The Pragmatic Case for a Universal Basic Income


ed dolan, an economist specializing in Eastern Europe, is the author of the textbook Introduction to Economics, now in its fifth edition.

July 28, 2014


Are you frustrated by the interminable quest to end poverty in the face of ideological division and widespread cynicism? Why not just cut to the chase, sending everybody in the country a monthly check that covers the rudimentary needs of even the poorest among us?

I know, I know: this sounds like fodder for high school debating societies, or perhaps a bumper sticker left over from the Occupy movement. And it is certainly not on Congress' immediate agenda — or for that matter, President Obama's. But slow recovery from the Great Recession, a stubbornly high poverty rate, stagnant wages and the threat of chronic unemployment fed by technological change and global competition have combined to spark new interest in an old, seemingly radical, idea. Besides, it isn't so radical. At one time or another, progressives, libertarians and conservatives have all supported some variant.

The concept goes by many names: unconditional basic income, basic income guarantee, demo-grant. I prefer "universal basic income," or UBI for short. Whatever you call it, though, the feature that distinguishes a UBI from other sorts of social safety nets is its universality. Unlike other income-support programs, it is not means-tested. Instead, a UBI would provide subsistence-level grants to everyone, regardless of need, earned income, age or job status.

A Scorecard

Step back for a moment. The ideal income-support program would:

  • Leave no one below the agreed-upon poverty level;
  • Focus support on those truly in need;
  • Sustain (or increase) incentives to work;
  • Be inexpensive to administer.

Unfortunately, there are trade-offs that make it impossible to achieve all four objectives simultaneously. A comparison of the alternatives suggests why.

The simplest form of income support is a "top-up" grant – that is, a policy that offers sufficient aid to bring families out of poverty, but no more. Just to have a number to work with, let's set that figure at $20,000 a year for a family of three, a number close to the current official poverty benchmark.

The top-up approach would, of course, satisfy the first requirement. And it would be well targeted, since households already above the poverty line would not be eligible. However, it has one huge drawback: It would leave many households with almost no incentive to work and earn.

For families that are poor to start with, the disincentive is obvious. If a family with no earned income gets $20,000 in government benefits, and one with $10,000 in earnings gets only a $10,000 benefit, why work at a low-wage job?

Less obviously, the disincentive extends to families who could earn enough to lift themselves just above the poverty line. Suppose a household with no income gets $20,000 in annual benefits. One member has a chance to get a job paying $25,000 per year. Is it worth it? Probably not. If $20,000 in benefits will be lost in the process, the net return to work falls to a measly $5,000. Child care, transportation and clothes for work could eat up most of that.

Moreover, despite the simplicity of its objective, a program that just topped-up income would not be easy to administer. The government would have to constantly nose around in families' finances and adjust benefits accordingly. That would be all the more difficult because the threat of losing benefits would provide a strong incentive to hide earnings.

The Taper Trap

One way to have it both ways would to be taper payments gradually as income increases, rather than cutting benefits by a dollar for each dollar of earnings. A tapered benefit was one of the central features of the negative income tax that Milton Friedman proposed in the 1960s. His idea: give people with no earned income a federal tax credit, paid in cash, that would be sufficient to live on. After that, families would be subject to a benefit-reduction rate of 50 percent. That is, as income rose, benefits would fall by 50 cents for each extra dollar earned. Once the benefit fell to zero, which would happen at double the poverty line, families would start paying regular income and payroll taxes only on additional earnings.

Friedman's negative income tax was not adopted in its original form, but many current income-support programs include some kind of taper to encourage work. For example, the federal earned income tax credit (EITC) applies a benefit-reduction rate of 34 percent to families near and just above the poverty line. Complicating matters a bit, the benefit-reduction rate changes with earnings: for families with incomes near the middle of the poverty range, the rate is zero. And the very poorest families with children receive a cash credit of more than one dollar per dollar earned, making their benefit-reduction rate negative!

Note, however, that tapered income support is more expensive than a simple top-up: Except for those who have no earned income at all, poor families are eligible for more than the minimum they need to cross the poverty line. Also, benefits continue for households with earned incomes modestly in excess of the poverty line.

To improve targeting and hold down costs, some safety-net programs impose cutoffs beyond which households are no longer eligible for benefits. Suppose we started with a program that gave $20,000 to families with no earned income, and reduced benefits by 25 cents for each dollar of earned income. That would provide a reasonable work incentive (an effective tax rate on earnings of 25 percent), but families with incomes all the way up to $80,000 would receive some benefit.

To reduce costs, the program could cut off benefits altogether when earned income, say, reached double the poverty line – that is, $40,000. But there's no free lunch here; far from it.

While this plan would maintain adequate work incentives for households with low earned incomes, it would introduce a new kind of disincentive – a cliff – at the cutoff point. A household with earned income of $39,999 would receive $10,000 in benefits, giving it disposable income of $49,999. However, one more dollar in earnings would make the household ineligible for any benefit. So, in this extreme case, the hapless earner would effectively pay $10,000 in taxes on the last dollar earned!

Benefit cliffs can create especially strong disincentives for second workers in a household. Imagine a family where one spouse earns $30,000. In our example, the family's income, including a $15,000 income-support benefit, would be $45,000. If the other spouse now gets a full-time job that brings in an extra $20,000, household income would rise only to $50,000 since all benefits would be lost. Child care, transportation, and other work-related expenses would erode the net $5,000 gain, leaving little to show for the extra work.

Work incentives under existing income-support programs are often even weaker than those in our hypotheticals. One reason is that many incorporate benefit cliffs. Medicaid, the Children's Health Insurance Program and the health insurance subsidies provided to low earners through Obamacare are all guilty as charged.

Note, too, that benefit-reduction rates of various programs are additive. For example, if a household faces a 20 percent benefit-reduction rate for food stamps and a 30 percent benefit-reduction rate for subsidized rental housing, it loses a total of 50 cents in benefits for every extra dollar earned.

Finally, we have to take account of the incentive effects of taxes. Although people with very low incomes pay no federal income tax, earned income is subject to a 7.65 percent payroll tax for Social Security and Medicare. Moreover, after a household reaches the threshold for federal income taxes, an additional 10 percent marginal rate applies.

For purposes of work incentives, what matters is a household's effective marginal tax rate, which is the sum of its marginal tax rates for income and payroll taxes plus the benefit-reduction rates on any income-support programs it qualifies for. For example, if a household faces a 20 percent benefit-reduction rate on food stamps, a 30 percent benefit-reduction rate on rent subsidies, a 7.65 percent payroll tax rate and a 10 percent income tax rate, its effective marginal tax rate would be 20+30+7.65+10, which equals 67.65 percent. That means its disposable income would rise by just $32.35 for each $100 of additional earned income.

In practice, effective marginal tax rates are moderate for families with earned incomes below the poverty level. Effective rates are much higher, typically over 50 percent, for families whose earned incomes fall in the range between the poverty level to twice the poverty level. And effective rates on second earners are typically even higher.

The Ubi Alternative

By now, some of the attraction of a universal basic income should be apparent. A UBI would be effective in raising household incomes at least to the poverty line as long as the benefit were set at that level or higher. Moreover, a UBI would not affect anyone's work incentives, since there would be benefit reductions as earned income climbed.

Then, too, a UBI would be administratively efficient and unobtrusive. It would require no verification of any personal trait or behavior other than the existence of the beneficiary. If the UBI were integrated with the existing federal income tax system, only households with no income at all would receive the full UBI benefit in cash. Those with low-to-moderate incomes would receive part of the benefit as a credit toward income and payroll taxes, and the rest in cash. Those with high incomes would get a tax credit equal to the UBI benefit, which would reduce the taxes they would otherwise owe.

Unlike other existing support programs, however, a UBI would go to everyone: households that didn't "deserve" help (however you choose to define that) would still get it. Moreover, by definition, universality would make the UBI more expensive than targeted programs.

A UBI would be effective in raising household incomes at least to the poverty level as long as the benefit were set at that level or higher. Moreover, a UBI would not affect anyone's work incentives, since there would be benefit reductions as earned income climbed.
Thinking About Affordability

Providing grants to everyone that would be sufficient to raise a typical family of three above the official U.S. poverty threshold would cost about 10 percent of GDP. And that would truly be a budget-buster if it were simply layered on top of today's system of taxes and transfer payments. But this is not what supporters have in mind.

Indeed, a UBI is only conceivable as part of a thorough overhaul of both taxes and government transfers. For the poor, the UBI would replace other safety-net programs. For middle- and upper-income groups, it would replace existing tax deductions, credits and preferences. Consider some back-of-the-envelope calculations of how large a basic grant we could afford without raising marginal tax rates for middle-class and wealthy taxpayers.

To simplify the matter, I'll separate the issue of health insurance from that of income support. This means we will neither expect people to use their UBI benefits to cover health care expenses nor look to cutbacks of existing health care subsidies as a source for financing the UBI. I'll also separate the issue of education support from that of income support. One way to do that would be to pay part of children's UBI benefit in cash to the parents, while putting the rest in trust. The trust could pay educational expenses of minor children, taking due account of parents' preferences. As children grew older, they could receive a gradually increasing allowance to spend at their own discretion. Trustees would release any remaining funds to the beneficiaries when they reached adulthood. That would give young adults a nest egg that they could use, if they chose, to pay for college.

If you're already a skeptic, you've no doubt noticed that, if transformed into top-up grants, the $1 trillion the government currently spends on safety-net programs would be more than enough to raise everyone above the official poverty line. But, of course, the top-up approach would create wretched disincentives to work and be a nightmare to administer.

Another pitfall: about a quarter of all means-tested spending comes from state and local government budgets. Conceivably, some combination of financial sweeteners and political bipartisanship could persuade all the states to make contributions to the UBI. But many Republican states' scorched-earth approach to Washington's largesse, which rejected heavy federal subsidies to expand Medicaid, suggest that this would be a very hard sell in places that already regard the safety net as socialism. Better, then, to count only on the federal portion of welfare spending as a source of funds.

Note, too, that, even before Obamacare, about a third of federal means-tested welfare spending went to Medicaid (which includes nursing homes for the elderly) and the Children's Health Insurance Program. That leaves only $500 billion to finance a UBI.

Divide $500 billion by a population exceeding 315 million – the "universal" in UBI means universal – and we end up with something like $1,600 per person annually. So, to fund a UBI that put everybody above the poverty line, we'd need to dig a lot deeper.

Dare We Speak the E-Word?

Middle-class entitlements delivered through the tax system, including the mortgage-interest deduction (worth $174 billion in 2013), the deferral of income taxes on retirement accounts ($145 billion) and the deduction of charitable contribution ($49 billion), are another potential source of funding for a UBI. All told, the termination of these tax preferences (even leaving intact the deductibility of employer-paid health care insurance) would add some $580 billion to the kitty.

Political suicide, you say? Perhaps. But not if voters truly understood where their interests lay. The math suggests that most middle-income households would gain as much from the UBI as they lost from the end of the big tax-based entitlements. It would only change the way the benefits were distributed.

All told, the $580 billion in additional revenue could add about $1,800 to the per capita UBI grant, bringing the running total to about $3,400 per person. To put that in context, consider the case of a couple in the 25 percent tax bracket (taxable income of $73,800 to $148,850 filing a joint return in 2014). Such a couple would be better off with the UBI and without the middle-class tax preferences unless they amassed more than $27,000 in itemized deductions in the eliminated categories.

Onward. We could also jettison the personal exemption, which reduced taxable income by $3,900 per person in 2013. Stripping it from middle income tax returns would (very conservatively) yield $585 billion in revenues, or enough to add another $1,800-plus to the UBI grant. Now the running total is about $5,200 per capita – still not quite enough to bring a family of three with no earned income above the poverty line.

Where Might Social Security Fit in?

Social Security retirement benefits are higher for individuals who pay in more. But the system is nonetheless redistributive because benefits received by lower-income retirees are relatively large in proportion to lifetime contributions. It would thus be natural to rethink the redistributive function of Social Security in conjunction with the introduction of a UBI.

One approach would start with a phase-in period, during which people who were already retired could opt either for the UBI or for their Social Security benefits, but not both. In the long run, as new people entered the system, the whole schedule of Social Security taxes and benefits could be adjusted accordingly.

About 35 million people currently receive Social Security retirement benefits, with 90 percent of them receiving more than $7,200 per year. That suggests fewer than 10 percent of retirees would opt for the UBI during the transition phase. If so, we would only need to spread the resources available for financing the UBI among just 286 million people, not the entire population of 316 million. Making the appropriate adjustments would bring the amount available for UBI grants to more than $5,800 per person.

Who Gains, Who Loses?

Count the virtues:

  • Except for small families with little or no earned income, the UBI would be enough to bring households above the poverty line (in 2014, from $19,790 for three to $40,090 for eight). Health care programs for low-income families would not change under our version of the UBI.
  • Replacing today's jumble of means-tested welfare programs with a UBI would sharply decrease effective marginal tax rates for poor and near-poor families, thereby enhancing work incentives.
  • Most middle-class households would receive more from the UBI than they would lose in tax benefits. No Social Security retirees would suffer a net loss. Those currently receiving the smallest Social Security benefits would gain by opting for the UBI.
  • Financing the UBI would not require raising anyone's marginal tax rates. However, middle- and upper-income households that currently take large itemized deductions could face an increase in the percentage of income they pay in taxes.
Far from producing a nation of goof-offs, a UBI would draw into the labor market many people who don't consider the net returns to work sufficient to justify the effort.
The Politics of a Ubi

Although many conservatives and progressives support a UBI, others fear that a UBI would produce a nation of layabouts. Once people no longer faced homelessness or starvation, why would they work? But these concerns are surely exaggerated. In the language of economics, a UBI would generate two conflicting incentives. On the one hand, the "substitution effect" would lead people – especially those in low-income households – to work more because the effective marginal tax rate would fall: substituting one hour of work for one hour of leisure would bring in more money than before. The "income effect," by contrast, would incline people to work less when their total incomes increase – as it would for the poor and near-poor.

Which effect would dominate? One way to test the strength of the income effect is to look at how people react to windfalls like inheritances and lottery winnings. These do not change the amount a person takes home from an added hour of work, so they have an income effect but no substitution effect. Studies of windfalls suggest that income effects do exist, but they are weak.

If a UBI were added on top of existing income-support programs, it would have only an income effect. In that case, it would probably reduce work effort, although not by much. However, a realistic UBI that replaced current safety-net benefits would create a strong substitution effect because poor and near-poor families would be able to keep a much larger share of their marginal earnings.

At the same time, the income effect of such a program would be small (or nil). That's because peoples' incomes would not increase by the full amount of the UBI, but only by the difference between the UBI and the value of the food stamps, housing vouchers or whatever that they get now.

On balance, then, a UBI that replaced means-tested welfare programs would be likely to have a large positive effect on work effort for those now living at or near the poverty line. Far from producing a nation of goof-offs, it would draw into the labor market many people who don't currently consider the net returns to work sufficient to justify the effort.

Libertarians and the Ubi

So far, we have focused mainly on conservative and progressive support for a UBI, but some libertarians also favor the concept. That might seem surprising; after all, it is almost an axiom of libertarianism that for Peter freely to give money to Paul is charity, but for John to point a gun at Peter and force him to give to Paul represents theft of Peter's property. Libertarians who support a UBI must thus begin by rebutting the presumption against forced giving.

The first libertarian defense of a UBI is purely pragmatic: while some ideal society might get along without any coercive redistribution of income, a UBI would at least be better than what we have now. As Matthew Feeney, the assistant editor of the libertarian webzine Reason 24/7, put it, "Rather than make the principled argument against the redistribution of wealth, libertarians would do better if they were to argue for a welfare system that promotes personal responsibility, reduces the humiliations associated with the current system and reduces administrative waste in government."

Libertarians can also defend a UBI as a counterweight to the injustices of private property. In principle, of course, libertarians are ardent defenders of private property. But some acknowledge serious flaws in the way the system now works. One such flaw is that a lot of wealth is acquired through government subsidies or by putting up legal barriers to competition – behavior that economists lump together under a term they coined for it: rent-seeking. That behavior is so pervasive that it is impossible to untangle the particular victims and beneficiaries of each law. But a moderate tax on all forms of income and property, returned to everyone as a UBI, would almost certainly lean against the rent-seeking wind.

Less-doctrinaire libertarians, many of whom prefer to call themselves classical liberals, suggest a third reason for supporting a UBI. Given the choice, they argue, freedom-lovers would prefer to live in a system where the minimal state, in addition to protecting people and their property from criminals and foreign enemies, provided everyone with a minimum income to fall back on in case of misfortune. We can find support for a UBI in one form or another in the writings of Adam Smith and John Stuart Mill, and of more recent libertarian thinkers, including Friedrich Hayek and Milton Friedman.

The Improbable Dream?

Hardly anyone sees a UBI as a perfect safety net. It offends conservatives by offering something for nothing. And it raises serious questions for progressives who worry there is more to poverty than a lack of income – that a UBI would not do enough to transform the culture of poverty that weighs down the underclass. But it has pragmatic advocates (including me) who believe that a UBI offers a better compromise than do other income-support programs among the mutually incompatible criteria of effectiveness in reducing poverty, maintenance of work incentives, administrative efficiency and accurate targeting.

A big worry, of course, is that a UBI would end up as budget-buster or require a raid on private wealth to finance it. However, as shown, it need be nothing of the sort – provided it were part of a bargain in which other antipoverty efforts (save medical care) were abandoned, and middle-income earners traded in a hodgepodge of tax breaks for the universal basic income grant.

The most encouraging sign is that the liveliest debates over a UBI today are taking place within, rather than between, the main ideological camps. At a time when macroeconomic forces and the politics of big money are leading to ever-greater inequality, perhaps America is still capable of finding common ground for a pragmatic antipoverty effort.

main topic: Income Distribution
related topics: Job Creation, Tax Policy, Technology