(2024-09-03) ZviM On The UBI Paper

Zvi Mowshowitz On the UBI Paper. Would a universal basic income (UBI) work? What would it do? Many people agree July’s RCT on giving people a guaranteed income, and its paper from Eva Vivalt, Elizabeth Rhodes, Alexander W. Bartik, David E. Broockman and Sarah Miller was, despite whatever flaws it might have, the best data we have so far on the potential impact of UBI.

They do not agree on what conclusions we should draw. This is not a simple ‘UBI is great’ or ‘UBI it does nothing.’

I see essentially four responses.

• The first group says this shows UBI doesn’t work. That’s going too far. I think the paper greatly reduces the plausibility of the best scenarios, but I don’t think it rules UBI out as a strategy, especially if it is a substitute for other transfers.
• The second group says this was a disappointing result for UBI
• The third group did its best to spin this into a positive result
• The fourth group was some mix of ‘if brute force (aka money) doesn’t solve your problem you’re not using enough’ and also ‘but work is bad, actually, and leisure is good.’

RTFP (Read the Paper): Core Design

Like all studies of UBI, this can only be done for a limited population, and it only lasts a limited amount of time.

The temporary form, and also the limited scope, means that it won’t cause a cultural shift and changing of norms. Those changes might be good or bad, and they could overshadow other impacts.

Also key is that you don’t see the impact of the cost on inflation or the budget. If you gave out a UBI you would have to pay for it somehow, either printing money or collecting taxes or borrowing. Here you get to measure the upside without the downside. Alternatively, we also aren’t replacing the rest of the welfare state.

And in this study, the people getting the money improve their relative status, not only their absolute position.

We can still learn a lot. I would treat this as a very good test of ‘give people money.’

For every one dollar received, total household income excluding the transfers fell by ‘at least 21 cents,’ and total individual income fell by at least 12 cents or $1,500/year.

There was a 2% decrease in labor force participation and 1.4 hour/week reduction in labor hours, and a similar amount by partners, a 4%-5% overall decline in income.

No impact on quality of employment

20% had college degrees, 92% had finished high school or a GED or a post-secondary program by the endline

Unemployment lasted 8.8 months in the treatment group versus 7.7 for controls. This is a sensible adjustment, if it is coming from a sensible base. There were ~0.3 additional months of unemployment per year for the treatment group.

Especially disappointing: “While treated participants exhibited more entrepreneurial orientation and intentions, this did not translate into significantly more entrepreneurial activity.

Also bad news: “We find a significant increase in the likelihood that a respondent has a self-reported disability

This is a remarkably high number, and it is some mix of way too many health problems and way too many incentives to claim to have them.

Participants appear to spend approximately the full amount of the transfers each month

Also note that this study started during Covid. That could make the results not too representative of what would have happened at another time.

*What did we get for the money?

Before looking at the reactions of others, I would say not enough. There was an entire array of variables that changed very little. The group was given roughly an extra year’s income, and ended up in the same place as before.*

They did see declines in stress and food insecurity in year one, but they faded by year two, potentially in part because the end of the study loomed

Eva Vivalt, lead author of this excellent study, has a good thread here going over many of the results. He is clearly struggling to find positive outcomes

Alex Tabarrok: Important thread on important paper.
tl/dr;
“You have to squint to find any positive effects other than people do more leisure when you give $”.

do think this study scores a big one for Give Parents Money rather than Give People Money. The money seems to have a much bigger impact on children and families, while also raising fertility.

Maxwell Tabarrok talks about this study as well as three others. He summarizes this study’s result as a 20 percent decline in work and not much else happening besides more leisure. He also covers the Denver UBI trial (also $1k/month), which was exclusively in the homeless, and showed that the money didn’t even do much for actual homelessness there – people found housing often but you see almost the same results in the control group.

Scott Santens is a strong advocate for UBI. He attempts to make the bull case.

There’s talk about how ‘cash can be anything’ and citing all the things people spent their cash on. Well, yes, and I do buy the ‘better cash than in kind benefits’ argument.

I appreciated the detailed analysis and willingness to note details that were unsupportive, and not trying to disguise his advocacy

But I found most of it to be a (honest) stretch.

Expectations

Another cool note is that the authors collected ex ante forecasts of what the researchers would find.

This confirms the study results were disappointing.

Work

The decrease in work is no surprise – the issue is what we got in exchange.

If you say ‘people working is bad and we should want them to do less of that’ then I am going to disagree. Yes, if the amount of production could be held constant while people worked less, that would be great, we should do that. But we should presume that the jobs in question are productive.

You want to spend less, impact quantity of working hours less, and get more other benefits. Instead, we found more impact on work (although not a catastrophic amount), and little other benefits. And that’s terrible.

Additional Reactions

John Arnold: Consensus among academics is that results of the OpenResearch UBI study were between mixed and disappointing. Yet most articles in the popular press (Forbes, Bloomberg, Vox, NPR, Quartz) characterize the results in a positive tone and ignore or bury the null/negative results. Other outlets that have written extensively about UBI (NYT, WaPo) have ignored the story. Would they have covered it had the results been more positive?

The NPR write-up focused on Sam Altman’s funding of the study, and on heartwarming anecdotes about individuals. It was indeed quite bad, attempting to put a positive spin on things

Again, that does not mean UBI is a bad idea. I don’t think this study showed that.

Ramez Naam: Yesterday, results from OpenAI’s basic income study came out, with disappointing results. $1,000 / month for 3 months had little impact on people’s lives.

• My response to this: Let’s focus on lowering the cost of living for people, particularly at the bottom of the income.
• How? By taking three of the most expensive things in America [Housing, Healthcare, and Eduction] and removing restrictions on supply, while introducing price competition.
• Ease housing permitting. Force price clarity & consistency in medicine. Embrace competition in education.
• cf Education Healthcare And Leisure

UBI as Complement or Substitute

Is this an additional redistributive transfer, or are we having our transfers take a different form?

Among others, Arnold Kling notes that our current redistributive system effectively imposes very high marginal tax rates on the poor, as earning more causes them to lose their benefits

I continue to see a strong argument for doing less of our existing mess of conditional transfers, often confusing in-kind transfers that lose a lot of the value, and instead spending that money on UBI

What this study does is look at the effects of more UBI spending on its own, which is different. I do think this made me more likely to support transfers that target families with young children, including the child tax credit, as money better spent than giving UBI to all adults.

On UBI in General

Post from Eliezer Yudkowsky explaining why you are only as rich as your access to the least available vital resource. Having lots of Nice Things and needs met does not matter if one is missing, here air to breathe

The Future May Be Different

• ...people who today are middle class but in the future have been made unemployable by AI and robotics.
• If this type of future does indeed come to pass, where large groups of people become ZMP (zero marginal product) workers without jobs, then everything is different.
• A study like this tells us little about that world.

I strongly believe we should continue to study schemes to Give People Money in various ways, especially over long periods of time.