Summary: Basic income advocates often encounter the objection, “We can’t afford it!” To counter that objection convincingly, they need to address several key questions:
- What resources can a basic income draw upon? Which existing income support programs would be rendered unnecessary if an adequate basic income were in place?
- How would a basic income mesh with other social programs, especially health care, child care, and social security?
- Who would be eligible for a UBI? The entire population? Citizens only? Adults only?
- Does an adequate basic income have to be generous enough to raise everyone out of poverty by itself, or would something less than that be enough?
In search of a baseline UBI
To many people, a universal basic income (UBI) sounds like a good idea — until you start thinking about whether we could afford it.
Robert Greenstein, President of the Center for Budget and Policy Priorities, is no enemy of robust government programs to help the poor, yet he is an outspoken critic of a UBI. Affordability is Greenstein’s number one concern. He argues that giving a UBI of $10,000 a year to the entire U.S. population of 327 million people would cost $3.27 trillion, about equal to the entire annual revenue of the U.S. government. It’s hard to imagine that such a UBI would advance very far, he says.
But is asking whether we could afford a UBI of any given size really the right question? A more sensible starting point might be to ask how much basic income we could afford if we used only what the government is already spending on income support, without raising taxes or total spending at all. The answer to that question gives us a baseline UBI that that we can use as a point of reference.
Determining how generous that baseline UBI would be involves two steps: First determining how much money could be made available by redirecting existing income support spending to a basic income (the numerator), and second, determining how many people would be eligible to receive benefits (the denominator). The baseline UBI is then equal to total funds divided by the eligible population.
What funds are available?
Robert S. Pfeiffer, creator of the website FederalSafetyNet.com, provides a helpful list of $754 billion in federal programs that have the primary purpose of aiding low-income Americans. By selecting appropriate items from that list, excluding others, and adding still others that do not appear on the list at all, we can make a plausible estimate of the numerator for the baseline UBI calculation.
Core welfare programs. The most obvious components of the numerator are a core group of means-tested federal programs that constitute what most people think of when they hear the term “welfare.” These include EITC, SNAP, TANF, housing assistance, and a handful of smaller programs. By Pfeiffer’s calculations, the federal share of these came to $225 billion for 2018. An adequate UBI would make all of these redundant.
Healthcare. The next question is what to do with the $400 billion now being spent on Medicaid — the largest of all current programs to aid the poor. In principle, we could try offering a UBI that was generous enough to allow even people living on the basic grant alone to pay for their own health care at market rates. Then we could throw the $400 billion into the UBI numerator. However, there are two serious problems with that approach.
One is that a basic income large enough to allow everyone to pay for their own medical care would have to be very large. Many discussions see a UBI of $10,000 to $12,000 thousand a year as enough to largely eliminate poverty, but U.S. healthcare spending alone comes to about $11,000 a year per person. A grant large enough to cover everyone’s unsubsidized health insurance and out-of-pocket costs of care, and still maintain a standard of living at the poverty level or above, would require a basic grant of more than $20,000. Dividing the $400 billion federal contribution to Medicaid among the entire U.S. population would yield only a little over $1,200 per person. Obviously, that would not be enough.
The second problem stems from the highly unequal distribution of healthcare spending. The healthiest half of the U.S. population accounts for just 3 percent of total spending, while the sickest 5 percent accounts for fully half.
Even if we did set the basic income at $20,000 a year, it is far from obvious that people would use the money to buy health insurance at market rates. Instead, people who were young and healthy would be tempted to spend their UBI on other needs. Some of them would be unlucky enough to get sick or would have costly accidents. What would happen to them? To turn them away when they sought care that they could not pay for would be inhumane. But to guarantee even emergency care for the uninsured would only increase the temptation for the young and healthy to forego insurance.
It would make more sense, in my opinion, to address the need for universal health care as a completely separate policy initiative. Elsewhere, I have described one way of doing so, universal catastrophic coverage. Money now spent on Medicaid should be put toward funding that or some other version of universal health care rather than used to increase the baseline UBI.
Middle-class welfare. Critics like Greenstein correctly point out that taking what is now spent on programs for the poor and redistributing it equally among the whole population would, other things being equal, increase overall inequality. To avoid such an outcome, we need to bring in another major category of federal spending, one that we could call middle-class welfare. It would be preposterous to replace current income support for the poor with a UBI while retaining income support for those who are already better off, and then giving them their basic income, too.
Most benefits for people with middle and upper incomes are delivered through the tax code. According to the Tax Policy Center, the three largest such items are housing subsidies, in the form of the mortgage interest deduction and the exemption of capital gains on the sale of homes ($70 billion in 2018); the charitable deduction ($54 billion); and tax benefits for retirement savings ($251 billion). The total from just these three programs would add $375 billion to the pool of funds available for a UBI.
But, it might be objected, wouldn’t working families struggle financially without these deductions? Not really. The reason is simple — tax deductions provide big benefits only to people who are rich enough to pay big taxes.
Consider the mortgage interest deduction. According to a study for the Tax Policy Center by Chenxi Lu and Eric Toder, in 2016 the mortgage deduction was worth just $193 a year, on average, to a family earning $40,000 to $125,000. Even for those with incomes in the $125,000 to $200,000 range, the mortgage deduction was worth just $1,007 a year. A very modest basic income would outweigh that. Only the very wealthy would lose more in deductions than they gained through a UBI.
But what about the charitable deduction? Isn’t that necessary to support essential nongovernmental programs for the poor? Again, not really. According to my analysis of the data, less than a third of all charitable giving actually reaches the poor. Much of the rest goes into covering salaries and overhead costs of religious institutions or into art museums, opera houses, endowments for prestigious universities, and other “charities” that are gifts from the wealthy to the wealthy. What is more, studies show that the deduction has a remarkably small effect on total giving.
As for retirement, I suggest that anyone who wants could have all or a part of their UBI go automatically into government savings bonds or a private mutual fund of their choice. For most people, that would do far more for retirement than the current tax benefits on 401k contributions.
State and local spending. Pfeiffer calculates that state and local governments spend $58 billion on income support for the poor, above and beyond federal programs. That figure does not include Medicaid or state and local tax benefits for the middle class. How should these funds be treated for purposes of a UBI?
One approach would be to require states to throw the money that they now spend for affordable housing, school nutrition, homeless shelters, and the like into the UBI pool. Doing so would increase the basic income for all recipients. However, that approach would involve significant redistribution of funds among the states, and would probably encounter serious political and even constitutional impediments.
Another approach would be to let states do whatever they want, either setting up their own mini-UBIs on the state level, or continuing their existing programs for the needy. In either case, state programs would somewhat reduce the minimum federal UBI grant needed to raise people out of poverty. It would also allow states to seek creative solutions to problems like substance abuse, mental illness, and domestic violence for which cash payments alone might not provide adequate relief.
For present purposes, I will follow the second approach and focus on a UBI funded solely from federal sources.
Who is eligible?
Having addressed the numerator of our baseline UBI, available funding , we turn now to the denominator — the number of people eligible to receive a grant. At first glance, that might seem obvious. If the “U” in UBI stands for “universal,” then everyone gets it. But who is “everyone?” Here are three of the most important eligibility issues:
Citizenship status. According to the Census Bureau, the U.S. population in 2018 was 327 million. Of those, about 44 million were foreign born. Data compiled by the Center for American Progress suggest that roughly half of the foreign born are naturalized U.S. citizens. Of the remainder about half are legal permanent residents and half are undocumented. The 327 million population includes all people residing in the United States, regardless of their legal status, but it does not include U.S. citizens residing abroad. Those are estimated to number about 9 million.
Who, then, is “everyone” for purposes of a UBI? An expansive definition would include all U.S. residents, regardless of legal status, plus U.S. citizens living abroad — about 336 million as of 2018. A narrow definition would include only resident U.S. citizens, of whom there are about 305 million.
The issue of who should be eligible has been widely debated by both critics and advocates of a basic income. Rather than re-litigate that issue here, I will simply calculate both high and low estimates for the baseline UBI, based on the Census population plus expats (336 million) and resident citizens only (305 million).
Children. According to the Census Bureau, 13.4 percent of the U.S. population, or about 73 million people, were under the age of 18 as of 2018. Some UBI advocates would like to see all of them receive a full UBI. Others think the UBI should be paid only to adults, or that children should receive only a fractional payment. Again, I will not try to resolve this contentious issue. Instead, I will offer alternative estimates for a baseline UBI, using two different assumptions regarding children.
In the first version, only adults are eligible for the UBI. In that case, I will assume that the federal government continues to spend the same amount on child support as it does now. The 2018 total for such purposes was $164 billion, including $42 billion in federal child support programs listed by Pfeiffer (child nutrition, Head Start, and WIC) plus $122 billion in child-related tax credits from the list provided by the Tax Policy Center.
It would be possible to continue those programs without change, but in my view, it might make more sense to spend all or part of the funds for a universal child credit of the kind proposed by my Niskanen Center colleague Sam Hammond. His proposal takes the form of a refundable tax credit that would be available to poor and middle-class families with children, but phased out for those with higher-incomes. If the entire $164 billion currently budgeted for child support programs were used in support of children in the bottom half of the income distribution, the credit could be quite large, as much as $8,000 per child per year.
The second variant for the treatment of children assumes that full eligibility begins at birth. The actual grant would be paid to the children’s parents or guardians until they reach the age of 18, perhaps with some restrictions to safeguard the children’s interests. In that case, the children’s own basic income grant would replace existing child support programs, so the $164 billion now spent on them would go into the pool of funds available for the UBI.
Social Security. A third major eligibility issue is how to integrate a basic income with Social Security. According to the Social Security Administration, some 67 million Americans received retirement or disability benefits as of 2017. Total benefits came to just under a trillion dollars — an average of about $14,000 per year per beneficiary.
It would be tempting to say that a basic income would make Social Security unnecessary. In that case, we could throw that trillion dollars into the basic income funding pool. However, although that would simplify matters, it raises some practical and political difficulties.
One is that most Social Security beneficiaries get payments that are higher than any plausible UBI. Another is that Social Security benefits are tied to payments into the system made over a lifetime, so that recipients have a sense of ownership over them. Together, these would create considerable resistance to ending Social Security.
Rather than folding Social Security directly into the UBI, then, I suggest the following rule: Everyone eligible for the UBI should have the right to take the baseline benefit, or to take his or her Social Security benefit, whichever is larger, but not both.
It is not easy to make an accurate calculation of how many people would opt for the UBI and how many would instead stay with Social Security. As a rough estimate, I will assume that 90 percent of current Social Security beneficiaries would opt to keep what they get now, and that the average annual Social Security benefit of the 10 percent who opt for the UBI would be $2,000 a year. If so, about 6 million current Social Security recipients would choose the UBI. The Social Security benefits they gave up would add about $12 billion to annual funds available, while the people who opted to stay in Social Security would reduce the number of people eligible for the UBI by about 60 million. The result would be a significant increase in the baseline UBI.
Results: The baseline UBI
We now have all the information we need to calculate some estimated baseline values for a UBI benefit. All variables use 2018 values.
The numerator of our calculations, “available resources,” consists of funds now used for programs that can safely be eliminated once the UBI is in place. The low estimate of available resources includes core welfare programs of $225 billion, middle-class tax benefits of $375 billion, plus savings in Social Security benefits of $12 billion, for a total of $612 billion. The high estimate includes an additional $164 billion now devoted to child benefits, for a total of $776 billion.
The denominator, “eligible recipients,” is the number of people eligible to receive UBI benefits. Social Security beneficiaries are given the option to take either their existing Social Security benefit or their UBI benefit, whichever is greater, but not both. For purposes of illustration, 90 percent of Social Security beneficiaries are assumed to opt to keep their current benefits, thereby forfeiting their UBI eligibility. Estimates are given for four versions of the eligibility:
- Census population plus expats consists of the entire U.S. census population of 327 million people, including resident citizens, legal permanent residents, and undocumented residents, plus an estimated 9 million U.S. citizens living abroad, less 60 million who opt to stay with Social Security, for a total of 276 million.
- Adult census population plus expats equals the above less 45 million children under the age of 18, for a total of 231 million.
- Resident citizens includes 305 million U.S. citizens of all ages residing in the United States less the 60 million social security adjustment, for a total of 244 million.
- Adult resident citizens include the above less their estimated 41 million children, or 204 million in all.
The four eligibility groupings give rise to four estimated values for the baseline UBI benefit. Estimates for the two all-age eligibility groups use the higher estimate of available resources. In those scenarios, existing child benefits are placed in the pool of funds available to support the UBI. Basic income benefits paid to parents in the names of their children replace current child benefits. Estimates for the two adult-only eligibility groups use the lower value for available funds. In that case, either current child benefits could continue, or the $164 billion now spent on them could be mobilized for a new child-only cash benefit program separate from the adult UBI.
Here are the four estimates of the baseline UBI benefit in tabular form:
Is it enough?
These estimates of the baseline UBI grant are likely to seem disappointingly small to many UBI advocates. Even the most generous of the estimates, the all-ages, citizens-only benefit of $3,167, provides only a quarter of the 2018 federal poverty level (FPL) for an individual ($12,140) and only half of the FPL for a family of four ($25,100). But here are some additional points to consider in deciding whether such a UBI would be enough:
- Benefits for families with children would be more adequate if the UBI for adults were combined with an additional credit for children, as discussed above. Adding an $8,000 child credit to the adult grant from Scenario 4 in the table above would produce a total family income at 67 percent of the FPL for a single parent with one child, 116 percent for a single parent with two children, 68 percent for two parents with one child, 88 percent for two parents with two children, and 102 percent of FPL for two parents with three children.
- The UBI benefit would not be the only source of income for most poor families. About a third of all poor adults work at least part time, and more than half of all poor people live in families where at least one member works. Those families do not need a UBI grant at 100 percent of FPL to reach an income at or above the poverty level. Furthermore, as I have argued elsewhere, replacing current income support policies with a UBI would increase incentives to work.  
- If funds currently spend on Medicaid, Medicare, the ACA and other health programs were applied to universal health care for people with low incomes, they would be relieved of the need to spend any part of their UBI grant on health care. That would greatly increase their ability to maintain an adequate standard of living.
- Under the scheme outlined here, $58 billion or more in state and local poverty programs would still be available to fill gaps in the federal UBI.
Keep in mind, too, that the goal of this commentary has not been to calculate the size of a UBI that would be enough by itself to fully eliminate poverty. Instead, the aim has been to estimate a baseline UBI that would draw only resources that the federal government already devotes to income support. The existing policies themselves fall short, leaving nearly 12 percent of the population in poverty as of 2018. All things considered, even the modest baseline UBI shown in the table would very likely leave fewer people in poverty than do current policies.
But understandably, many UBI advocates would prefer something more generous. If so, there are two ways to proceed.
One would be to increase the available funding. Based on the parameters of Scenario 3 from the above table, increasing the basic benefit to $10,000 per year would take some $1.7 trillion a year beyond what is now being spent on federal income support programs. That is a lot less than Greenberg’s estimate of $3.27 trillion, but it is still equal to about half of all current federal revenues.
Greenberg assumes that the additional spending would need to be financed by new taxes. There are plenty of proposals — carbon taxes, wealth taxes, value added taxes or whatever. Some readers, inspired by modern monetary theory, will point out that the government could, if it wanted, spend the money for a UBI without raising taxes or borrowing at all. Still others might prefer to fund a more generous UBI by cutting back on other government spending for programs that they dislike.
Alternatively, the basic income could be modified in a way that made it less generous for people with higher incomes. The simplest way to do that would be to make the basic income itself subject to tax. Benefits could be taxed as ordinary income, or some special formula could be applied, say, a 20 percent benefit reduction rate for households with household incomes above the poverty level. As has been noted more than once, a basic income that is subject to a positive benefit reduction rate is analytically equivalent to a negative income tax. In a recent paper for the NBER, Hilary Hoynes and Jesse Rothstein show that a UBI is just one special case in a spectrum of income support plans that includes negative income taxes, wage subsidies, and others.
“We can’t afford it!” seems to be the most frequent objection to a universal basic income. It is true that any kind of meaningful UBI would be a major fiscal undertaking. Yet, the objection from affordability is not conclusive, and critics’ estimates of the cost of a UBI are often overstated. UBI advocates can use several strategies to rebut the charge of unaffordability.
First, they can show that the money already being spent on income support would go a long way toward providing a meaningful UBI. Even a basic benefit well below the often-cited levels of $10,000 or $12,000 a year could make a bigger dent on poverty rates than does the current alphabet soup of EITC, SNAP, TANF, and the rest. One key is to make sure that a UBI replaces not just income-support programs for the poor, but also the “middle class welfare” lavished on higher-income households through the tax code. Another is to make sure that people are not allowed to “double dip” by receiving both the UBI, and Social Security benefits that are greater than the UBI.
Second, UBI advocates can refine their concept of eligibility. Critics too often just multiply the U.S. census population of 327 million by a basic benefit of $10,000 and consider the case closed when that produces a huge number. Instead, we need a conversation about residency, citizenship, age and other characteristics that define the eligible population.
Third, UBI advocates can pay closer attention to coordinating basic income with programs for health care and child support. A good case can be made for designing separate programs to deal with both areas, rather than folding them into a UBI.
Fourth, UBI advocates need to be clear about whether basic income grants should be subject to taxation or other forms of benefit reduction.
In my view, a calm dialog about details like these is the proper way to confront the dismissive “We can’t afford a UBI!”