Universal basic income has cycled through serious policy discussion multiple times. The 1960s negative-income-tax debates, the Nixon Family Assistance Plan, the 1970s Mincome experiment in Manitoba, and the 2010s revival driven by tech-sector concerns about automation all shared the same core question: what happens when you give people money with no work requirement? The 2020s evidence base is now substantial enough to answer most of the predictable questions.

The Stockton Experiment

SEED (Stockton Economic Empowerment Demonstration), launched in 2019, gave 125 randomly selected Stockton residents $500 per month for 24 months with no strings attached. The control group received nothing. Results published by the Stanford Basic Income Lab in 2021: recipients showed measurable improvements in full-time employment (up 12 percentage points relative to control), mental health, financial stability, and goal-setting. There was no evidence of work disincentive — if anything, the cash enabled people to take jobs they otherwise could not have accepted because of transportation or childcare logistics.

The Stockton results were small in absolute terms but consistent with what theory predicts when modest cash transfers are provided to working-age low-income adults. They are not consistent with the "people will stop working" narrative that has dominated US welfare debates for fifty years.

GiveDirectly's Kenya Experiments

GiveDirectly's randomized controlled trials in Kenya, beginning in 2011 and continuing through ongoing long-term studies, have the largest sample sizes and longest follow-up windows of any cash-transfer research. The findings:

  • Cash transfers produce substantial improvements in food security, asset accumulation, and child outcomes.
  • The improvements persist for years after the transfers end, suggesting the cash enables productive investments rather than short-term consumption smoothing.
  • Spillover effects on non-recipients in the same community are generally positive — cash transfers create local economic activity that benefits people not receiving the transfers.
  • Work effort is largely unchanged or slightly higher; the recipients do not exit the labor force.

The Kenya experiments are not perfect analogues for US policy — extreme-poverty cash transfers produce different responses than moderate-poverty ones. But the basic finding that recipients use the money productively and do not stop working has held across very different income levels and cultural contexts.

The OpenResearch Study

The OpenResearch (formerly Y Combinator Research) study, launched in 2019 with funding from Sam Altman's foundation, provided $1,000 per month to 1,000 randomly selected low-income participants in Illinois and Texas for three years. The control group received $50 per month.

Results published in 2024: the cash group worked roughly 1.3 hours per week less than the control group, slightly increased self-employment, and used the time partly for caring for children and family members. Earnings from work fell by about $1,500 per year on average, but the cash transfer ($12,000) more than offset this. Mental health, financial security, and food security all improved. There was no evidence of substantial work disincentive — the small reduction in work hours was consistent with people taking time to search for better jobs and care for family.

This is the largest and most rigorous US basic-income experiment to date. The results broadly track the smaller studies. The work-hour reduction is real but modest, and the welfare gains substantially exceed the labor-market losses.

What the Evidence Now Supports

Across the Stockton, Kenya, and OpenResearch studies, plus smaller experiments in Finland, Spain, Brazil, and elsewhere, the empirical record converges on several findings:

  • Cash transfers do not produce mass work withdrawal.
  • Recipients use the money for productive purposes more often than not — paying down debt, investing in skills, starting businesses, covering transportation and childcare to enable work.
  • Mental health improvements are consistent across studies.
  • Effects on children — school attendance, nutrition, behavioral outcomes — are typically positive.
  • The argument that cash transfers create dependency is not supported by the data.

What the Evidence Cannot Tell Us

The studies do not directly address full UBI at population scale. Small experiments cannot reveal what would happen if everyone in a country received a basic income simultaneously. The macro effects — on labor supply, on inflation, on tax incidence — could differ substantially from the partial-equilibrium estimates the experiments provide.

Funding is the other unresolved question. Stockton's SEED was philanthropically funded. OpenResearch is philanthropically funded. National UBI would require tax revenue at scales that would shift the tax incidence substantially, and the political economy of those shifts is the part the experiments cannot illuminate.

The Variations That Aren't Quite UBI

Several real-world programs implement parts of the UBI logic without universality:

  • The Alaska Permanent Fund Dividend ($1,000-3,000 per year per resident, funded by oil revenue, paid since 1982).
  • Expanded child tax credit (US 2021, $3,000-3,600 per child, expired after one year).
  • Earned Income Tax Credit (substantial cash transfer to low-income workers, conditional on work).
  • Various state and local guaranteed-income programs (over 50 documented as of 2024).

The 2021 expanded CTC's effect on child poverty — a one-year drop from 12% to 7%, the largest single-year improvement on record — is the closest US has come to a partial UBI for families with children. Its expiration in 2022 produced an immediate increase in child poverty back toward pre-2021 levels. The political-economy lesson is that broad cash transfers can produce large measurable welfare gains; sustaining them across political cycles is harder than implementing them.

The Honest Reading

The empirical evidence on cash transfers has converged toward a clear conclusion: they work, they do not produce mass work withdrawal, and the welfare gains exceed the labor-market costs in most rigorous evaluations. Full UBI at national scale remains untested, but the component pieces — modest cash transfers, no work requirements, adequate amounts — have been tested enough that the conventional "people will stop working" objection is no longer the strongest argument against UBI. The remaining serious arguments are about funding, scale, and political durability. These are real arguments and they do not have settled answers, but they are different arguments from the ones that dominated welfare debates for the previous fifty years. The intellectual ground has shifted; whether the political ground will follow is the open question.