The Earned Income Tax Credit

Constructing an EITC with Minimal Marriage Penalties

In the module below, you can play around with the EITC outlined above and compare it to the current EITC. You'll notice that the new EITC is superior to the current one and does, in fact, eliminate all marriage penalties except for those inherent to phaseouts.

One thing that you'll notice is that the reformed married EITC is quite a bit larger than the existing married EITC: the max benefit is over $1,000 more and the phaseout is extended substantially. Importantly, this is not because I've slyly tried to fit a big benefit boost in for married people—this is the bare minimum required to eliminate marriage penalties (except for those intrinsic to phaseouts). Marriage penalties, it turns out, save a lot of money. And by all appearances, Congress has prioritized saving money over ensuring marriage is not discouraged.

This means that, with the EITC's current design, the only way Congress can get out the pickle it has put itself into—to get out of the business of discouraging marriage—is either to raise benefits for married people or to slash benefits for single people. The latter option is clearly undesirable: Congress would be throwing millions of single parents into poverty. So it's either a big EITC boost for married couples, or we consider a major re-thinking of the EITC.

There are more issues

Even if you're on board with an EITC boost for married people (it's not a bad idea in itself!), a big problem remains. So far, we've only dealt with two curves: the one-child single EITC and the two-child married EITC that a couple in which each partner has one child would jump up to. This obscures the complexity of the reform, because the EITC doesn't consist of only two curves: it consists of eight. The structure of the EITC is frankly a bit absurd. Words you'd generally want to associate with a policy are things like "simple", "elegant", and "easy to understand". The EITC instead brings to mind words like "convoluted", "ugly", and "perplexing". Just look at it:

The way in which the EITC differentiates benificiaries by the number of children that they have makes it difficult to construct a benefit schedule that is both free of marriage penalties and easy to understand. Remember that when we dealt with the one-child EITC, we were really solving for the problem of when a person with one child marries someone else who also has one child, and so they bump up to the married two child EITC. The fact that we have to add together children to determine the married EITC complicates things substantially, because it creates a large number of child combinations (16, to be exact). A couple can have two children, for instance, not only when one partner has one child and the other also has one, but also when one has two children and the other has none. So our two-child married EITC will need to take into account both possibilities, or else we'll end up with marriage penalties for one of the two groups. And that's only for two children: there's separate EITC's for one, two, and three+ children, so we have to consider married couples with one, two, three, four, five, and six+ children.

To truly eliminate marriage penalties, the only way to handle this is to deal with the worst case. i.e. the case that necessitates the largest benefit. If you design the EITC around a child combination other than the worst case, then you get marriage penalties. So, for instance, the three child benefit needs to be designed around a two-child parent marrying a one-child parent, because adding up the single two-child and single one-child EITC's results in a larger benefit than adding up the single three-child and single childless EITC's. If we didn't do this, and instead constructed the three-child married EITC to accomodate a three-child parent marrying a childless person, then any couple in which one spouse has two children and the other one child would face substantial marriage penalties.

At the same time, the phase-in for the single three-child EITC is steeper than the phase-in for either the two-child or one-child EITC. Which means that our married three-child EITC is going to need to phase in at least as fast as the single three-child EITC, or else we'd create marriage penalties for three-child—zero-child couples who have incomes that put them in the phase-in region.

This necessity to deal with the worst-case itself creates some complications. Notice that if a parent with two children married another parent with two children, then they have four children. The current EITC doesn't have a married four-child EITC—it shoves them into the 3+ children EITC. But if we were to construct the married 3+ children curve around the idea of the couple having three children, then couples with more than three children would face marriage penalties. And as the number of children increases (up to six), these marriage penalties grow larger and larger, until they're stupid large. Under the current EITC, when a parent with three or more children marries another parent with three or more children, they can face marriage penalties of over $10,000!

This means that, if we really want to eliminate marriage penalties, we'll either need to create married EITC's for people with four, five, and six or more children, or we'll have to set the married three+ children EITC equal the worst case possibility of 6+ children. The former would cost less but would substantially increase the complexity of the EITC, so I'll consider the latter possibility. Either way, however, the largest married EITC (for 6+ children) is going to have to be much, much larger than the current largest EITC.

All of these considerations together result in the below, ghastly EITC:

The childless EITC is constructed straightforwardly from the process we used above to build a (mostly) marriage-penalty free EITC (i.e. it's "double" the single EITC). The married two-child EITC is exactly the one we constructed in the previous section, only with a steeper phase-in to prevent marriage penalties among two children–zero children couples. The married three or more children EITC shows the worst-case possibility of a parent with three or more children marrying someone else with three or more children. And the married one-child EITC—the EITC for a person with one child who marries a childless person—takes the single one-child EITC and increases the maximum value by $600, the largest benefit a childless person can receive.

I hope it is evident by now that the current structure of the EITC, based around differentiating beneficiaries by the number of children that they have, makes it painfully difficult to build an EITC benefit structure that doesn't have marriage penalties. Fortunately, there's no law of the universe that says we have to construct it this way. It's just a policy choice, so we can reformulate the EITC along whatever lines we like. And, as I'll explain in the next two sections, it's not only marriage penalties: there are a many reasons why we should restructure the EITC.

Solution 1: Simplifying the EITC

So, you've come to terms with the fact that the EITC as currently structured makes it impossibly difficult to eliminate marriage penalties. You might be feeling sad that such a vaunted policy turns out to be irredeemably flawed, destined to encourage family dysfunction. Now, I will add salt to the wound, and lay out a number of other ways in which the EITC as currently constructed is deeply broken.

In this section I'll operate within the EITC's basic logic—that it should include a phase-in to "encourage work" and a phase-out to "target" the benefit at low-and-working class people and reduce the sticker price. In the next section, I'll challenge that logic and argue for a more substantial reform.

Along with the infeasibility of eliminating marriage penalties, there are three additional reasons why the current practice of stratifying benificiaires by the number of children that they have is bad:

1. It's needlessly complicated

The tax code is already complicated and confusing.

As the Treasury Deparment notes: "It can be challenging to distinguish between the portion of a credit that offsets an individual tax liability versus the portion that is refundable...Whether a tax credit results in a refund instead of a reduction in tax liability depends on all of the taxpayers' characteristics, such as income, deductions, and other credits claimed, and is not simply driven by the credit itself...The IRS’s risk assessments on refundable tax credits have consistently concluded that overclaims are not rooted in internal control deficiencies, but instead are due to the complexities of verifying eligibility, including unavailability of relevant third party data, for refundable tax credits within the time periods prescribed by the tax system."^[Treasury]"Agency Financial Report: Fiscal Year 2022" (2022). p 191. Treasury Department. Archived from the original on January 3, 2023.

According to the Treasury Department, "faced with the complexities of claiming certain refundable credits...50 percent of taxpayers claiming refundable credits use paid preparers".^[Treasury]"Agency Financial Report: Fiscal Year 2022" (2022). p 192. Treasury Department. Archived from the original on January 3, 2023. The complexity of the EITC, in other words, effectively siphons hundreds of million of dollars intended for low-income people into the coffers of H&R Block, Turbotax and the like.

The complexity of the EITC also results in lower-income people being audited at considerably higher rates than middle and upper class people.^[TRAC]"IRS Audits Few Millionaires But Targeted Many Low-Income Families in FY 2022" (January 4, 2023). TRAC. Syracuse University. Archived from the original on January 11, 2023. Headlines do tend to exaggerate this somewhat: according to the Congressional Research Service, it is true that the audit rate for people who receive the EITC is several times higher than for people who do not, but this represents and increase from a 0.3% audit rate to a 1% audit rate.^[CRS-EITC]Crandall-Hollick, Margot L. (June 13, 2022). "Audits of EITC Returns: By the Numbers". Congressional Research Service. Archived from the original on June 14, 2022. In 2022, the IRS conducted 310,081 refundable credit exams.^[Treasury]"Agency Financial Report: Fiscal Year 2022" (2022). p 192. Treasury Department. Archived from the original on January 3, 2023.

This complexity not only produces headaches for EITC claimants, but also results in additional costs to the government in the form of improper payments and additional overhead costs required to audit and recover these payments. In 2022, for instance, the Treasury Department estimated that roughly $18 billion in EITC payments (32% of the total) was sent out improperly, representing 10% of the tax gap.^[Treasury]"Agency Financial Report: Fiscal Year 2022" (2022). p 48. Treasury Department. Archived from the original on January 3, 2023. The comparatively simpler child tax credit, by contrast, had an improper payment rate half that and made up only 3% of the tax gap. The IRS then has to expend significant resources detecting and auditing improper payments, which is made considerably more challenging by the EITC's eligibility rules, which depend upon complicated family relationships and residency arrangements that can be difficult to verify. Indeed, over 90% of improper payments result from "Inability to Authenticate Eligibility: Data Needed Does Not Exist", i.e. payments in which the IRS was simply unable to determine whether eligibility criteria have been met.^[TAS]"Taxpayer Advocate Service 2018 Annual Report to Congress: Improper Earned Income Tax Credit Payments" (2018). p. 91–104. Taxpayer Advocate Service. Internal Revenue Service. Archived from the original on January 19, 2022.

Both the IRS and the Treasury have repeatedly made it clear that the root cause of overpayments is the complexity of the EITC and of the tax system more broadly, not fraud. As the Treasury notes, "the IRS has continuously deployed costly enforcement tools to lower the error rate" with "minimal impact".^[Treasury]"Agency Financial Report: Fiscal Year 2022" (2022). p 245. Treasury Department. Archived from the original on January 3, 2023.

Finally, the complexity of the EITC and of the tax system more broadly results in many people not claiming the EITC despite being eligible for it. Both the IRS and non-government studies consistently find that roughly 20% of people eligible for the EITC do not claim it.^[IRS]"EITC Participation Rate by States Tax Years 2012 through 2019" (2022). Internal Revenue Service. Archived from the original on November 27, 2022.^[Hamad]Hamad, Rita; Gosliner, Wendi; Brown, Erika M.; Hoskote, Mekhala; Jackson, Kaitlyn; Esparza, Elsa M.; Fernald, Lia C. H. (December 2022). "Understanding Take-Up Of The Earned Income Tax Credit Among Californians With Low Income". Health Affairs. 41 (12). DOI: 10.1377/hlthaff.2022.00713. Around 25 million people claim the EITC annually,^[IRS-2]"Statistics for Tax Returns with the Earned Income Tax Credit (EITC)" (2022). Internal Revenue Service. Archived from the original on November 27, 2022. which means that over 6 million people who are eligible for the EITC do not claim it.

2. It's incoherent—the EITC is trying to be both a wage subsidy and a family benefit

3. Childless workers receive little—and may even be hurt by the EITC

The childless EITC is a pittance compared to the EITC's for families. You have to squint just to see it on the chart. The consequence of this is severe: according to the Center on Budget and Policy Priorities (CBPP), more than 5 million childless workers are taxed into poverty each year.^[Marr-2]Marr, Chuck; Huang, Yixuan (March 2, 2020). "Childless Adults Are Lone Group Taxed Into Poverty". Center on Budget and Policy Priorities. Archived from the original on December 21, 2022. The childless EITC is simply much too small to offset the income and payroll taxes owed by childless people near the poverty line, and as a result they end up in poverty after their taxes are subtracted.

This is a feature that is unique to childless people: they are the only group that can be taxed into poverty.^[Marr-2]Marr, Chuck; Huang, Yixuan (March 2, 2020). "Childless Adults Are Lone Group Taxed Into Poverty". Center on Budget and Policy Priorities. Archived from the original on December 21, 2022. People with children who have incomes above the poverty line are spared this fate by the child tax credit and the much larger EITC's that they are eligible for. The childless EITC, by contrast, just isn't up to the job.

This is a problem that should be remedied: no one who works enough to keep themselves out of poverty should see the federal government shove them back down into it. The simplest solution, naturally, is to raise the childless EITC, as was done for one year under the American Rescue Plan Act of 2021 and as has been proposed in the Build Back Better Act and the Working Families Tax Relief Act. This is a worthwhile reform, and it should be supported. This page, however, is principally about eliminating marriage penalties, and these targeted reforms of the EITC would not do that (in fact it would make them worse). So instead we will remedy this deleterious feature of the EITC within a broader reform that is designed to simultaneously reduce marriage penalties while also addressing all of the criticisms laid out above.

The Proposal

In this proposal, I simplify and rationalize the EITC. We'll collapse it down to just two benefits: a single benefit and a married benefit. We'll construct them as we did in the section above to minimize marriage penalties. This redesign results in an EITC that is:

1. Pro-Marriage: It largely eliminates marriage penalties.
2. Simple: It makes the program easy to understand, simplifying an already complicated tax code and reducing improper payments
3. Coherent: It gives the EITC a coherent purpose: it's a wage supplement for working class people, simple as that.
4. A fair shake: There's no reason to largely exclude low-income childless people from the EITC—they need help too.
5. Pro-Work: The EITC is designed to encourage work, so it is profoundly strange that we've largely excluded childless people—who make up a large share of prime working age people—from it. The redesign extends the EITC's work encouraging properties to childless people.

Reformed EITC

Comparison with the Family Security Act

You might notice that this reform is similar in nature to the EITC proposed by Mitt Romney in his Family Security Act. Both proposals simplify the EITC and accompany it with an increased child benefit. And as I also will advocate later in this page, Romney proposes eliminating the head of household filing status and putting the savings towards the child benefit. There are, however, some differences between the two proposals, particulary as it relates to the EITC. For reference, this is Romney's proposed EITC:

Family Security Act EITC

With Romney's proposal, it should be noted, he went to lengths to ensure that the policy is revenue neutral, whereas I have been more lackadaisical with funding. This placed particularly strong constraints on his EITC reform, as he relied on it to provide nearly half of the funding for the FSA's enhanced child benefit. While I don't know his exact thought process, these funding constraints seems like they could be a major reason he didn't fully consolidate the program into one single and one married benefit, instead maintaining separate benefits for childless people and for parents. He likely needed to do that because raising the childless benefit up to the value of the benefit for parents would substantially reduce the amount of revenue his EITC reform would save, and thus push his proposal into the red; conversely, dropping the benefit for parents down to the childless benefit would be such a substantial benefit reduction compared to the current EITC that virtually all working class people would see their benefits cut substantially even with the larger child benefit.

Under similar budgetary constraints, the EITC reform I proposed might need to be similarly altered. These alterations, however, have major drawbacks. While Romney's EITC generally reduces marriage penalties, it is still a good ways from eliminating them. The phase-out, of course, inherently imposes marriage penalties, but there are also marriage penalties for many couples in which each partner has children of their own. These marriage penalties are generally not as severe as those that currently exist, but they can be as large as $1,000.

The reductions to the EITC for parents are also a bit overzealous (alternatively, the child benefit increase is not large enough). Even though the FSA substantially raises the child benefit, the EITC reductions are so severe that about 7 million single-parent families would see their benefits reduced, according to the CBPP.^[1]Marr, Chuck; Cox, Kris; Hingtgen, Stephanie; Sherman, Arloc; Calame, Sarah; Cook, Jabari (July 6, 2022). "Romney Child Tax Credit Proposal Is Step Forward But Falls Short, Targets Low-Income Families to Pay for It". Center on Budget and Policy Priorities. Archived from the original on October 21, 2022. The FSA would still reduce child poverty substantially,^{[note-number]}The Niskanen Center estimates that Romney's original proposal would reduce child poverty by a third and deep child poverty by half (Hammond, Orr 2021), while the FSA 2.0 would reduce child poverty by 13% (Orr, McCabe 2022). but it would certainly have a greater impact if either (1) the EITC were not so sharply cut, or (2) the EITC reductions were fully offset by a child benefit larger than that proposed.

The latter option would be preferable, as a child benefit is in general superior to the EITC, but it would require raising more revenue than a less sharp reduction to the EITC. In any case, the structure of Romeny's EITC is a big step in the right direction, and it is likely only for budgetary reasons that he did not go the full ways to the structure I have proposed.

Solution 2: Child Benefits & the Standard Credit

So we've got a reformed EITC that mostly eliminates marriage penalties. However, within the EITC's basic design principle there's just no way to completely eliminate them: phase-outs necessarily produce marriage penalties. Fortunately for those of us who would prefer not to discourage marriage, there are good reasons to doubt two fundamental aspects of the EITC: that it should include a phase-in and a phase-out.

1. Phase-Outs Discourage Marriage

As we've already established, there's just no way to eliminate marriage penalties without eliminating phase-outs.

2. Phase-Outs Discourage Work

The EITC is designed to promote work by topping up the wages people make from work. In economic terms, the EITC makes use of the substitution effect: by increasing the wage of an additional hour of work, the EITC increases the cost of not working, and thus increases the incentive to work more (to "substitute" additional wages for additional hours of leisure). It must be noted that there is a countervailing income effect—as income rises, people demand a higher quantity of goods, including a higher quantity of leisure–but research suggests that the substitution effect dominates.

In more intutitive language, all this is saying is that if your wage is, say, $15 an hour, then the phase-in of the one-child EITC increases this to $20.10 an hour. When you're considering whether to work additional hours, that $20.10 an hour makes the additional work more appealing than the $15 an hour wage would.

However, this only holds in the phase-in region. The plateau has little effect on work incentives, and the phase-out reverses the incentives. In the phase-out, additional hours of work reduce the amount of the EITC. For example, the one-child EITC phases out at a rate of $210.60 per $1,000 in additional earned income. This means that if you work enough to earn an additional $1,000, your income in fact only rises by $789.40.

This puts the substitition and income effects into reverse. The EITC phase-out reduces additional wages, and thus reduces the cost of not working additional hours, and thus reduces the incentive to work more. Again, the income effect works in the opposite direction, but the substitutiton effect appears to dominate, which accords with intution: very few people would suspect that taxes increase the incentive to work, and all the EITC phase-out is doing is imposing an effective tax (you can read more about effective marginal tax rates on our page Marginal Tax Rates).

Importantly, other programs like SNAP ("food stamps") and Supplemental Security Income (cash payments to disabled people who are unable to work) also impose effective taxes. As a result, they can compound on top of each other to produce very large effective tax rates: an additional $100 of earned income, for instance, can cause a $21.06 loss in the EITC, and a $24 loss in SNAP benefits, and a $50 loss in SSI benefits, resulting in only $3.96 of actual new income (and this is before income and payroll tax). Together these create large work disincentives—additional work results in very little additional income.

There is really only one way to solve this: don't impose phase outs on working class people.

3. The phase-out is really just an inferior way to fund the program

All Together

We've now established that phase-outs necessarily discourage marriage and discourage work. And the only arguments in their favor—that they reduce the cost of the EITC and "target" benefits to lower-income people—turn out to be a mirage. It is just much less effective to concentrate effective taxes onto a small group of people who fall into the phase-out than it is to give everyone the benefit and spread out the taxes.

So if we take the reformed EITC I proposed in the section above and eliminate the phase-out, we end up with something like this:

This is a big improvement, but you'll probably notice that there's something pernicious about this. We're giving everyone a cash benefit—except for poor people! This is, on its face, absurd. Poor people are precisely the people who most need financial assistance—it is the very definition of poverty that they do not have enough money. Just as it'd be nonsensical to establish a food bank that's available to everyone except those who are starving, it's ludicrous to provide a cash benefit to everyone except for those in poverty.

So if we eliminate the phase-in, we end up with the below tax credit:

What we have here is a standard credit. While this might be a significant departure from the EITC, it's really quite similar to something that is widely accepted in our tax system: the standard deduction.

The standard credit is not a perfect policy. It has the hallmark problems of a benefit delivered through the tax code rather than through direct spending. You can read about these problems [[[here]]]], and Matt Bruenig has written extensively about them with regards to the child tax credit at the People's Policy Project. In short, tax benefits can end up excluding poor people because people with low incomes are not required to file tax returns, but without a tax return the IRS can't send them a check. This is one reason it's become an increasingly popular proposal to replace the child tax credit, administered by the IRS, with a monthly child benefit administered by the Social Security Administration, which doesn't have this problem.

I'm proposing a standard credit rather than a direct spending program, however, because it strikes me as something that politicians and the general public would be more amenable to. As the political scientists Christopher Ellis and Christopher Faricy demonstrate in their book The Other Side of the Coin: Public Opinion toward Social Tax Expenditures, "public support for a wide variety of social benefits is significantly higher when those benefits are framed as being delivered through tax breaks rather than framed as direct government spending".^{[note-number]}Ellis and Faricy have summarized the thesis of their book in an article for Fortune: "America’s messy tax code is actually quite popular" (Archived). For the full book, see (Ellis, Faricy 2021).

The standard credit fits this mold: it's a "tax cut", not a direct spending program. And as tax cuts go, a standard credit feels in nature a lot like the standard deduction, which is very popular. Additionally, since the standard credit does not necessarily have to replace the EITC—you can easily have both at the same time—a standard credit also offers the alternative pathway of first instituting it as a non-refundable credit, which could make it more amenable to Republicans and the handful of Democratic politicians who are still deeply wed to the idea of phase-ins. Then, with the non-refundable standard credit in place, it becomes a simple and cheap reform to later make it refundable.

Supplemental Security Income