Well, no, we’ve never been able to test UBI. That would require the entire population of significant geographic areas to receive UBI levels of income in a way they start believing it’s a safe thing to expect for the foreseeable future, and to model how it’s funded rather than just how it pays out.
What we’ve done is frequently means test the experiments, deliberately select low income people, but only a tiny portion of a larger low income population. Also, the participants know very well that the experiment might be a few months or a year, but after that they’ll be on their own again, so they need to take any advantage it gives them. So all the experiments prove is that if you give some, but not all, low income people a temporary financial benefit, they can and will out compete others without the benefit.
UBI might be workable, or it might need certain other things to make it workable, or it might not be workable, but it’s going to be pretty much impossible to figure it out in a limited scope experiment.
The Alaska permanent fund is about as close to UBI as we’ve gotten, but the amounts are below sustenance living so it’s not up to the standard either.
What we’ve done is frequently means test the experiments, deliberately select low income people, but only a tiny portion of a larger low income population.
So what you’re saying it we explicitly looked at the most extreme examples and seen how UBI has greatly benefitted the people in those extreme situations, and every single time the experiments are conducted the results are pretty consistent, but we can’t extrapolate that it won’t work in less extreme situations because… reasons…
Because you still have the element of differential compared to others. In true UBI, the UBI recipient would represent the ‘low point’ for any citizen. Let’s take Seattle for example as they recent had an ‘experiment’ about UBI. If you had true UBI, then 750,000 people would all get same benefit, of which 75,000 were unemployed. In the UBI experiment, 100 of that 75,000 people had the benefit temporarily, and have an advantage over 74,900 people without that benefit, and the experiment only influences 0.01% of the population in general and then only by a meager amount, so the general local economy won’t even register the activity as a blip. Those 100 people can have a breather but know that time is short. So they take advantage to maybe take a class, get nice interview clothes, and show up better prepared for a job than maybe the other dozen applicants that couldn’t afford to buy the clothes, take time off for the right interview, or take that class. They might not have any particular advantage if everyone had UBI, and the experiment measured success in terms of relative success over those not in the cohort.
So we are missing:
What is the behavior if UBI is taken for granted as a long term benefit for the forseeable future, rather than a temporary benefit.
What is the competitive picture if 100% of the population have the same benefit rather than 0.01%. i.e. how much of it was success owing to better resources versus success owing to others needing to fail to allow that relative success.
What is the overall economic adjustment if 100% of the population has this income and participants in the economy may adjust
What does it look like when the funding model in terms of taxation resembles what is needed in a UBI
Just like all sorts of stuff in science, at scale does not necessarily map to small scale observations. Especially in economic and social science. That’s not to say UBI is definitely not going to work, it’s that we can’t know how it will work/not work until done “for real” at the appropriate scale.
It’s pretty typical, select a few people out of thousands to receive a temporary benefit and extrapolate to UBI. Sure they didn’t talk about the other population, but Internet search does generic demographics for the city (750k total/75k on unemployment).
I appreciate the thoroughness and acknowledgement about limitations:
There is an obvious research evidence
gap in the evaluation of an experimental,
sustained UBI, which is considered the
‘gold standard’ for evidence. There is a
shortage of evidence that meets most or
all of the definitional features of a UBI, and
the interventions covered by this report
vary significantly. To arrive at conclusions
at what may occur if all core features
were unified into UBI policy, reviews have
synthesized evidence from interventions
that may not meet the most stringent
definitions of universality or unconditionality.
But even with the gaps of experiments missing some pieces, the meta analysis still has a really strong positive outlook on UBI despite the imperfections of the numerous studies.
The problem is that the imperfections are systematic and the same imperfections inflict every experiment.
To use an extreme, let’s say you just up and gave someone 10 million dollars. Would they do pretty well? Probably would live it up pretty well or die trying. Assuming they avoided vices that would kill them, they would probably have whatever sort of house they want, any sort of car they want, and so on. They might engage in some noble occupation now that they don’t care about money that sounds nicer than most jobs to make money. You could repeat this same experiment dozens and dozens of time and come away with the same conclusion, that giving people money without conditions improves their situation. At 10 million, you probably wouldn’t see the same data about ‘gainful employment’ as you do with $5,000 dollar experiments, where they know that won’t get them going, but other facets should reproduce.
What if you gave 340 million people 10 million dollars? Well, the economy would adjust to deal with that new normal. You would have massive inflation to compensate for the glut of cash. The market for Ferrari’s jumps a thousand-fold, but they aren’t going to be making a thousand-fold more cars, they’ll just price them up to the point where even millionaires can’t afford them anymore.
What if you gave 340 million people 4 dollars a month? Nothing would happen, it’s too meager to move the needle.
So a UBI would be somewhere in the middle. Now the question is whether there is a sweet spot with the right impact. Economically, it’s likely that their opportunities in ‘objective’ terms may be about the same before and after, but they might feel less crappy about it. Though over long enough time they might get pissed that their UBI payment isn’t going to be even sustenance income even if it started that way.
The other common trend in these is some romanticism of some jobs over others. He was stuck mowing public parks, but with the UBI experiment, he was able to better himself and get a more respectable job. Good for him, but who is now mowing those parks? Oh someone who wasn’t in the UBI experiment. So if everyone has UBI, you can’t have that nice story for everyone, someone is still going to be doing that “lame” mowing job.
UBI might be a part of dealing with a hypothetical labor surplus when we just don’t know what people should be doing, but it’s going to be rough. If it’s too low, you’ve doomed people relying on UBI to forever be stuck with crappy standard of living because there would be no jobs for some of them to have the chance to supplement the income. If it’s too high, then why would anyone ever mow that public park?
I suspect shorter and shorter work weeks has to be a mandated solution, rather than some fixed money amount injected per capita. Spread those crap jobs over more people. Instead of unrelenting 40 hours of landscaping a week, a couple hours a week and surplus labor means you get 20-fold more people doing the job. Though first we have to actually have an obvious labor surplus to get that going.
See my reply to your other reply. That the programs are all limited in scope and duration and thus cannot possibly speak to scale. The meta analysis says the data that is available is appealing, but acknowledges that almost none of them really hit the criteria for a real UBI in terms of scope, scale, and duration. Particularly:
Only a handful of the interventions covered
by this review are truly unconditional and
universal. In an exhaustive review, Gentilini
and colleagues21 identify only a small
number of schemes that reach everyone
within a geographic region without meansbased or demographic targeting, and
regardless of work history. These included
national schemes in Mongolia and Iran,
dividend transfers in Alaska and the Eastern
Band of the Cherokee Nation, a one-off
transfer to all citizens in Kuwait, and pilots
financed by private contributions and
the non-governmental organizations in
Kenya and Namibia, and by the national
government in India. Several of these
programs are either short-term, or not set to
a level that would meet basic needs
I don’t know that any of them manage to hit both long-term and enough to meet basic needs. Alaska is long term, but it’s well below basic needs, for example.
We still do not have data that would speak to a whole society with UBI.
It seems like we’re in a Catch 22 scenario. Models and experiments only give so much information by the very nature of models and experiments simplifying much more complex problems, and in order to collect the kind of holistic data that would speak to a societal level would require an experiment that is functionally identical to just full implementation. Like experiments can only get so big or go on so long before it just becomes the actual thing itself and is no longer an experiment.
And in regards to your other comment about dialing in from the extremes, instead of making everything a uniform number, we can use an formula and a handful of variables to arrive at a local number. A function that takes in common cost of living costs, such as food, shelter, clothing, utilities, transport, and generates a number that will adequate cover those expenses for a given area plus some extra because something unexpect can always happen.
With respect to formula, one aspect that proves difficult is that as you derive the right number, the right number changes. Since currency is kind of a synthetic mathematical trick we play on ourselves to “do economy”, the things being modeled change when we try to force the numbers to be pleasant. Psychology plays a role to potentially make people feel better even with objectively similar circumstances (eg getting a 2% raise along with 2% inflation the person feels like they made some progress despite sitting still.
In any event, I don’t have data either, but I just strongly suspect a numerical manipulation of money balances won’t suffice and we will have to intervene with things like universal healthcare, housing initiatives, labor regulation, and some means of mitigating the phenomenon of billionaires. Easier said than done, but broadly just thinking we have to mind the details explicitly.
Well, no, we’ve never been able to test UBI. That would require the entire population of significant geographic areas to receive UBI levels of income in a way they start believing it’s a safe thing to expect for the foreseeable future, and to model how it’s funded rather than just how it pays out.
What we’ve done is frequently means test the experiments, deliberately select low income people, but only a tiny portion of a larger low income population. Also, the participants know very well that the experiment might be a few months or a year, but after that they’ll be on their own again, so they need to take any advantage it gives them. So all the experiments prove is that if you give some, but not all, low income people a temporary financial benefit, they can and will out compete others without the benefit.
UBI might be workable, or it might need certain other things to make it workable, or it might not be workable, but it’s going to be pretty much impossible to figure it out in a limited scope experiment.
The Alaska permanent fund is about as close to UBI as we’ve gotten, but the amounts are below sustenance living so it’s not up to the standard either.
So what you’re saying it we explicitly looked at the most extreme examples and seen how UBI has greatly benefitted the people in those extreme situations, and every single time the experiments are conducted the results are pretty consistent, but we can’t extrapolate that it won’t work in less extreme situations because… reasons…
Because you still have the element of differential compared to others. In true UBI, the UBI recipient would represent the ‘low point’ for any citizen. Let’s take Seattle for example as they recent had an ‘experiment’ about UBI. If you had true UBI, then 750,000 people would all get same benefit, of which 75,000 were unemployed. In the UBI experiment, 100 of that 75,000 people had the benefit temporarily, and have an advantage over 74,900 people without that benefit, and the experiment only influences 0.01% of the population in general and then only by a meager amount, so the general local economy won’t even register the activity as a blip. Those 100 people can have a breather but know that time is short. So they take advantage to maybe take a class, get nice interview clothes, and show up better prepared for a job than maybe the other dozen applicants that couldn’t afford to buy the clothes, take time off for the right interview, or take that class. They might not have any particular advantage if everyone had UBI, and the experiment measured success in terms of relative success over those not in the cohort.
So we are missing:
Just like all sorts of stuff in science, at scale does not necessarily map to small scale observations. Especially in economic and social science. That’s not to say UBI is definitely not going to work, it’s that we can’t know how it will work/not work until done “for real” at the appropriate scale.
I’m gonna need citations if you’re going to make claims about the data of experiments.
Well, here is the Seattle one: https://www.businessinsider.com/seattle-ubi-guaranteed-basic-income-low-income-poverty-housing-employment-2024-4
It’s pretty typical, select a few people out of thousands to receive a temporary benefit and extrapolate to UBI. Sure they didn’t talk about the other population, but Internet search does generic demographics for the city (750k total/75k on unemployment).
Or maybe Meta Analysis is more convincing.
https://basicincome.stanford.edu/uploads/Umbrella Review BI_final.pdf
Page 15 is rather quite enlightening.
I appreciate the thoroughness and acknowledgement about limitations:
But even with the gaps of experiments missing some pieces, the meta analysis still has a really strong positive outlook on UBI despite the imperfections of the numerous studies.
The problem is that the imperfections are systematic and the same imperfections inflict every experiment.
To use an extreme, let’s say you just up and gave someone 10 million dollars. Would they do pretty well? Probably would live it up pretty well or die trying. Assuming they avoided vices that would kill them, they would probably have whatever sort of house they want, any sort of car they want, and so on. They might engage in some noble occupation now that they don’t care about money that sounds nicer than most jobs to make money. You could repeat this same experiment dozens and dozens of time and come away with the same conclusion, that giving people money without conditions improves their situation. At 10 million, you probably wouldn’t see the same data about ‘gainful employment’ as you do with $5,000 dollar experiments, where they know that won’t get them going, but other facets should reproduce.
What if you gave 340 million people 10 million dollars? Well, the economy would adjust to deal with that new normal. You would have massive inflation to compensate for the glut of cash. The market for Ferrari’s jumps a thousand-fold, but they aren’t going to be making a thousand-fold more cars, they’ll just price them up to the point where even millionaires can’t afford them anymore.
What if you gave 340 million people 4 dollars a month? Nothing would happen, it’s too meager to move the needle.
So a UBI would be somewhere in the middle. Now the question is whether there is a sweet spot with the right impact. Economically, it’s likely that their opportunities in ‘objective’ terms may be about the same before and after, but they might feel less crappy about it. Though over long enough time they might get pissed that their UBI payment isn’t going to be even sustenance income even if it started that way.
The other common trend in these is some romanticism of some jobs over others. He was stuck mowing public parks, but with the UBI experiment, he was able to better himself and get a more respectable job. Good for him, but who is now mowing those parks? Oh someone who wasn’t in the UBI experiment. So if everyone has UBI, you can’t have that nice story for everyone, someone is still going to be doing that “lame” mowing job.
UBI might be a part of dealing with a hypothetical labor surplus when we just don’t know what people should be doing, but it’s going to be rough. If it’s too low, you’ve doomed people relying on UBI to forever be stuck with crappy standard of living because there would be no jobs for some of them to have the chance to supplement the income. If it’s too high, then why would anyone ever mow that public park?
I suspect shorter and shorter work weeks has to be a mandated solution, rather than some fixed money amount injected per capita. Spread those crap jobs over more people. Instead of unrelenting 40 hours of landscaping a week, a couple hours a week and surplus labor means you get 20-fold more people doing the job. Though first we have to actually have an obvious labor surplus to get that going.
Cool. And here are 150+ other experiments and program of varying sizes and time frames pretty much all with generally positive results.
https://basicincome.stanford.edu/experiments-map/
It seems to me like there actually is quite a lot of data to support the claim that UBI is consistently good when implemented.
See my reply to your other reply. That the programs are all limited in scope and duration and thus cannot possibly speak to scale. The meta analysis says the data that is available is appealing, but acknowledges that almost none of them really hit the criteria for a real UBI in terms of scope, scale, and duration. Particularly:
I don’t know that any of them manage to hit both long-term and enough to meet basic needs. Alaska is long term, but it’s well below basic needs, for example.
We still do not have data that would speak to a whole society with UBI.
It seems like we’re in a Catch 22 scenario. Models and experiments only give so much information by the very nature of models and experiments simplifying much more complex problems, and in order to collect the kind of holistic data that would speak to a societal level would require an experiment that is functionally identical to just full implementation. Like experiments can only get so big or go on so long before it just becomes the actual thing itself and is no longer an experiment.
And in regards to your other comment about dialing in from the extremes, instead of making everything a uniform number, we can use an formula and a handful of variables to arrive at a local number. A function that takes in common cost of living costs, such as food, shelter, clothing, utilities, transport, and generates a number that will adequate cover those expenses for a given area plus some extra because something unexpect can always happen.
With respect to formula, one aspect that proves difficult is that as you derive the right number, the right number changes. Since currency is kind of a synthetic mathematical trick we play on ourselves to “do economy”, the things being modeled change when we try to force the numbers to be pleasant. Psychology plays a role to potentially make people feel better even with objectively similar circumstances (eg getting a 2% raise along with 2% inflation the person feels like they made some progress despite sitting still.
In any event, I don’t have data either, but I just strongly suspect a numerical manipulation of money balances won’t suffice and we will have to intervene with things like universal healthcare, housing initiatives, labor regulation, and some means of mitigating the phenomenon of billionaires. Easier said than done, but broadly just thinking we have to mind the details explicitly.