US guaranteed income pilots: what Stockton, OpenResearch, Cook County and Baby's First Years found

Since 2019 the United States has become the world's busiest laboratory for "guaranteed income": monthly, unconditional cash paid to a selected group of lower-income residents for a fixed period, usually by a city, county or charity. These pilots are not universal basic income — they are targeted, temporary and small — but they are where most of the recent rich-country evidence on unconditional cash comes from. This study examines four of the most informative: the Stockton Economic Empowerment Demonstration (SEED), the OpenResearch Unconditional Income Study, the Cook County Promise Guaranteed Income Pilot and the Baby's First Years trial. Read together, they tell a more sober story than early publicity suggested. Cash reliably eases financial hardship and short-term stress. The largest and most rigorous trial found a modest reduction in work and wellbeing gains that faded after the first year. And a carefully designed trial of cash for mothers of newborns found no detectable effects on children's development or on maternal and child health at age four.

Key findings

• Stockton (SEED). From February 2019, 125 residents received $500 a month for two years; 200 people formed a control group. In the first year, before the pandemic, the share of recipients in full-time work rose from 28% to 40%, against 32% to 37% in the control group. The peer-reviewed evaluation reports lower income volatility, lower mental distress and better energy and physical functioning. The sample was small, and the employment comparison is descriptive.
• OpenResearch. From November 2020 to October 2023, 1,000 low-income adults aged 21–40 in two US states received $1,000 a month and 2,000 controls received $50 a month. Recipients' income excluding the transfers fell by about $1,900 a year relative to controls, labour-market participation fell by 4.2 percentage points and hours by one to two a week. Wellbeing improved in the first year and then reverted to control levels.
• OpenResearch, health and finances. The transfer produced large but short-lived improvements in stress and food security and increased use of hospital, emergency and dental care, but did not improve self-reported physical health, clinical biomarkers or mental health after the first year. Household spending rose by at least $300 a month; net worth barely changed.
• Cook County. From December 2022 to January 2025, the county paid $500 a month to 3,250 low- to moderate-income families selected by lottery, funded by the American Rescue Plan Act. Early findings published in April 2025 are participant survey results — 75% reported feeling more financially secure, 56% reduced stress — without a published comparison with non-recipients.
• Baby's First Years. 1,000 mothers with incomes near or below the poverty line were randomised at birth to receive $333 or $20 a month. At age four (891 families), there were no statistically significant effects on four pre-registered child outcomes — language, executive function, social-emotional problems and high-frequency brain activity — and no improvement in maternal mental health, maternal or child body mass index, or reported child health.
• The pattern. Small, advocacy-linked pilots tended to report broad benefits; larger, pre-registered trials found narrower effects. Cash improves material circumstances and short-run security; claims that it transforms work, health or child development are not supported by the strongest evidence.

Research question and scope

This study asks what the recent American guaranteed income pilots show about four outcomes: work and earnings; financial security and spending; physical and mental health; and, for families, children's development. It also asks how much confidence each pilot's design allows.

We focus on four programmes chosen for the strength or influence of their evidence: SEED, because it launched the movement; OpenResearch, because it is the largest randomised trial; Cook County, because it is the largest publicly funded pilot of its kind and moved towards a permanent programme; and Baby's First Years, because it is the most rigorous test of cash for young children. We do not attempt to survey the many smaller municipal pilots, whose evaluations vary widely in quality. We compare results with the Finnish and German experiments, each covered by its own atlas study.

Background

Guaranteed income is not universal basic income

A universal basic income is paid to everyone, individually, without a means test, indefinitely. The American pilots differ on almost every one of these dimensions. They select lower-income participants, usually by lottery among applicants; they pay for a fixed term, typically 12 to 36 months; they are often funded by philanthropy or one-off federal pandemic money; and they cover a tiny share of the local population. Their findings are best read as evidence on what an unconditional cash supplement does for low-income individuals, not as evidence about a society-wide basic income.

Earlier American evidence

The United States ran large income-guarantee experiments before. Between the late 1960s and early 1980s, a series of negative income tax experiments tested income guarantees that were withdrawn as earnings rose. As Verho and colleagues summarise in their study of Finland, those experiments found that the guaranteed income led participants to reduce their work effort, with women affected more than men. Their designs differed from today's pilots — they varied guarantee levels and withdrawal rates, and they targeted working families — but they established the question that every later trial has had to answer.

Why a new wave of pilots

Recent pilots were motivated by income volatility among low-wage workers, by the experience of pandemic cash payments, and by advocacy for a permanent federal income floor. Stockton's mayor launched SEED in 2019 as the nation's first mayor-led guaranteed income initiative. Federal pandemic funds from the American Rescue Plan Act later allowed counties such as Cook County to fund larger programmes. Several pilots were designed and funded by organisations that support guaranteed income; this does not invalidate their findings, but it makes pre-registration and independent analysis especially important.

Method and sources

Types of evidence

The pilots produce three kinds of evidence, which deserve different weight:

1. Pre-registered randomised trials with administrative data (OpenResearch, Baby's First Years). These compare a treatment group with a randomly selected control group, specify outcomes in advance and combine surveys with records such as tax, credit or health data.
2. Smaller randomised pilots with mixed methods (SEED). These have a control group but small samples, and rely more on surveys and interviews.
3. Descriptive participant surveys (Cook County's April 2025 early findings). These report what recipients say, without a comparison group, and cannot estimate causal effects.

Control-group payments

Several trials pay the control group something — $50 a month in OpenResearch, $20 a month in Baby's First Years — to keep them engaged in surveys. This reduces attrition and makes the comparison one between a large and a small transfer, not between cash and nothing. It slightly understates the effect of cash relative to no cash, but by little given the amounts.

Sources

For SEED we use the first-year report released in March 2021 and the peer-reviewed article by Stacia West and Amy Castro in the Journal of Urban Health (2023). For OpenResearch we use the employment paper by Vivalt and colleagues (NBER Working Paper 32719, 2024; published in the Quarterly Journal of Economics in 2026), the health paper by Miller and colleagues (NBER Working Paper 32711) and the consumption paper by Bartik and colleagues (NBER Working Paper 32784). For Cook County we use the county's official programme page. For Baby's First Years we use the study website, the age-four papers by Noble and colleagues (NBER Working Paper 33844, 2025) and Duncan and colleagues (JAMA Pediatrics, 2025), the family-process paper by Magnuson and colleagues (Nature Communications, 2025), the family-investment paper by Gennetian and colleagues (Nature Human Behaviour, 2024) and the infant brain-activity paper by Troller-Renfree and colleagues (PNAS, 2022).

Findings: Stockton (SEED)

Design

SEED gave 125 Stockton residents $500 a month, unconditionally, beginning in February 2019. Recipients had to be at least 18, live in Stockton and live in a neighbourhood with a median income at or below $46,033, the city's median household income; there was no limit on individual household income. The evaluation team mailed invitations to 4,200 randomly selected residences. Of the 125, 100 formed the core research sample, 25 formed a "storytelling" cohort who spoke publicly about the programme, and five additional places were reserved in case of medical attrition. Another 200 residents formed the control group. The peer-reviewed evaluation describes 131 people in the treatment condition and 200 in the control condition, with payments from February 2019 to January 2021 — the programme was extended during the pandemic. The evaluation was led by Stacia West and Amy Castro Baker and funded by the Robert Wood Johnson Foundation's Evidence for Action programme.

Results

The first-year report covers February 2019 to February 2020, before the pandemic. Its headline employment result is that 28% of recipients had full-time employment at the start and 40% a year later, while in the control group full-time employment rose from 32% to 37%. The report itself cautions that these findings cannot point to larger labour-market trends. Recipients also became better able to cover an unexpected $400 expense with cash or its equivalent: the share able to do so rose from 25% to 52%, against 25% to 28% among controls. Food was the largest category of tracked spending; less than 1% of tracked purchases were for tobacco and alcohol. Around 40% of the money loaded onto the prepaid cards was transferred to bank accounts or withdrawn as cash, so tracked spending covers only part of the total.

The Journal of Urban Health article reports that the treatment group had lower income volatility, lower mental distress, better energy and physical functioning, greater agency to explore opportunities related to employment and caregiving, and better ability to weather pandemic-related financial volatility.

Weight of evidence

SEED was influential far beyond its size. With about 100 core recipients and 200 controls, it could detect only large effects, and the full-time employment figures are simple before-and-after shares rather than regression-adjusted estimates with confidence intervals. The atlas treats SEED as credible evidence that $500 a month reduced financial volatility and distress for its participants, and as weak evidence on employment.

Findings: the OpenResearch Unconditional Income Study

Design

The OpenResearch study is the largest randomised guaranteed-income trial in the United States. It randomised 1,000 low-income individuals aged 21–40 to receive $1,000 a month unconditionally for three years, from November 2020 to October 2023, and 2,000 to a control group receiving $50 a month. Participants lived in two US states, with household incomes no higher than 300% of the federal poverty level. The researchers combined detailed surveys, administrative records and data from a mobile phone app. The study was supported by private funding and implemented by non-profit organisations, with the Inclusive Economy Lab at the University of Chicago among its partners.

Work and earnings

The transfer caused total individual income excluding the transfers to fall by about $1,900 a year relative to the control group, and labour-market participation to fall by 4.2 percentage points. Participants reduced their working hours by one to two hours a week, and their partners reduced hours by a comparable amount. The largest increase in time use was leisure. Despite detailed questions about job amenities, the researchers found no effect on the quality of employment, and their confidence intervals rule out even small improvements. They observed no significant effects on degree attainment, although point estimates were larger for younger participants. The authors summarise the result as "a moderate labor supply effect that does not appear offset by other productive activities."

Spending and finances

Bartik and colleagues find that the transfer increased measured household spending by at least $300 a month, with the largest increases for housing, food and car expenses. Recipients were more likely to move house and neighbourhood. Asset values rose modestly, mainly in financial assets, but higher debt offset the gains, leaving net worth close to unchanged. Self-reported financial health and credit scores improved, while credit limits, delinquencies, utilisation, bankruptcies and foreclosures did not change. The authors conclude that large temporary transfers raise short-term consumption and improve financial health but may not durably improve the financial position of young, low-income households.

Health

Miller and colleagues report that the transfer produced large but short-lived improvements in stress and food security, and increased use of hospital, emergency department and dental care. It did not improve self-reported physical health, clinical biomarkers or mental health after the first year, and the researchers can rule out small effects on many outcomes. They found no effects on the health of children in participants' households.

Wellbeing

Measures of subjective wellbeing were higher among treated participants in the first year but then returned to control-group levels. This contrasts with the German pilot, where wellbeing gains were stable for three years, and is one of the most important open puzzles in the literature.

Findings: Cook County Promise Guaranteed Income Pilot

Design

Cook County, Illinois, which includes Chicago, held an open application period in October 2022 and selected participants by lottery in November 2022. The pilot paid $500 a month, unconditionally, to 3,250 low- to moderate-income families for two years; payments began in December 2022 and continued through January 2025. It was funded through the county's allocation under the American Rescue Plan Act.

Early findings

In April 2025 the county announced early findings from participant surveys. 75% of participants reported feeling more financially secure; 94% had experienced a financial emergency or unexpected expense and used programme funds to manage it; 73% believed the payments would continue to affect them after the programme ended; 56% reported reduced stress; and 70% said the programme had a positive effect on their mental health.

Weight of evidence and policy follow-through

These figures describe recipients' own perceptions; the county page does not report a comparison with non-recipients, so they cannot be read as causal effects. The county engaged the University of Chicago's Inclusive Economy Lab to evaluate the pilot and said in April 2025 that quantitative and qualitative findings would be published in fall 2025; as of October 2026 the county's programme pages still reported only the April 2025 survey results. Their policy significance lies elsewhere. In November 2025 the Cook County Board of Commissioners approved $7.5 million for guaranteed income in its fiscal year 2026 budget, and the county's Bureau of Economic Development was designing the next phase, expected to be complete in late 2026. Cook County is therefore one of the few cases in which a pandemic-funded pilot has led to continued public funding — on a much smaller scale than the original pilot's 3,250 families at $500 a month.

Findings: Baby's First Years

Design

Baby's First Years recruited 1,000 mothers with incomes at or near the federal poverty line from 12 hospitals in four communities — New York City, greater New Orleans, the Twin Cities and the Omaha metropolitan area — between May 2018 and June 2019. Mothers were randomised to receive either $333 or $20 a month, paid onto a debit card from the child's birth until three months after the child's sixth birthday — 76 monthly payments in all. The study is funded by the National Institutes of Health and a consortium of foundations, collects data at 12, 24, 36, 48, 72 and 96 months of age, and began its age-eight data collection in July 2026.

Early brain-activity results

At age one, Troller-Renfree and colleagues (2022) reported that infants in the high-cash group showed more power in high-frequency bands of brain activity measured by electroencephalography (EEG) — a pattern associated in other research with later cognitive skills. The authors noted that effect sizes were similar to those of many scalable education interventions, but that statistical significance varied with the analytic specification. The study attracted wide attention as evidence that poverty reduction could change early brain development.

Results at age four

The age-four results are less encouraging. Analysing 891 families, Noble and colleagues (2025) find no statistically significant effects of the cash transfers on the four pre-registered primary outcomes — language, executive function, social-emotional problems and high-frequency brain activity — nor on three secondary outcomes. In JAMA Pediatrics, Duncan and colleagues (2025) report that the $333 monthly transfer did not improve maternal physical or mental health or the physical health of four-year-old children; effect sizes for maternal depressive symptoms, anxiety, maternal body mass index, child body mass index percentile and overall child health were all small and statistically insignificant. The authors offer several explanations: no causal link between income and these outcomes, effects that may appear later, or an income increase of about 18% that is too small to offset the structural disadvantages associated with poverty.

Family investments and processes

The cash did change how families spent money and time. High-cash families had net incomes about $235 a month higher than low-cash families over the first four years, reducing the share at or below the official poverty line from 66% to 61%. They spent on average $47 a month more on child-specific goods — books, toys and clothes — and mothers spent about 13 more minutes a week (a 7% increase) reading, telling stories and playing with the child. Magnuson and colleagues (2025) find, over the first three years, a moderate increase in household income and reductions in poverty, but no statistically significant improvement in reported economic hardship or in the quality of play with infants, and small, mostly insignificant increases in parental psychological distress.

Comparing the pilots

Sources: SEED (2021); West and Castro (2023); Vivalt et al. (2026); Miller et al. (2024); Cook County (2025); Noble et al. (2025); Duncan et al. (2025).

Reading the pattern

Why do Stockton and OpenResearch differ on work?

SEED reported more full-time work among recipients; OpenResearch found less work. Several explanations are possible, and the evidence does not let us choose between them with confidence. SEED's sample was small and its employment result descriptive, so chance cannot be ruled out. OpenResearch paid twice as much for longer, which strengthens income effects. The trials covered different periods: SEED's first year preceded the pandemic, while OpenResearch began in November 2020. And OpenResearch used administrative records, which are less prone to reporting differences between groups. On the evidence available, the atlas gives more weight to the OpenResearch result.

Why do wellbeing effects fade in some trials and not others?

OpenResearch found that wellbeing gains disappeared after the first year; the German pilot found stable gains over three years; Finland measured wellbeing only once. One hypothesis is that the German participants, who were employed with moderate incomes, used the money to build savings and autonomy, while the American participants faced financial pressures that a temporary transfer did not resolve. Another is that measurement and sample differences explain much of the gap. Neither has been tested directly.

A caution about selective attention

Early pilot results that were positive received wide publicity; later null or negative results from larger trials have received less. The atlas treats all the results above as part of one body of evidence. The editorial note that accompanied the Baby's First Years health paper in JAMA Pediatrics, titled "There Are No Sacred Deer in Science", makes the same point about the duty to publish and discuss null results.

What the evidence does not show

• It does not show the effects of a permanent or universal programme. All four pilots were time-limited and targeted. People respond differently to a temporary windfall than to a permanent entitlement, and a universal programme would also change prices, wages and taxes.
• It does not show that cash has no long-run effects on children. Baby's First Years continues to age eight. Effects on education or earnings could appear later, or not at all.
• It does not show that guaranteed income "works" or "fails" in general. The answer depends on the objective. If the goal is to reduce hardship and volatility, the evidence is supportive. If the goal is to raise employment, improve health or close developmental gaps, the strongest trials do not support it at the amounts tested.
• It does not show causal effects for Cook County. The published early findings are participant self-reports without a comparison group.
• It does not settle the labour-supply question for all groups. OpenResearch covers adults aged 21–40 below 300% of the poverty level in two states; the German pilot covers single, employed adults; neither speaks directly to older workers, families with many children or regions with high unemployment.

Implications for institution design

1. Match the instrument to the objective. Unconditional cash is a good tool for reducing financial volatility and hardship. It is not, on current evidence, a reliable tool for improving health or child development by itself. Policies that promise all of these at once set themselves up for disappointment.
2. Expect a modest work response and plan for it. The best American evidence suggests that $1,000 a month reduces work by one to two hours a week, for participants and their partners. This is far from the collapse critics predicted and far from the increase some advocates claimed. Income floors should be designed and costed with such effects in mind.
3. Pre-register, use records and publish everything. The trials that changed the evidence base were those with pre-registered outcomes, administrative data and commitments to publish null results. Public funders of pilots should require these standards.
4. Distinguish pilots from programmes. Cook County's move from a pilot to continued funding shows how quickly politics can run ahead of evaluation. Permanent programmes should be built with evaluation designed in from the start, for example through phased rollout by lottery.
5. Count financial resilience as an outcome. Lower income volatility, better ability to handle emergencies and higher credit scores are real gains, even where health and employment do not change.

Open questions

• Will the Baby's First Years results at ages six and eight show effects that were not visible at age four?
• Why did OpenResearch find fading wellbeing effects while the German pilot found stable ones?
• What will rigorous evaluations of Cook County's pilot, comparing recipients with non-selected applicants, show?
• How do effects change with the size of the payment relative to income? Few trials vary the amount within the same population.
• Would results differ if payments were permanent and universal within a community, allowing local spillovers to be measured as in the Kenyan experiment?

Sources

1. Stockton Economic Empowerment Demonstration (2021). Preliminary Analysis: SEED's First Year. PDF
2. West, S. and Castro, A. (2023). "Impact of Guaranteed Income on Health, Finances, and Agency: Findings from the Stockton Randomized Controlled Trial." Journal of Urban Health 100(2): 227–244. doi:10.1007/s11524-023-00723-0
3. Vivalt, E., Rhodes, E., Bartik, A. W., Broockman, D., Krause, P. and Miller, S. (2026). "The Employment Effects of a Guaranteed Income: Experimental Evidence from Two U.S. States." Quarterly Journal of Economics 141(4): 2675–2734. doi:10.1093/qje/qjag042 (first circulated as NBER Working Paper 32719, 2024)
4. Miller, S., Rhodes, E., Bartik, A. W., Broockman, D., Krause, P. and Vivalt, E. (2024). Does Income Affect Health? Evidence from a Randomized Controlled Trial of a Guaranteed Income. NBER Working Paper 32711. doi:10.3386/w32711
5. Bartik, A. W., Rhodes, E., Broockman, D., Krause, P., Miller, S. and Vivalt, E. (2024). The Impact of Unconditional Cash Transfers on Consumption and Household Balance Sheets: Experimental Evidence from Two US States. NBER Working Paper 32784. doi:10.3386/w32784
6. Cook County Bureau of Economic Development. "Cook County Promise Guaranteed Income Pilot." cookcountyil.gov; and Cook County (2025). "President Preckwinkle Unveils Next Phase of Guaranteed Income in Cook County." 8 April 2025. cookcountyil.gov
7. Baby's First Years. Study website and publications. babysfirstyears.com
8. Noble, K. G., Magnuson, K., Duncan, G. J. et al. (2025). The Effect of a Monthly Unconditional Cash Transfer on Children's Development at Four Years of Age: A Randomized Controlled Trial in the U.S. NBER Working Paper 33844. doi:10.3386/w33844
9. Duncan, G. J., Magnuson, K. A., Kunin-Batson, A. S. et al. (2025). "Cash Transfers and Their Effect on Maternal and Young Children's Health: A Randomized Clinical Trial." JAMA Pediatrics 179(8): 867–875. doi:10.1001/jamapediatrics.2025.1612
10. Magnuson, K., Duncan, G. J., Yoshikawa, H. et al. (2025). "Effects of unconditional cash transfers on family processes and wellbeing among mothers with low incomes." Nature Communications 16: 7517. doi:10.1038/s41467-025-62438-x
11. Gennetian, L. A., Duncan, G. J., Fox, N. A. et al. (2024). "Effects of a monthly unconditional cash transfer starting at birth on family investments among US families with low income." Nature Human Behaviour 8: 1514–1529. doi:10.1038/s41562-024-01915-7
12. Duke Center for Child and Family Policy (2025). Regular, Monthly Unconditional Cash Gift Increases Families' Time and Money on Children from Infancy through Age 4. Research brief, March 2025. PDF
13. Troller-Renfree, S. V., Costanzo, M. A., Duncan, G. J. et al. (2022). "The impact of a poverty reduction intervention on infant brain activity." Proceedings of the National Academy of Sciences 119(5). doi:10.1073/pnas.2115649119
14. Verho, J., Hämäläinen, K. and Kanninen, O. (2022). "Removing Welfare Traps: Employment Responses in the Finnish Basic Income Experiment." American Economic Journal: Economic Policy 14(1): 501–522. doi:10.1257/pol.20200143
15. Bohmann, S., Fiedler, S., Kasy, M., Schupp, J. and Schwerter, F. (2025). "Pilotprojekt Grundeinkommen: kein Rückzug vom Arbeitsmarkt, aber bessere mentale Gesundheit." DIW Wochenbericht 15/2025. doi:10.18723/diw_wb:2025-15-1