Skip to content

Basic Income for the Homeless: Findings from Three Experiments

Basic Income for the Homeless: Findings from Three Experiments

What happens when you give cash to people who are experiencing homelessness?

This is the second piece in an ongoing series in which we explain how local housing costs cause homelessness, why solutions to homelessness should be funded from a land value tax, and the merits of redistributing land values through a universal basic income.

Part 1: “In a good economy, homelessness goes up”: The Paradox of Progress and Poverty 

Part 2: Basic Income for the Homeless: Findings from Three Experiments 👈 (you are here)

Part 3: Universal Land Dividend: Why LVT and UBI are even better together

Part 4: Citywide Land Dividend: Can a LVT-funded UBI help solve homelessness?

Cities grappling with rampant homelessness are frantically searching for policies to help people get housed and stay housed. These include building public housing, subsidizing housing costs, and the Housing First approach of providing permanent and supportive housing directly to people experiencing homelessness.

More recently, a few cities have experimented with providing direct cash transfers to people experiencing homelessness. Referred to as a basic income (BI), these programs are defined by their ‘no strings attached’ approach whereby recipients are guaranteed a one-off or periodic cash payment, safe in the knowledge that they will not lose this support if they get a job or find housing. For example, a non-profit run by Google recently announced a new program in San Francisco, where families at risk of homelessness will be randomly selected to receive $1,000 per month over the course of a 5-year pilot program.

To provide some insights into the effects of these programs, this article describes the key findings from three recent experiments with providing BI to homeless individuals in Vancouver BC, Los Angeles CA, and Denver CO.

Basic Income: Benefits and Downsides
Readers may be familiar with the idea of a universal basic income (UBI), such as the $1,000 per month promised to all Americans by Andrew Yang during his 2020 presidential campaign. Because the above programs were provided only to a small number of people experiencing homelessness, they are not truly universal, and we therefore refer to them only as BI programs. However, the merits of these BI programs often mirror those of UBI programs, which we highlight here.

BI payments do not abate if recipients’ income or living standards improve, reducing the need for invasive means-testing, and avoiding situations where recipients find that an additional $10 of gross income gets cut down to only a few dollars in their pocket by the combination of taxes and reduced benefits (often called welfare cliffs). Likewise, the receipt of cash can protect welfare recipients from discrimination, which is a longstanding issue for households using Section 8 rental vouchers.

Guy Standing argues that BI creates an automatic mechanism for ensuring that everyone in society enjoys a share of the benefits of social & economic progress. This helps to compensate those who are harmed by social progress, such as those households struggling to afford rents because their city has become more hotly-demanded. In the case of BI paid to a city’s homeless, this can be likewise seen as compensation for “the structural processes that render housing unaffordable to an increasing number of people(Clarke 2022).

Applying this lens to the specific case of BI used to address homelessness, Andrew Clarke (2022) concludes that BI can help with the income side of the problem, essentially by helping poor households afford a minimum standard of shelter. However, he argues that such an approach still “leaves BI recipients to rely on the market to meet their needs and that it must still be paired with ‘supply side’ interventions to ensure sufficient availability of stable and affordable housing for “low-income households and vulnerable groups like the homeless.” The relative importance of these two sides of the problem will be further unpacked in a later article.

Bearing these merits and downsides in mind, the rest of this article examines the measured benefits of three BI experiments in Vancouver BC, Los Angeles CA, and Denver CO. We note again that these were not truly universal BI, as they were only provided to homeless individuals. However, their results may still be indicative of the specific benefits of BI for people experiencing homelessness.

Experiment #1: New Leaf Project in Vancouver BC
In Vancouver, British Columbia, the New Leaf Project distributed a one-off payment of CAD$7,500 to 50 people who had been homeless for fewer than two years and who they determined did not have a severe addiction or mental illness. Recipients were randomly selected and compared to a control group of non-recipients.

Over the year following the BI payment, recipients spent 99 fewer days unhoused while also reducing their use of shelter services. They had increased spending on necessities such as food, rent and transportation, with no increase in spending on temptation goods like alcohol, drugs, and cigarettes. Perhaps a unique feature of paying the BI as a large lump-sum rather than smaller monthly transfers, recipients spent more on durable consumer goods, and increased their savings by $1,200, equal to 16% of the BI.

Recipients of the New Leaf BI used fewer shelter services, which was estimated to have reduced the public cost of providing these services by $8,277 per BI recipient. This was $777 more than the cost of the BI, suggesting that such a program could help to cut public costs while also improving outcomes for recipients. However, it may be the case that when New Leaf recipients purchased housing for themselves, they outbid other individuals vying for the same stock of housing, displacing those individuals onto the street and into shelter services and undermining the cost-effectiveness of programs like New Leaf. As we discuss below, future research should attempt to trace the extent to which such a mechanism occurs.

Recipients of the New Leaf BI spent fewer days homeless and used fewer shelter services, while spending more on rent, food and transit. Source: Dwyer et al. (2023).

Experiment #2: LA & Bay Area Miracle Money (M$)
For one year starting in May 2022, the Miracle Money experiment gave $750 per month to 100 people experiencing homelessness in Los Angeles and the San Francisco Bay Area. Recipients were randomly selected from a pool of people who had expressed interest in participating, enabling direct comparison to a treatment group of similarly unhoused individuals. Participants were surveyed every three months on a range of outcomes relating to their housing status, health and wellbeing.

Six months into the experiment, Miracle Money (M$) recipients had more secure housing: the share reporting that they had spent at least one night unsheltered over the prior month had fallen from 30% to 12% (compared to 23% for the control group). Recipients were also more likely to report that they had enough money to meet their needs, with the M$ being spent mostly on food (37%), housing (20%), transportation (13%), clothing (11%), and healthcare (6%). Only 2% was spent on alcohol, cigarettes or drugs. A full set of results from this BI experiment are expected to be published in 2024.

Recipients of Miracle Money spent less time unsheltered compared to people stuck on the waitlist. Source: Miracle Money California – Interim Report (2023)

Experiment #3: Denver Basic Income Project
Over in Colorado, the Denver Basic Income Project (DBIP) randomly assigned 800 homeless in Denver to three groups. Starting in October 2022 and running for one year, Group A received $1,000 per month, Group B received $6,500 up-front plus $500 per month, and Group C received $50 per month.

While full results are expected in June 2024, the six-month follow-up study found that all groups were less likely to be sleeping outside, with the greatest reduction among Groups A and B, who received the most money. Both of those groups also indicated reduced reliance on homeless support services, compared to a mild uptick for Group C. Likewise, the share of recipients who were living in their own home/apartment increased substantially, by 26% for Group A, 35% for Group B, and 20% for Group C. Mark Donovan, Executive Director of DBIP, emphasizes the effectiveness of BI I in reducing homelessness, stating in an interview with RSF that “basic income is housing policy.”

DBIP participants who received more money (Group A and B) were more likely to find permanent housing. Source: Brisson et al. (2023)

Other positive outcomes were also observed. Recipients of the Denver BI indicated improved feelings of safety while sleeping, and reductions in feelings of distress and anxiety. Six months after receiving their $6,500 lump-sum transfer, Group B were significantly more likely to be working full-time, up from 21% to 35% over that period. While these preliminary results are highly promising, a full set of results from this groundbreaking program are expected in June 2024.

DBIP is currently working to extend this pilot program so that longer-term outcomes can be observed, with a view towards demonstrating replicability and scalability of the basic income approach to mitigating homelessness.

A Brief Word of Caution

Preliminary results from these BI experiments are highly promising. Providing a guaranteed basic income to homeless individuals appears to help them become housed, reduce their reliance on homelessness support services, purchase groceries and transportation, find gainful employment, and improve their sense of wellbeing. Clearly, policymakers should give serious consideration to expanding public support for such programs.

However, it is important to note that these results are measured at the level of the homeless individuals themselves, rather than providing proof that a BI can reduce the structural amount of homelessness in the city overall. Like a game of musical chairs, it is possible that a homeless recipient of BI may simply outbid an individual who was previously at the margins of being unable to afford rent, pushing that new person out into homelessness instead. If this is the case, then such BI programs will accidentally become a game of whack-a-mole, helping a series of homeless individuals back onto their feet, but never fundamentally reducing the level of homelessness in a given city.


Early experiments provide strong indications that a guaranteed basic income can help homeless individuals become housed and rebuild happy and productive lives. Further research should attempt to measure the impact of such policies on the overall level of homelessness at the city level as well, to provide evidence into whether a truly universal BI could also be effective in helping combat many cities’ homelessness problems.

As explained in a previous article, there is a strong case for using a land value tax (LVT) to fund homelessness solutions, including policies such as a BI for homeless individuals, or a UBI for all city residents. In a subsequent article, we will unpack the various theoretical routes by which such a LVT-funded UBI could help serve as a comprehensive solution to homelessness.

Leave a Reply

Your email address will not be published. Required fields are marked *