The Headlines:
How Giving $750 a Month to L.A.'s Homeless Proved Basic Income Works. - Julius Miller/LA Magazine, December 20, 2023.
$750 a month was given to homeless people in California. What they reported spending it on is more evidence that universal basic income works. - Natalie Musumeci/Business Insider Business Insider, December 20, 2023,
$750 a month, no questions asked, improved the lives of homeless people. - Doug Smith/Los Angeles Times, December 19, 2023.
What the Researchers say:
“It dispels this myth that people will use money for illicit purposes." - Ben Henwood, who leads the study, as quoted in the Los Angeles Times.
“People used the money better than we could have used it for them. When we trust people, they can resolve their issues. They know what they need.” - Kevin Adler, founder of Miracle Messages, the San Francisco nonprofit that helps distribute the funds to homeless study participants, as quoted in the San Francisco Chronicle.
What’s the Fuss About?
An ongoing study that provides a monthly unconditional stipend to homeless individuals in San Francisco and Los Angeles. Researchers just released a preliminary report on key findings so far (Miracle Money California, November 2023 Interim Report). Some excerpts:
In May of 2022, we began recruitment into a randomized controlled trial to study the impact of Miracle Money (M$), which is a basic income and social support intervention for people experiencing homelessness delivered by the non-profit, Miracle Messages. Since then, 103 people experiencing homelessness in either the San Francisco Bay Area or Los Angeles County started receiving $750 per month for 1 year, based on random selection.
This report provides a description of those who began receiving monthly income and how they spent their money. It also compares how the 69 people who have received at least 6 monthly payments are doing compared to a control group (n=86) who accessed usual homeless services.
Two key findings thus far include that people who received M$ for 6 months are:
Less likely to be unsheltered as compared to those who accessed usual services;
Closer to having enough money to meet all of their basic needs as compared to those who accessed usual services.
Here are a couple charts from the report:
A Few Questions and Comments:
Recruitment began in May 2022. Since then, 103 people started receiving payments. But only 69 have received payments for at least six months. What happened to the other 34 individuals? Why haven’t they received payments for six months? Did they have breaks in payments and, if so, why? Did they start late and, if so, why? Were some replacements for initial participants who dropped out? What is the participant attrition rate so far? A high attrition rate would cast a shadow on these findings.
Actually, “findings” is too strong a word. The data is all self-report and thus subject to desirability bias. The homeless individuals in the study know the researchers want the intervention to be successful. They form relationships with the “phone buddies” who ask them all these questions. I imagine some participants would hesitate to tell the whole truth and nothing but.
The researchers describe the control group as “those who accessed usual services”. What does that even mean? What services count as the '“usual services”? When did they receive these services? Did they access these services more than once?
Note that in the above charts, the researchers are using a “waitlist control” group. That means the control group is not receiving social support services from the nonprofit that disburses funds to participants in the “treatment” condition.
Per the study protocol, participants who receive a monthly stipend are also matched with a “phone buddy” for ongoing social support, which may include referral services in case of urgent need. Since the intervention has two parts, there needs to be a way to disentangle the effects of each part. The best way to do that would be to have a control group that gets the buddy support only. That was the initial intention in the study protocol and I’m not sure why they settled for a waitlist control only.
What percentage of the waitlist controls participated in the follow-up surveys? That’s important to know because waitlist controls may be less motivated to complete ongoing surveys than basic income participants with a phone buddy. And you can’t generalize about the control group unless most of them complete the surveys.
The question is not whether receiving $750 monthly payment helps individuals meet their basic needs, but whether a basic income program does a better job than other programs for the homeless. This study sheds no light on that issue.
The study also does not address whether an unconditional (“no strings attached”) program would be better than a conditional cash transfer program. Ironically, in the study protocol, the researchers point to a systematic review of cash transfer programs as evidence that a “guaranteed basic income” has been successful in alleviating poverty and other societal ills. But the review in question - Cash transfers: what does the evidence say? - notes “evidence that making transfers conditional on certain behaviours or actions can positively affect the outcomes relating to the conditions on which the transfers are conditioned”.
The researchers have absolutely no business saying that preliminary self-report results “dispels any myths” about anything. Too little information, too many questions unanswered. And too much agenda-driven self-promotion.
Yet the news media plays along, boosting the boosters, and asking no hard questions.