A Conversation with O-Lab Directors Hilary Hoynes and Jesse Rothstein
In recent years, the idea of establishing a Universal Basic Income (UBI) has emerged as a central topic of interest in social policy circles and in political debates. Democratic presidential candidate Andrew Yang has made the UBI a centerpiece of his 2020 campaign, and the idea has received support from high-profile figures in the tech world and on both the right and the left, including Charles Murray, Elon Musk, David Graeber, and Chris Hughes. The idea is not new – Richard Nixon proposed a version of the UBI in his first term – but the renewed enthusiasm for such a program raises new questions about what such a program would mean for existing safety net programs, who stands to gain and lose through such a proposal, and how much it would truly cost to set one up.
O-Lab Directors Hilary Hoynes and Jesse Rothstein addressed these questions in the 2019 Annual Review of Economics. Looking at an array of forms the UBI could take, the costs associated with various programs, and its potential effects on the economic security of recipients and on the labor market generally, Hoynes and Rothstein offer notes of caution about whether the benefits of UBI programs would justify the costs involved, particularly considering the amount of money flowing to middle- and upper-income families.
Hoynes and Rothstein also point out that, while attention to the issue has increased dramatically over the last several years, the evidence base has not kept pace. And while there are some examples of income transfers with UBI-like features, there are no large-scale examples of programs that meet the test of (1) true universality (e.g. the benefit is available to all adults regardless of demographics, family status, income, or employment status), (2) basic coverage (providing enough income to cover one’s basic needs), and (3) cash support (paid in the form of cash).
In this O-Lab Q&A, Hoynes and Rothstein expand on their research and offer insight into where the recent enthusiasm for UBIs has come from, what the evidence suggests about using UBI as a new component of the social safety net, what politics might be involved, and whether other programs might accomplish the same goals more efficiently and equitably. Throughout the discussion, examples are provided of recent and ongoing experimentation with the idea.
Let’s begin at the simplest level: what is it we’re talking about when we talk about the universal basic income? At the broadest level, how would you define it?
Hilary: In our paper, we break down the UBI into the U-, the B- and the I-. The I is simple - it’s income, so it’s something that’s always cash, not in-kind support. The B means that a true UBI would cover enough to meet your basic needs. The U is the one Jesse and I found to be less consistent across the programs we looked at. Universality could mean everybody’s eligible regardless of their income - that’s the one we thought to be the most pure version - but universal could also mean, demographically, everybody’s eligible, as opposed to targeting it to, for example, the aged, or single mothers, etc. At the end of the day, the one element that was consistent within all UBIs is that it covers you if you don’t have earnings.
What are some examples of UBI programs that are out there now? Do we have good models for what this can look like?
Jesse: The closest existing analogy is the Alaska Permanent Fund, where Alaska takes the royalties that it gets from leasing public land for oil drilling and distributes those royalties as a flat amount to everyone who lives in Alaska. It’s not “basic” in the sense that we defined it, since it’s nowhere near enough to live on, but it’s a flat amount and it goes to everybody regardless of how much they work, their income, and who they are.
Beyond that, there are a lot of pilot studies; there aren’t a lot of universally implemented programs. There’s a pilot study being done by Y-Combinator that’s picking a small number of families and giving a payment every month.
Hilary: But even though those are small pilots - the one in Stockton and the Y-Combinator one - they’re small, but also, to be eligible you have to have income below some level. And so they’re still targeted. Nobody’s testing, even in a demonstration project, what it means to randomly give people - in the universal sense - income.
Why do you think the idea has been generating such interest over the last couple years, even before Andrew Yang raised it in his presidential campaign?
“There’s a sense that the labor market has faltered for a long time and that we need to figure out new ways to provide support. That’s leading to the desire to do something.
Jesse: I can think of a couple explanations. One is just the sense that the labor market and all of our labor market institutions aren’t working, and that a lot of the things we were worried about before, we’re not as worried about now. Particularly, previously we were really trying to encourage people to get jobs, and now we have a sense that maybe there aren’t enough jobs, so maybe we ought to not be pushing people so hard to go find them. So, I think there’s just a sense that the labor market has faltered for a long time and we need to figure out new ways to provide support - that’s leading to the desire to do something.
Beyond the distinction of cash vs. in-kind support, how would this really differ from the kind of social safety net we have today?
“A simple, but profound difference from the existing system is in thinking about a program that provides more benefits and further up the income distribution.
Hilary: We do have cash programs, so the idea of using cash isn’t new. But the idea of [setting up a program to truly meet basic needs] – probably not so much, because the US has never been that generous. Aside from those two things though, I think the critical difference is that the cash benefit programs that we do have - or even food stamps, which is about as close to cash as there is without being cash - are phased out practically on the first dollar of earnings. By design they’re much more targeted to go to those who have income at the poverty level or below. So a simple, but profound difference, is in thinking about whether we should have a program that provides more benefits further up the income distribution before you phase them out. That seems to be a first order issue that, regardless of how you design the program, comes up in contrast to what we do in our current system.
Are we generally talking about layering this on top of existing safety net programs, or using UBI to replace them?
Hilary: I think this is more of an unknown. There are two different groups of people who like the idea of the UBI - there’s the group of people that says “our current system is a patchwork of programs with all sorts of hassles and high marginal tax rates and we’d be better off to just throw it away and replace it with the UBI.” That has profound distributional consequences if it’s revenue neutral. But then there’s another group that is advancing a more principled argument and is thinking of it more as an add-on to existing programs. And, in the pilot studies, the UBI is completely separated from the rest of the social safety net in terms of how they’re being implemented.
“You can’t replace Medicare with UBI. But all of the money is in Medicare and Medicaid, so if you’re not going to replace those, then you’re not going very far towards paying for the UBI.
Jesse: Right. On the argument that we should just replace the existing safety net with a UBI - there are some programs you could imagine replacing. If you had a UBI, for example, you might not need unemployment insurance. Or maybe food stamps. But there are other programs that UBI just doesn’t come anywhere close to replacing, and those tend to be the expensive programs. You can’t replace Medicare with UBI. But all of the money is in Medicare - and Medicaid - so if you’re not going to replace those, then you’re not going very far towards paying for the UBI.
Hilary: Exactly. So if you take that same amount of money that we’re currently spending on food stamps, TANF, SSI - all these programs, and then you turn it into a universal program, the people at the bottom who are currently getting those programs are going to get a lot less. So that’s the sense in which it could have very profound distributional consequences, if it was revenue-neutral.
With those distributional consequences in mind, what is the best argument for this kind of program? Is it simply the political argument that people will be more supportive of a universal program?
Jesse: There is an argument that people will feel more positively about a universal program if they don’t feel like it’s only going to people who are somehow not worthy, and they’re more inclined to support it if they feel like they’re benefitting from it. I’m skeptical that they’re going to be that much more supportive once they get the tax bill though.
Hilary: Right, and it’s really unknown. We don’t have much evidence on this, so it’s hard to know. But I think a little bit more nuanced layer of this might be: you could have a not-universal program that’s phased out higher up the income distribution than where a lot of our programs are phased out, that would be less expensive and could have some important implications for labor supply and have implications for providing support in a broader sense than we’re doing now.
That sounds a lot like bringing back a welfare program and just making it available further up the income distribution.
Hilary: Yes. But maybe with some re-branding. A lot of people don’t think about EITC as welfare, for example (although it’s criticized for other reasons). And so, for example, could we move away from this very old negative view of welfare in America by bringing in something new?
On that point, I have never heard any discussion of Temporary Assistance for Needy Families (TANF) in current UBI discussions. It does seem likely that, once you get into experimentation and political compromise, that’s approximately where you would land.
Jesse: I think TANF is easily the most stigmatized of our existing programs, and I think one of the key virtues in the minds of UBI’s backers is that it would reduce stigma. And so, if you back UBI, it’s because you believe it won’t turn into TANF.
Hilary: And TANF feels so broken that it is difficult to think of how to fix the existing program rather than to replace it with something else.
Your paper suggests that a UBI would likely disincentivize work for some recipients, but you also discuss the opportunities it creates for people to pursue training, education, or other personal investments. Can you talk about why this is such an important aspect to consider?
Hilary: When you think about Unemployment Insurance, one of the arguments for it is not just to give income protection when you lose your job, but also that it gives you the time to search and find the next job. So any program that provides income support provides that mechanism, and the more generous it is, the more it allows time and the space for a person to get that training or education. So it doesn’t seem to me to be something that’s unique to UBI. The point we make in the paper - which Jesse really articulates - is that if you have a program that isn’t phased out or that’s phased out higher up or less steeply, you’re obviously going to have less of a work disincentive and maybe, if we we’re now worried that all the jobs are going away, we’re not as worried about work disincentives.
Jesse: I think the way I think about it is: the virtue of the UBI is that it provides a true safety net. You don’t have to worry that, if you need it, it won’t be there for you. Because everybody has it. And that hopefully allows people to take some risks - that could be going to school, taking some time out of the labor market, trying out a new business, and that potentially is a positive thing. In terms of the labor supply effects, you have to be concerned about the “compared to what” question; a UBI reduces labor supply modestly compared to no UBI, holding everything else the same. But relative to existing programs - TANF and food stamps - maybe it raises labor supply a little bit, because we think there are some disincentives in the existing programs. On the other hand, EITC increases labor supply. So it’s not quite clear what the net effect is.
When people outside economics and policy hear proposals along the lines of Andrew Yang’s $1,000/month proposal, what are some of the important questions they should be asking in order to assess the merits of the proposal?
Hilary: The first thing is: do a back-of-the-envelope calculation. Just do the math. $1,000 a month times 12 months a year times X million people. What is that going to take? What are the taxes that will take? It’s a lot.
Jesse: There is an argument that knowing this is there will make people happier and make people less stressed out about money and more comfortable and feel valued in society. And those are all potentially positive. So it’s not nothing. But it’s expensive.
Any final takeaways?
“Either you come up with a whole lot of new money or you have to take money away from people who are the neediest in society. But you can’t talk about how great a UBI program is without thinking about which of those things you’re going to do.
Hilary: I just think there’s a lot of discussion about the benefits, which I agree with, with less discussion of what the costs are. And you need to be thinking about both to figure out if it’s a good policy choice.
Jesse: Either you have to come up with a whole lot of new money to pay for it or you have to take money away from people who are the neediest people in our society or some combination of those two. But you can’t talk about how great the program is without thinking about which of those things you’re going to do.
In recent years there has been growing momentum to test variations of basic income programs in the United States and internationally. Below, we describe the general features of three such programs:
The “Making Ends Meet” Experiment
Following a lengthy pilot period, Y-Combinator Research, the non-profit arm of the technology company Y-Combinator, plans to formally launch a study of a UBI program issuing payments of $1,000/month to 1,000 individuals over three years. The program will be rolled out across two states and is targeted at individuals between the ages of 21 and 40 whose total income does not exceed the Area Median Income and who are not receiving SSI, SSDI, or housing vouchers.
The Stockton Economic Empowerment Demonstration (SEED)
Formally launched in 2019, the SEED program is experimenting with providing a $500/month cash grant to 125 city residents living in the neighborhoods with an average income below the city average of $46,033/year. The program has no strings attached - families receive a pre-loaded debit card which they can either use to make purchases directly, withdraw cash, or use to transfer funds into another account. Participating families were randomly selected from an invited pool of participants living in targeted neighborhoods. A control group of 200 participants will not receive the stipend, allowing the city to compare outcomes for both groups following the end of the 18-month intervention period.
Finland Basic Income Experiment
Between 2017 and 2019, the Finnish government experimented with a version of Basic Income targeted at 2,000 unemployed 25-58 year olds. Participants were randomly selected for inclusion, with 5,000 recipients of standard unemployment benefits serving as a control group. The goal of the experiment was to understand how shifting to an unconditional cash payment system (in this case of €560/month) would affect labor market participation and a range of other outcomes. Importantly, unlike the existing Finnish Social Security program, cash support for participants did not end upon receipt of employment and participants had no administrative or reporting requirements.