U.S. Has Run 105 Local Guaranteed Income Experiments. Researchers Reviewed Them.

Over the past decade, cities and counties across the United States tried a bold idea. Give people cash, with no strings attached, and see what happens. These programs went by the name guaranteed income.

A new study steps back from that noise and asks a quieter question. Not whether cash works, but what these programs were actually built to test. The answer reshapes how we should read every result that comes out of them.

The paper is by Sarina Rodriguez and six colleagues. It ran in the journal Basic Income Studies in 2025. The team compiled 105 guaranteed income programs from across the country. Between them, those programs reached more than 40,000 people. Then the researchers lined the programs up side by side and compared them on the basics. Who qualified. Who paid for it? How many people got? Who ran it? How long did it last? And how success was measured.

What the study found

The first is a definition. Guaranteed income is an umbrella label. It covers cash programs that are typically unconditional and targeted at a specific group. Think expectant mothers, or residents below a certain income. Hold onto that word specific. It does a lot of work later.

From there, a few clear patterns emerge. Just over half of the programs set an income limit. To get the money, you had to be poor enough to qualify. That makes these targeted programs, not payments for everyone. Even more striking, 84 percent came with a place-based rule. Your address determined whether you were eligible. Live in the right city or county, and you are in. Live one town over, and you were out.

The programs are also clustered on the map. California ran the most by a wide margin, with 28 of the 105. Sixteen more operated at the county level. The rest were scattered thin across the country.

The last finding is the most revealing. Many of these programs were launched with the language of community. Organizers spoke about dignity, belonging, and local repair. But when it came time to fund them and judge them, the yardsticks changed. Funders and evaluators leaned on economic measures and public health measures instead. Did people work more? Did stress go down? Did health improve? The story a program told about itself at the start was often not the story it was measured against at the end.

It would be easy to read this paper as a bookkeeping exercise. It is not. It quietly pulls the floor out from under a lot of loose talk about basic income.

Start with the gap between framing and measurement. A program can be sold as a way to knit a neighborhood back together and then be graded on whether recipients found jobs. When that happens, the people holding the checkbook decide what counts as success. The original goal fades from view. A community that set out to build trust and belonging gets a report card about employment and blood pressure. Those are worthy things to measure. But they are not the thing the program promised. And because different programs used different yardsticks, the results do not add up to a single clear lesson. We end up with 105 stories that resist being told as a single story.

Now return to that word specific. These programs were targeted and local. More than eight in ten depended on where you lived, and more than half depended on how little you earned. That is worth saying plainly. This is not a test of universal basic income. Universal means everyone, rich and poor, in every zip code, no questions asked. What the country actually ran was a large set of programs for low-income people in particular places. Treating those results as proof for or against a universal program is a mistake. It compares one thing to a very different thing and hopes no one notices.

The geography deepens the problem. When a single state holds the largest share of programs, its conditions shape the evidence. California has its own housing costs, its own politics, and its own safety net. A result that holds in Stockton or Los Angeles may not travel to rural Ohio or the Mississippi Delta. The map is not neutral. It bends what we think we know.

Then there is the sheer fragmentation. One hundred and five programs, each with its own rules, its own payment size, its own length, and its own scorecard. That variety is a strength for local communities, who get to design what fits them. It is a weakness for anyone trying to draw a national conclusion. You cannot cleanly pool studies that never agreed on what they were measuring.

This paper does not tell us whether a guaranteed income works. It tells us something more useful and less comfortable. The evidence base is patchier, more local, and more targeted than the confident headlines suggest.

That is not a reason for despair. If we want to know what cash can and cannot do, the first step is knowing what we have actually tried. So far, we have run many small, targeted, place-bound programs and then talked about them as if they were one grand experiment in giving everyone money. They were not.

The value of this study is that it draws the real map. Anyone who wants to argue about basic income, for it or against it, should start here. Know what was tested before claiming what it proves.

Rodriguez, S., Kagawa, R., Koundinya, V., Choe, D., Vaitla, B., Volzer, A., and Brinkley, C. (2025). “Guaranteed Income: A Policy Landscape Review of 105 Programs in the United States.” Basic Income Studies, 20(1), 93 to 123.