Giving People Cash Reduces Homelessness. Be Precise About Who It Reaches.
- In the first randomized controlled trial of direct cash for homelessness, a single unconditional payment of CAD$7,500 reduced the number of days recipients spent homeless over the following year, moved them into stable housing faster, and increased spending on food, rent, and transportation — with no increase in spending on alcohol, cigarettes, or drugs.1
- The transfer largely paid for itself. By reducing nights in shelter, it generated an estimated net societal saving of about CAD$777 per recipient — meaning the public spent roughly the same whether it handed someone $7,500 or left them in the shelter system.1
- The catch is the sample. That trial deliberately screened out people with serious substance use or mental health conditions and enrolled only people who had recently become homeless. It is strong evidence that cash works for newly homeless, lower-acuity adults — and silent on the chronically homeless population that dominates the policy debate.
- A larger U.S. study, the Denver Basic Income Project, paid up to $1,000 a month for a year and saw the share of participants in their own housing rise from roughly one in ten to nearly half, alongside hundreds of thousands of dollars in avoided public-service costs.3
- Cash is not a replacement for supportive housing. It is one of the most cost-effective tools available for prevention and early intervention — and it is consistently underused because the public does not trust people experiencing homelessness to manage money, a belief the evidence directly contradicts.
For most of the modern history of homelessness policy, one idea has been treated as obviously naive: just give people the money. The objection is rarely stated as a research question. It is stated as a certainty — that people experiencing homelessness will spend cash on alcohol or drugs, that money without supervision is money wasted, that the problem is too complex for something so simple. For a long time there was little rigorous evidence either way, which let the assumption stand unchallenged. That has changed.
In September 2023, a team led by Ryan Dwyer and Jiaying Zhao at the University of British Columbia published the first randomized controlled trial of direct cash transfers to people experiencing homelessness in the Proceedings of the National Academy of Sciences. The result was not subtle, and it ran directly against the conventional wisdom.
The Vancouver trial: cash worked, and it did not go to vices
The study, run in partnership with the nonprofit Foundations for Social Change through its New Leaf Project, gave a one-time, unconditional payment of CAD$7,500 to 50 people experiencing homelessness in Vancouver. Another 65 people formed a control group and received no cash. Everyone was tracked for a full year.
Over those twelve months, the people who received cash spent fewer days homeless than the control group and moved into stable housing faster. They increased their spending on food, clothing, rent, and transportation. They still had money left — recipients retained savings rather than burning through the lump sum. And on the outcome that anchors the entire skeptical case against cash, there was no effect at all: cash recipients showed no increase in spending on what researchers call "temptation goods" — alcohol, cigarettes, and drugs — relative to the control group.
The cost finding is the one that should change how funders think. By reducing the number of nights recipients spent in shelters, the cash transfer generated an estimated net societal saving of roughly CAD$777 per person over the year. Read that carefully: the public did not pay extra to give people $7,500. The reduced shelter use offset most of the transfer. Society spent about the same money and got better outcomes — people housed faster, with dignity and choice, instead of cycling through emergency beds.
The caveat that the headlines skipped
This is where precision matters, because the Vancouver study is routinely cited as if it proved cash works for "the homeless" as a category. It did not, and the researchers were careful to say so.
The trial enrolled only people who had become homeless relatively recently, and it screened out anyone with serious problems related to substance use, alcohol, or mental health. That was a deliberate, defensible design choice — it isolated the effect of cash for a population most likely to convert money into stability quickly. But it also means the sample looked very different from the chronically homeless population that policy debates usually center on: people who have been unhoused for years, often with co-occurring serious mental illness and addiction.
So the honest reading is narrow and powerful at the same time. For adults who have recently fallen into homelessness and are not in acute behavioral-health crisis, a lump sum of cash is one of the fastest, most efficient interventions ever tested. It is, in effect, a prevention and rapid-exit tool. What the trial cannot tell us is whether the same payment would produce the same result for someone who has been on the street for five years with untreated schizophrenia — and it would be a misuse of the evidence to claim that it does.
Denver: a bigger test, with a messier design
The largest U.S. effort to date is the Denver Basic Income Project, which began in November 2022 and was evaluated by researchers at the University of Southern California. It enrolled more than 800 people experiencing homelessness and split them across three payment groups: one receiving $1,000 a month for a year, one receiving an initial $6,500 followed by $500 a month, and a third receiving $50 a month — a total of $600 over the year.
The one-year results, released in 2024, pointed the same direction as Vancouver. Across all groups, the share of participants living in housing they rented or owned rose from roughly one in ten at the start to nearly half by the end of the year. Nights spent unsheltered fell. Participants reported greater financial stability and less reliance on emergency assistance. The project estimated public-service savings — fewer emergency room visits, shelter nights, and jail bookings — in the range of several hundred thousand dollars.3
Denver also illustrates the limits of what these studies can prove. There was no true no-cash control group; even the comparison arm received $600. And notably, that smallest-payment group improved too. The cleanest interpretation is that cash helps people exit homelessness — but Denver cannot cleanly tell us how much cash, or how the dose changes the result, because every participant got something. That is an honest limitation, not a debunking. It is also a reminder that pilot evaluations, however encouraging, are not the same as the kind of large, well-controlled trial that would settle the dosing question.
Why a tool this effective stays on the shelf
If cash is cheap, fast, and effective for a meaningful slice of the homeless population, the obvious question is why it remains a pilot rather than a policy. Part of the answer is structural: cash does nothing to fix a housing market with too few affordable units, and money cannot be drawn down from a shelter someone cannot find. But the larger barrier is public belief.
The same Vancouver research team measured public attitudes and found deep mistrust: most people surveyed assumed cash recipients would spend the money on temptation goods — the precise outcome the trial showed did not happen.1 Polling consistently finds a majority of Americans oppose a federal guaranteed income, and social-perception studies rank people experiencing homelessness as among the least trusted groups in society. The researchers also found that this mistrust is movable: showing people the actual evidence, or framing transfers in terms of taxpayer savings, measurably increased support.
That gap — between what the evidence shows and what the public assumes — is the real obstacle. Programs get designed around the fear of misuse rather than the data on outcomes, which is how you end up with restricted, surveilled, conditional aid that costs more to administer than it saves.
What this means for practice
The evidence base is still young. The most rigorous trial is Canadian and small; the largest U.S. study is a pilot without a clean control; ongoing trials such as the Miracle Money study in California, pairing monthly payments with peer phone support, have not yet reported final outcomes. Anyone who tells you cash transfers are a proven solution to homelessness writ large is overstating what exists.
But the direction of the evidence is consistent, and the practical implication is concrete. For people who are newly homeless or at imminent risk — the population coordinated-entry systems are often slowest to serve, because they are not "acute" enough to prioritize — direct cash is a strong, cost-neutral intervention that prevents a short crisis from hardening into chronic homelessness. For CoCs and funders, the question is not whether to replace supportive housing with cash. It is whether prevention and rapid-exit dollars are being delivered in the most effective form available, or whether they are being routed through programs designed around a mistrust the evidence does not support.
Give people the money, the old objection went, and they will waste it. The first time someone actually ran the experiment, that is not what happened. It is worth letting the evidence, rather than the assumption, set the policy.
Frequently asked questions
Does giving people cash actually reduce homelessness?
In the first randomized controlled trial of direct cash for homelessness, a one-time unconditional payment of CAD$7,500 reduced the number of days recipients spent homeless over the following year and moved them into stable housing faster. A larger U.S. pilot in Denver pointed the same direction, with the share of participants in their own housing rising from roughly one in ten to nearly half.
Do people experiencing homelessness spend cash on alcohol and drugs?
No — that is the outcome the Vancouver trial directly contradicted. Recipients increased spending on food, clothing, rent, and transportation and retained savings, with no increase in spending on alcohol, cigarettes, or drugs relative to the control group. The public's assumption that money would go to vices is the precise belief the evidence disproves.
Does cash work for everyone experiencing homelessness?
No, and this is the key caveat. The Vancouver trial deliberately screened out people with serious substance use or mental health conditions and enrolled only people who had recently become homeless. It is strong evidence that cash works for newly homeless, lower-acuity adults — and silent on the chronically homeless population that dominates the policy debate.
Does giving people cash cost taxpayers more?
In the Vancouver trial it largely paid for itself. By reducing nights in shelter, the transfer generated an estimated net societal saving of about CAD$777 per recipient, meaning the public spent roughly the same whether it handed someone $7,500 or left them in the shelter system — while getting better outcomes.
Is cash a replacement for supportive housing?
No. Cash does nothing to fix a housing market with too few affordable units, and it is silent on the chronically homeless population. The post frames it as one of the most cost-effective tools for prevention and early intervention — a strong, cost-neutral option for people who are newly homeless or at imminent risk, not a substitute for supportive housing.
Sources & footnotes
- Dwyer, R., Palepu, A., Williams, C., Daly-Grafstein, D., & Zhao, J. (2023), "Unconditional cash transfers reduce homelessness," Proceedings of the National Academy of Sciences 120(36), doi:10.1073/pnas.2222103120 — Vancouver New Leaf randomized controlled trial, including the public-attitudes survey.
- Foundations for Social Change / New Leaf Project program documentation.
- Denver Basic Income Project one-year evaluation, University of Southern California Homelessness Policy Research Institute (2024).
- Urban Institute, "Guaranteed Income as a Mechanism for Promoting Housing Stability" (2022) and Housing Matters analysis of direct cash transfers.
- Miracle Friends / Miracle Money RCT protocol, USC Suzanne Dworak-Peck School of Social Work and Miracle Messages (Trials, 2024) — outcomes pending. Figures are drawn from peer-reviewed and program-published reports; pilot estimates are described as such.