to be successful one would need to guarantee UBI to everyone, without means testing, for credibly unlimited duration, linked to a reasonable index to protect against inflation.
I would like to see a reason stated for any of these assertions. We’re not doing UBI, or unlimited durations (at first at least). We also will be forced to do some means testing because of IRS limitations on 501(c)(3)s, but it will be only a 30-minute application. Our initial experiments will not be indexed to inflation because we still need to figure out the optimal amount to help people.
What do you mean by expensive? It seems clear we’re currently spending a lot more money a lot less impactfully. Our Pilot Budget is $182,000.
The US seems like a poor testing ground for it.
The second half of the post describes in detail why Arizona & homelessness is likely the highest-impact opportunity anywhere in the developed world. I’d like to hear the reasoning.
Hmm, you are certainly infinitely more familiar with the topic. I basically followed Scott Alexander’s blog posts about UBI which is not what you are trying to do here. I agree that giving money directly makes more sense than most alternatives. I am guessing that there is a subset of the population for whom it will work, and won’t just go into, say, drugs and gambling without a noticeable change in quality of life or mortality, but identifying it might be tricky, so you have to accept that a fair amount of money does not do any good. There was a local experiment with a one-time transfer that had some positive results https://forsocialchange.org/new-leaf-project-overview but the homeless situation here is getting progressively worse still.
So we’ve got two major types of guaranteed income experiments, those on homeless individuals, and those on the general (impoverished) public.
I agree with you that the New Leaf Project experiment (that I cited quite a bit) was quite positive, and also that the problem is still getting worse. Although It’s important to note that that experiment only had 115 participants, so it covered only a drop in the bucket.
The real question is, “If guaranteed income was scaled up to cover the entire population in homelessness or in danger of becoming homeless, would the results of the experiment stay true?”. Would it eradicate (mostly at least) homelessness? This has never been done so far, but considering the evidence, it seems to at least be worth trying. Especially because guaranteed income is way less expensive than all other forms of assistance. We spend ~$80,000 on average servicing each homeless person, giving them even $1,000 a month, which is <1/5 the cost. And it looks so far to be substantially more effective.
I am guessing that there is a subset of the population for whom it will work, verkeepingrsus keep neutral or potentially harm.
The important question, as you accurately point out, is, ” How big is that subset of society”.
I think at first, definitely not. I see it playing out like this:
While other forms of assistance are going about things like normal, someone (us) builds a big enough guaranteed income program to provide half or most of the homeless population in an area with guaranteed income.
When that program happens, hopefully, most homeless people attain far better situations within a year, and the existing assistance services find themselves with more resources available to assist fewer people (the ones in highly bad mental states/addiction).
Using guaranteed income to help people in danger of falling into homelessness drastically reduces the rates of people falling into homelessness, and leaves the existing service providers with less and less to do over time, twiddling their thumbs.
Peace and quiet.
Only then, will funding slowly get transitioned to guaranteed income (not all, some 20-30% will probably still be needed for intensive addiction, psychiatric services, and disaster relief). At this point, there will probably need to be major policy changes.
Is that ambitious, yeah, but it all relies on #2. Somebody needs to try it.
I would like to see a reason stated for any of these assertions. We’re not doing UBI, or unlimited durations (at first at least). We also will be forced to do some means testing because of IRS limitations on 501(c)(3)s, but it will be only a 30-minute application. Our initial experiments will not be indexed to inflation because we still need to figure out the optimal amount to help people.
What do you mean by expensive? It seems clear we’re currently spending a lot more money a lot less impactfully. Our Pilot Budget is $182,000.
The second half of the post describes in detail why Arizona & homelessness is likely the highest-impact opportunity anywhere in the developed world. I’d like to hear the reasoning.
Hmm, you are certainly infinitely more familiar with the topic. I basically followed Scott Alexander’s blog posts about UBI which is not what you are trying to do here. I agree that giving money directly makes more sense than most alternatives. I am guessing that there is a subset of the population for whom it will work, and won’t just go into, say, drugs and gambling without a noticeable change in quality of life or mortality, but identifying it might be tricky, so you have to accept that a fair amount of money does not do any good. There was a local experiment with a one-time transfer that had some positive results https://forsocialchange.org/new-leaf-project-overview but the homeless situation here is getting progressively worse still.
So we’ve got two major types of guaranteed income experiments, those on homeless individuals, and those on the general (impoverished) public.
I agree with you that the New Leaf Project experiment (that I cited quite a bit) was quite positive, and also that the problem is still getting worse. Although It’s important to note that that experiment only had 115 participants, so it covered only a drop in the bucket.
The real question is, “If guaranteed income was scaled up to cover the entire population in homelessness or in danger of becoming homeless, would the results of the experiment stay true?”. Would it eradicate (mostly at least) homelessness? This has never been done so far, but considering the evidence, it seems to at least be worth trying. Especially because guaranteed income is way less expensive than all other forms of assistance. We spend ~$80,000 on average servicing each homeless person, giving them even $1,000 a month, which is <1/5 the cost. And it looks so far to be substantially more effective.
The important question, as you accurately point out, is, ” How big is that subset of society”.
Luckily, this is the part where there is infinitely more research and data available, including large-scale experiments to look at. There’s a map with dozens of studies, the results of a large random experiment in Stockton California, and the results from the most comprehensive UBI experiment ever.
If anything, please watch this 15-minute TED Talk. I think you’ll find it fascinating.
Is it intended as a substitute for other forms of assistance?
I think at first, definitely not. I see it playing out like this:
While other forms of assistance are going about things like normal, someone (us) builds a big enough guaranteed income program to provide half or most of the homeless population in an area with guaranteed income.
When that program happens, hopefully, most homeless people attain far better situations within a year, and the existing assistance services find themselves with more resources available to assist fewer people (the ones in highly bad mental states/addiction).
Using guaranteed income to help people in danger of falling into homelessness drastically reduces the rates of people falling into homelessness, and leaves the existing service providers with less and less to do over time, twiddling their thumbs.
Peace and quiet.
Only then, will funding slowly get transitioned to guaranteed income (not all, some 20-30% will probably still be needed for intensive addiction, psychiatric services, and disaster relief). At this point, there will probably need to be major policy changes.
Is that ambitious, yeah, but it all relies on #2. Somebody needs to try it.