Universal Basic Income “Arithmetic”

Benjamin M. Leff

Last week, Eduardo Porter wrote a column pointing out that there is some interest currently – both internationally and in the United States – in a “universal basic income” (or “UBI”).  Under a UBI, the government provides each citizen with an annual cash payout of a certain amount.  The idea appeals to thinkers on both the left and the right, for slightly different reasons.  Porter argues that it’s a bad idea for a number of reasons, but he argues that “the first hurdle is arithmetic.”  He then goes on to argue that the cost to provide a universal basic income of $10,000 each for 300 million American citizens would be $3 trillion (pretty simple math so far), and that is “nearly all the tax revenue collected by the federal government.”  So, obviously, a nonstarter.

Daniel Hemel, a brand new assistant professor over at University of Chicago blogging at Whatever Source Derived, does a little “back-of-the envelope calculation,” in which he points out how silly Porter’s arithmetic is.  It’s ridiculous to think of instituting a universal basic income without simultaneously changing the tax code.  If the tax code stayed exactly the same, then a universal basic income would either be a tremendously expensive social program or a tremendous tax cut for everyone, depending on how you want to look at it.  But Hemel points out that we wouldn’t keep the tax code exactly the same if we instituted a universal basic income.  Instead, we could probably cut the personal exemption and the standard deduction.  Who needs a zero tax rate if the first $10,000 you earn is a gift straight from the government?  You also don’t need the earned income tax credit or child tax credit, since the universal basic income is basically a refundable credit available to everyone.  Then, Hemel suggests cutting a bunch of other deductions to pay for the UBI, like the deduction for state and local taxes, the mortgage interest deduction, and some others, and he produces $1.119 trillion dollars of savings, which would fund a UBI of $3,450 per person.  Not the $10,000 per person that would completely eliminate poverty for any family with children and almost completely eliminate poverty overall, but not a bad start.

But Hemel hasn’t gone far enough either, because he hasn’t considered a complete overhaul of the income tax.  When commentators like Porter complain that a UBI is too costly because it is not targeted at the poor it seeks to help, they ignore the fact that a UBI doesn’t have to represent any cost at all for middle and high earners – the overall cost depends on the structure of the rest of the income tax.  So, if one were willing to reform the income tax along with instituting a UBI, one could easily target the benefits of the UBI to the poor simply by taxing away the grant for higher-income earners.  The reformed tax system could be designed to replicate the tax burden on middle income earners, so there is no “cost” to the system for their UBI grant, and no “cost” to them for anyone else’s.  If the UBI is designed to be more generous to low earners than the current system, then a small overall tax increase could be borne by high earners to pay for the increased benefit for low earners, but that increase could be achieved in any number of ways, all of which would be simpler than the current system.

For Eduardo Porter’s benefit, here’s some arithmetic to illustrate what I mean.  According to the Tax Policy Center, taxpayers in the fourth quintile of income in America (three fifths of American’s make less than them, one fifth makes more than them), pay an average effective income tax rate of 6.2%.  Imagine a taxpayer, Above-Average Joe, who makes $100,000 per year (pretty much right in the middle of the fourth quintile).  Now imagine a very simple income tax with a universal basic income of $10,000 per person.  In this very simple income tax, there are no deductions or exemption, and only two rates: 15% on all income up to $150,000, and 40% on income that exceeds that amount.  Above-Average Joe would get his $10,000 and pay 15% of all his income (including the UBI grant), and so he’d pay $16,500 in income tax.  The first $10,000 he pays would just return his UBI grant to the government, so to speak, but the next $6,500 he pays is out of his other earnings.  His effective income tax rate then is 6.5%, just a little more than the average effective income tax rate for his quintile under the current tax.  In other words, no one has to pay for Joe’s $10,000 UBI grant; he’s paying for it himself.  In the end, he wouldn’t get anything from the government and would pay about the same overall income taxes as he currently pays.

Now imagine a second person, Doing-Great Jane, who earns $200,000, which puts her right in the middle of the top quintile.  Under the simple income tax described above, she’s still going to get her UBI grant of $10,000, but she’s going to pay it right back to the government as part of her $44,000 tax bill.  Once you subtract out her UBI grant, she’s paying $34,000 out of her other earnings, which gives her a tax rate of 17%.  Again, according to the Tax Policy Center, the average effective income tax rate for people in the top quintile of income is 15.8%, so the overall effect of the new system is a relatively small tax increase for her.  That small tax increase for her (and for people richer than her) would fund the significant tax decrease for people at the bottom of the income spectrum.  In that way, the UBI is “targeted” to the poor, near poor, and struggling middle class.

A few caveats about the “arithmetic” above.  First, this system of a $10,000 UBI grant with a two-bracket tax is just an example.  It might raise less money for the federal government than our current income tax, or it might raise more.  If it raised less, and we wanted for it to be “revenue neutral,” then we could choose to make the UBI grant smaller, or to make one of the rates higher.  The point is to illustrate that it makes most sense to talk about the “cost” of a UBI together with a discussion of the structure of the tax system.  It is simply not true that the fact that a UBI is “universal” means that it is not “targeted” at the poor or near poor.

Second, the “effective tax rate” I used for comparison to current rates is the rate for the income tax alone, and excludes federal payroll taxes, which fund social security and medicare, among other things.  For most taxpayers, payroll taxes are a bigger burden than income taxes.  Given that fifty-nine million Americans received social security benefits in 2015, it would be silly not to take social security and medicare into account in some way when designing a UBI system.  Charles Murray, of the American Enterprise Institute, has been explaining his plan, which incorporates social security, medicare, and almost everything else you can think of, over at the Wall Street Journal.  But dealing with the interaction with social security is too complicated to do here, as is any discussion of the treatment of children under a UBI, which is another can of worms.

I see discussions of UBI increasing in the near future, and I could imagine such discussions being tied to the perpetual but elusive goal of a dramatically simpler, but still progressive, tax code.  A simple progressive tax code is a goal shared by the vast majority of Americans, whether they vote for Republicans or Democrats, and could be a unifying issue for us all.