Universal Basic Income and Marginal Rates

By: Benjamin M. Leff

Last week I posted some preliminary thoughts about a Universal Basic Income (UBI), arguing that it doesn’t make much sense to talk about a UBI outside the context of reforming the tax code, because a UBI without tax reform is just a mind-bogglingly large tax cut.  After the post, I got a tweet from Daniel Hemel about the effect of such reform on marginal rates: “Totally agree w/your UBI analysis, though to keep effective TR constant we would have to raise marginal TR on middle class.”  He’s right that to talk about a UBI in the context of fundamental tax reform without talking about marginal rates is silly.

First, a quick primer on marginal rates for non-tax readers.  Last week, I pointed out that one could institute a UBI and keep “effective” tax rates the same for middle-income taxpayers.  The effective (or “average”) tax rate is just the percentage of income you pay in tax. If you earn $100,000 and pay $6,000 in tax, your effective tax rate is 6%.  Easy.  The “marginal” tax rate, on the other hand, is the amount of tax you pay on the last dollar you earn.  We sometimes call it your tax “bracket.”  As I pointed out last week, a person in the fourth quintile of income, on average, has an effective income tax rate of about 6%.  But, most people in that quintile are likely to be in the 25% tax bracket, and so (with a bunch of caveats described below) have a marginal tax rate of 25%.  Standard economic theory tells us that the incentive to earn money or not depends not on one’s effective tax rate, but on one’s marginal tax rate.  So, people who have a marginal tax rate of 25%, at least theoretically, have their incentive to work reduced a little (obviously, the fact that they get to keep 75% of their marginal dollar continues to provide a substantial incentive to work).  Again, at least theoretically, there is no way to tax income without reducing, at least somewhat, the incentives to work.  There is tremendous empirical uncertainty about how much disincentive is created by various levels of taxation at various levels of income (and I am resisting a thousand parenthetical caveats), but the fundamental tenet of standard economic theory – that higher marginal rates create a greater disincentive to work than lower rates – is pretty well accepted.  And that idea has pretty strong intuitive appeal, at least when rates get high.

So, if one were reforming the tax system to include a UBI, what effect would this reform have on marginal rates?  This is where I get to talk about an article that figuratively blew my mind when I first read it: Daniel Shaviro’s Effective Marginal Tax Rates on Low-Income Households.  (When I told my wife that Shaviro’s article had blown my mind, she said, “Compare it to Carlos Castaneda,” and I said, “More!  It blew my mind more than Castaneda.”)  Shaviro’s article simply explains that the phase-out of means-tested government benefits operates exactly like an increase in marginal tax rate.  So, for example, imagine someone with a housing subsidy that caps their rent at 30% of income.  For every dollar they earn, their subsidy decreases (or rent increases) by 30 cents.  That 30 cents of every dollar functions exactly like a marginal tax, except that to determine the overall marginal tax, all benefit phase-outs and all actual taxes have to be combined.  Shaviro includes the following amazing table, which admittedly is his most extreme example of the high marginal rates that could be created by the interaction of multiple benefit phase-outs operating on a poor family:

Table 1[1]
Estimated marginal tax rates for a one-parent,

two-child household residing in a high-TANF benefit

state and receiving a federal housing subsidy.

Income Range ($)                                     Marginal Rate (%)

0-1,550 …………………………………………………..-6.7

1,550-1,650 …………………………………………….. 21.2

1,650-9,800 ……………………………………………. 52.4

9,800-12,850 ………………………………………….. 89.6

12,850-14,350 ………………………………………… 109.2

At 14,350 …………………………… “Notch” loss of $1,800

14,350-14,700 …………………………………………… 78

At 14,700 …………………………… “Notch” loss of $2,250

14,700-15,050 ………………………………………….. 61.3

15,050-19,550 ………………………………………….. 78.5

At 19,550 …………………………… “Notch” loss of $1,000

19,550-25,000 ………………………………………… 78.5

It starts out pretty well, with a negative marginal rate, meaning that for every dollar the parent earns, the government gives them almost seven cents.  That negative marginal rate is because of the earned-income tax credit, which provides a wage subsidy for low-income workers.  But as soon as the parent earns more than $1,550 for the year, they’re hit with a pretty high marginal rate.  It peaks for income between $12,850 and $14,350, in which range the parent is subject to an effective marginal tax rate of 109 percent.  109 percent.  That means for every dollar that the parent earns, they lose a dollar and nine cents from the government.  Economists can debate whether small increases in marginal rates may or may not have an adverse effect on labor participation, but it is hard to imagine that one would have much incentive to get up in the morning and slog to a low-paying job if the result is that for every dollar you earn, your family loses that whole dollar plus nine additional cents.

If you think marginal rates this high are bad (and how could they not be), then there are two solutions: (i) don’t provide any government benefits to the poor in the first place, in which case you don’t have to take them away as income grows (a solution I don’t support for obvious reasons); or (2) don’t phase them out.  Not phasing out a cash government benefit targeted to the poor is – you guessed it – a universal basic income.  (ta da!)

But, of course, as I discussed last week, to pay for a UBI, you have to tax income, and if you want to match the effective tax rates of middle income taxpayers, you need to increase marginal tax rates to tax back the UBI grant.  Last week I suggested that you could do it with two tax rates, 15% and 40%.  If the 15% rate applied on the first dollar earned, then some taxpayers who currently pay at a zero rate (because of the personal exemption and standard deduction) or a 10% rate (because that’s the lowest bracket) would have a nominally higher marginal rate under this UBI, even if they pay less taxes overall (or even end up with extra money in their pockets).  But, to the extent that some taxpayers are paying very high marginal tax rates under current law because of the phase-out of means-tested benefits, these new marginal tax rates, even if nominally higher than under the current tax code, would still be much lower than current effective marginal rates.  And, most importantly, to the degree that the UBI replaces means-tested government benefits, the marginal rates are lower for low and middle-income workers, standardized, transparent, and steadily progressive.  Those all sound like good things, right?

[1] There are some caveats about this chart: first it assumes that the family receives a housing subsidy, which almost no one does; second, the article is almost twenty years old, and so the specifics have changed; third, I can’t even begin to talk about “notch losses,” which are even crazier than a 109% marginal rate.