.org/ [charity's own]
• Denver Basic Income Project, Year One report (Jun 2024) — https://hpri.usc.edu/wp-content/uploads/2024/07/CHHR_DBIP_Report.pdf ; research page — https://www.denverbasicincomeproject.org/research [charity's own]
• Chelsea Eats: index — https://www.hks.harvard.edu/centers/taubman/programs-research/rappaport/research-and-publications/special-collections/covid-19-relief-chelsea-ma ; experimental impacts (Dec 2022) — https://www.hks.harvard.edu/sites/default/files/Taubman/RIGB/Chelsea%20Eats%20Experimental%20Impacts%20120622.pdf [research, working paper]; Agarwal, Cook & Liebman, JAMA 2024 — https://jamanetwork.com/journals/jama/fullarticle/2821454 [research]; Lee et al., Nature Food 2026 (metadata only) — https://www.hks.harvard.edu/sites/default/files/2026-02/Lee_et_al-2026-Nature_Food.pdf [unread]
• Compton Pledge, Balakrishnan et al., NBER w33209 — https://www.nber.org/papers/w33209 ; funder page — https://policiesforaction.org/node/805 [research]
• Baby's First Years: Troller-Renfree et al., PNAS 2022 — https://escholarship.org/content/qt9j61j5ps/qt9j61j5ps.pdf ; Gennetian et al., NBER w30379 — https://www.nber.org/papers/w30379 ; J-PAL summary (24 Nov 2025) — https://www.povertyactionlab.org/node/8137695 ; Magnuson et al., Nature Communications 2025 — https://www.nature.com/articles/s41467-025-62438-x ; Duncan et al., JAMA Pediatrics 2025 — https://jamanetwork.com/journals/jamapediatrics/fullarticle/2834896 ; Raffington et al., Nature Human Behaviour 2026 — https://www.nature.com/articles/s41562-026-02568-4 ; publications list — https://www.babysfirstyears.com/publications ; home — https://www.babysfirstyears.com/ [research; charity's own for the site]
• Corinth & Mayhew, AEI (Feb 2026) — https://cosm.aei.org/wp-content/uploads/2026/02/Corinth-and-Mayhew-Guaranteed-Basic-Income-Pilots-1.pdf [research, think tank]
• Peng, Hahm, Noonan & Woronkowicz, IU systematic review (draft Jan 2025) — https://culturalaffairs.center.iu.edu/programs/aeilab/effects-of-recent-guaranteed-income-pilots-in-the-united-states-a-systematic-literature-review-and-quality-assessment-draft-january-2025.pdf [research, draft]
• University of Chicago Inclusive Economy Lab, Policymakers' Report (Oct 2025) — https://urbanlabs.uchicago.edu/attachments/40baa6d7e49551dc92e9a4204af9cb10dbc566f9/store/596a5de8256dd10415e8c3125c8cd78d670908a4ed0672f3d0c824a3c293/Policymakers%27+Report.pdf [research, synthesis]
• Economic Security Project, Cash Works (Sept 2026) — https://economicsecurityproject.org/wp-content/uploads/Cash-Works.-So-Do-Its-Recipients.pdf [charity's own, advocacy]
• Jones & Marinescu, AEJ:EP 2022 — https://www.aeaweb.org/articles?id=10.1257/pol.20190299 [research]; Guettabi, ISER 2019 record — https://scholarworks.alaska.edu/uaa_iser_reports/175 (PDF blocked) ; UAA write-up (28 Jun 2019) — https://www.uaa.alaska.edu/news/archive/2019/06/iser-research-pfd-effect-socio-economic-well-being.cshtml [news]
• GiveDirectly, Financials — https://www.givedirectly.org/financials/ ; by year — https://www.givedirectly.org/financials-by-year/ [charity's own]

Corrections, 2026-11-06

From the re-verification of 6 November (source/verify/inequality-2026-11-06-additions.md): the 2017 giving-circle study does compare members with "donors not involved in a GC" (this file said no comparison group). Karlan and McConnell's 14% comes from telling donors that larger givers are named in a newsletter. Karlan and List's 19% and 22% are for the match offer at any ratio, pooled. Census P60-290 (September 2026) revised SPM child poverty to 5.1% (2021) and 12.0% (2022) on new thresholds.

No tracking on any link.Ticks stay on this device.What changedContested