Thinking Out Loud About the Advanced Child Tax Credit – Part I

First of all, it is worth going back to first principles and considering why children even factor in the tax code at all.  Many countries don’t incorporate children in their tax codes – the taxable unit is the individual.  An individual’s wages are taxed through a Pay As You Earn (PAYE) system; individuals also pay taxes through sales tax, value added taxes (VAT) or goods and services tax (GST).  Benefits to families are traditionally disbursed via health and human services agencies rather than the tax system.  But the U.S. has opted to make the family unit the taxable unit, and thus incorporated provisions relating to spouses, children, and other dependents throughout the Internal Revenue Code (IRC).  Having adopted the family as the taxable unit, the dependency exemption (IRC §§ 151 and 152) – temporarily suspended and supplanted by an enhanced child tax credit (IRC § 24) — reflects the basic principle that taxation is based on one’s ability to pay, and that ability to pay should take into consideration the size of one’s family and basic living costs associated with that family.  The dependency exemption, however, only helps people who have taxable income.  It is worthless to those families who have no taxable income, which is not a problem if all you are doing with your tax system is taxing income.  (There is another problem with the current dependency exemption and definition of “household,” namely that it is based on a 1950s, Leave-it-to-Beaver concept of the family: middle income, White, mother and father, two kids.  More on that in future posts.)

As many tax scholars have noted, a straight exemption or deduction gives a greater benefit to those with incomes high enough to fully absorb the deduction/exemption, and less benefit to those on the lower income scale whose income is less than the amount of the deduction/exemption.  Similarly, the “value” of the exemption/deduction varies depending on the marginal tax rate of the taxable unit.  Converting the exemption/deduction to a tax credit avoids the latter problem, and introducing an element of refundability addresses the first problem.  Congress added the Child Tax Credit to the IRC in 1997, later adding limited refundability; it temporarily suspended the dependency exemption and replaced it with an expanded CTC in 2017 in the Tax Cuts and Jobs Act, and now has introduced a fully refundable CTC, severing the connection between taxable income and the child benefit.  For 2021 Congress also authorized a 6-month “proof of concept” of monthly issuance of the CTC in advance.  These last changes have been lauded as having the ability to reduce child poverty from 13.7 to 11.3 percent.

Of course, the income tax has never been solely about taxing income – it is an instrument for driving social policy and economic and personal behavior.  Not only are a limited number of household structures recognized as “households” for purposes of the income tax, but those households are awarded all sorts of tax credits, deductions, and allowances that reward certain behaviors – the earned income tax credit (EITC), the American opportunity tax credit (AOTC), the child and dependent care tax credit (CDCTC), and the head of household (HOH) filing status, to name a few.  It seems like a natural progression to include child poverty reduction in the mix of these social policies run through the tax code.

Why would we do this?  Well, the IRS is, when it puts its collective mind to it, an efficient machine for processing returns and disbursing benefits.  Each year, it processes over 150 million individual tax returns and disburses 102 million refunds totaling $282 billion.  I’ve probably written over a thousand pages criticizing the IRS’s failure to address problems in processing those returns, and made scores of recommendations, but the fact remains the IRS knows how to process and move the majority of returns through its pipeline.  The three rounds of Economic Impact Payments over the last 18 months have clearly demonstrated that the IRS can do this work.

The IRS’s “efficiency” was one reason the EITC was expanded.  Originally designed as a way to offset the regressive nature of the payroll tax (applied to the first dollar of earned income), the EITC has been expanded over the years to be one of the largest federal anti-poverty programs for families.  Although it requires earned income, it is refundable; thus, households who have no taxable income can still benefit from it, and over 26 million households do, to the tune of $64 billion for tax year 2018.  

One of the arguments for running such a social benefit program through the tax code is the ease of application.  Traditional benefit programs like the Supplemental Nutrition Assistance Program (SNAP, or food stamps) and Temporary Assistance to Needy Families (TANF, or “welfare”) have cumbersome application processes which require submission of documentation, and even in-person interviews, before eligibility is determined, as well as periodic recertifications requiring additional documentation.  There is the stigma of welfare attached to the application process; hence historical participation rates have been low with respect to these programs.  Moreover, the application and review processes result in relatively high direct administration costs.

Policymakers viewed the application process via the annual tax return as relatively stigma-free, and certainly much less burdensome than having to meet the documentation requirements of state-administered welfare programs.  And the EITC is anonymous – no one need know you receive it, unlike SNAP, where you segregate your food out based on SNAP coverage and present your SNAP card to the cashier at the food store, and all the people in line behind you see it.  The low cost of administering the program via tax return processing, especially with the near-universal use of e-filing, is another efficiency, compared to the personnel-intensive application process for traditional safety-net programs.  These features result in a relatively high participation rate for the EITC vis a vis other programs.  (For a comparison of programs, see Figure 2.1.5 from my 2016 legislative recommendation about family status provisions in the code here.)  The ease-of-application has downsides, however:  high noncompliance rates.  Treasury estimates the EITC improper payment rate is about 25%. This has led to disproportionately high audits of low income taxpayers, as well as an increasing use of summary assessment authority (aka “math error”) under IRC § 6213 and highly punitive penalties, including the 2 year/10 year ban under IRC §32(k).

The other downside to running social benefit programs through the tax code is its annual and retrospective nature, as well as the calendar year aspect of the annual determination, which may or may not conform to patterns of family relationships.  The tax return is filed after the close of the tax year, and eligibility for all of its provisions, including the EITC and other family-status credits, is determined retroactively.  This is a major difference with traditional social welfare programs; these programs accept applications throughout the year, and eligibility is determined on a prospective basis, meaning the benefit commences from the date of application, once the agency has reviewed the application and supporting documentation and determined eligibility.  Further, the agency’s determination means that for a period of time, until recertification, the applicant is presumptively eligible for the benefit unless someone else comes along and files a competing claim.  Under presumptive eligibility, claw back of benefits is rare, particularly during the period of dispute resolution.  If someone else was actually eligible for claiming the benefits during the presumptive period, but didn’t claim those benefits, well tough.  The presumptively eligible person has been defined as the correct person to receive the benefits, even if they aren’t.  Thus, noncompliance is defined out of existence.  Even if we get the wrong result, it isn’t the wrong answer because we have declared it isn’t the wrong answer.  Therefore, we have no improper payments.

Of course, it isn’t as bad as I make it sound, as it exists in other social benefit programs.  First of all, as noted earlier, traditional safety net programs have intensive and administratively burdensome and costly application processes; if an applicant makes it through this process (many don’t even try), there is a fair amount of confidence that the person actually is eligible for the benefit.  Second, the recertification process (e.g., 12 months in Washington, DC for SNAP, unless you are homeless in which case recertification is 6 months) ensures a regular check on eligibility.  Third, the process accommodates the fluidity of families at the lower income levels.  You can come in at any point in the year and apply for benefits; you don’t have to wait until the end of the year to apply to receive retroactive benefits; you receive the benefits when you need them.

It is these features that have led some advocates to propose presumptive eligibility for the AdvCTC.  The proposals generally require a taxpayer to check a box on their tax return that affirmatively states, under penalties of perjury, that you are eligible to receive the AdvCTC going forward on a monthly basis, generally starting the month after the end of the filing season.  It isn’t clear what definition of eligibility the advocates are proposing – is it the Qualifying Child of IRC §§ 151 and 152, which requires both a specified relationship and sharing a principal residence with the child for more than 6 months of the year, or is it some other test, like the “main carer” test that Canada and Australia have adopted.  (Canada Revenue also has a presumption that the mother of the child is the main carer, which is really troubling.  See Moritz v. Commissioner, here.)

The proposals adopt the approach of traditional benefits programs, with a few alterations.  For the AdvCTC, presumptive eligibility would be based on an affirmative statement and a checkbox on the return; unlike traditional benefit programs there is no application nor is any supporting documentation submitted.  Further, unlike traditional benefit programs, there is no obligation for the presumptively eligible claimant to update the IRS about any change of circumstances.  Instead, the burden is placed on newly eligible claimants to make a claim, and if the agency finds that this new claimant is eligible going forward, the AdvCTC payment will be paid to that claimant from date of application. Like traditional benefit programs, eligibility is based on the date the new claimant applies, not when they started serving as the primary caregiver or when the child actually resided in the new home.

Unlike the current design of the AdvCTC, in many cases the only way someone could claim the CTC on that year’s tax return is if no one else had claimed the AdvCTC for the child for that particular month.  There is no reconciliation to determine whether AdvCTC had been paid to an ineligible person during the year, absent fraud; as noted earlier, improper payments are defined away because the person is presumptively eligible.  Thus, there are no claw backs.  There can only be additional payments made with the tax return, if no one else has received a payment for that child for that month.  If the child lived with you and you were the main carer of that child during a month for which someone else received the AdvCTC, tough for you.  You should have known about the AdvCTC and come in and claimed it.  Some proposals have a grace period for the first year of implementation; my experience is it will take at least 3 years for new rules to be understood.  On the other hand, if a child is unclaimed for any given month during the year, you may receive the CTC retroactively if you claim it on your return (and you meet the 6 months-and-a-day principal residence test).

Perhaps this system works for traditional benefit programs – it seems to me the combination of intensive application process, the requirement that the presumptively eligible person report change of circumstances, and the recertification process (with application and documentation) provide some checks and balances to get to a comfort level that for the most part, the right households are getting the benefits.  But these traditional benefit programs have high operational costs and low participation rates.  Why would we want to import this process into the tax benefits system that has low operational costs and high participation rates?  Why would we bring in features that the tax benefits system was actually designed to eliminate?

There’s another problem here – people are forgetting that unlike the traditional social safety net programs, the Child Tax Credit is almost universal – that is, it isn’t just targeted to the lowest income levels or those who have no income.  In Tax Year 2018, the CTC was claimed on almost 39.4 million tax returns for almost $81.5 billion.  Over 26 million taxpayers claimed the Additional CTC, for about $36 billion; 19.5 million of these taxpayers received $19.5 billion as refunds (i.e., not offset against taxes owed).  (See Figure D, IRS 2018 Statistics of Income.) That is a huge swath of taxpayers.  The near-universality of the CTC brings some complications.

Will we regress and have more than one definition of a qualifying child instead of working toward a single definition?   For example, how will the IRS administer the EITC, the Child and Dependent Care Credit, and head of household status if the presumptively eligible person did not reside with the qualifying child for more than 6 months of the year?  Will the actually eligible person receive these three tax benefits but not the CTC, because the presumptively eligible person received the AdvCTC throughout the year?  Or will the presumptively eligible AdvCTC person also be presumptively eligible for the other family status provisions, compounding the harm to actually eligible persons?  (There will have to be a form on which to claim the CTC where someone opted out of the monthly payments or where no one claimed the AdvCTC for the child for a given month.)  

Advocates for the presumptive approach argue that by defining away improper payments, you avoid clawbacks, theoretically reduce political pressure by eliminating or reducing improper payments,  and eliminate the need for the IRS to reconcile competing claims at tax filing time.  I find these arguments naïve.  First, I can guarantee you that the Treasury Inspector General for Tax Administration (TIGTA) will be conducting an audit to determine whether there is any congruence between who were presumptively eligible and who were actually eligible.  When that hits the press, the noise will be great about the amount of payments that went out to people who did not care for the child.  Second, there will be competing claims regardless of how you design it – taxpayers who are actually eligible will be claiming the CTC on their returns and the IRS will have to send out notices stating someone else received the AdvCTC for those months; then the IRS will have to respond to those angry phone calls, and there will be calls to the Taxpayer Advocate Service and congressional offices, and news media will pick up stories about people who weren’t able to get the CTC because the noncustodial ex-spouse received it.  Third, what about those taxpayers who don’t want a monthly benefit and opt out?  Will they find, at the end of the year, that someone claimed the child and therefore they will not be eligible for a retroactive benefit?  Will they be forced to come in during the year and prove their eligibility for something that they don’t want to claim or deal with until the end of the year?  Or are we paternalistically saying that everyone has to get this payment monthly, even if for your circumstances it is better to receive it at year end?

Advocates also say that, as with the traditional benefits system, there will be a dispute resolution process throughout the year.  I believe it is possible to have an expedited administrative dispute resolution process to resolve competing claims throughout the year, but I have my doubts about having an expedited Tax Court or other judicial procedure during the year.  Right now, in speedy cases it can take over a year for the Tax Court to resolve an EITC dispute.  Remember that while the dispute is ongoing, the presumptive recipient is continuing to receive the AdvCTC; even if the dispute is resolved in favor of the new applicant, does the benefit start when the court decision is final or does it relate back to the date of application?  What if that is after the end of the tax year?  How does that impact an ultimately successful taxpayer’s eligibility for the EITC and head of household status for the tax year in question?

The final point I would like to make is that advocates are actually trying to resolve the right problems; they just have come up with the answers that aren’t an easy fit for a system that is administered through the tax code.  We do need to deal with improper payments – but it is best done by looking at the Payment Integrity Information Act (PIIA) and really analyzing whether overpayments of programs like the EITC or AdvCTC are, in fact, improper payments after all.   TIGTA has reported that Treasury and the IRS requested the Office of Management and Budget exempt IRS from PIIA reporting requirements.  (Read Procedurally Taxing’s blog about this here.)  And we need to have a more flexible definition of the family unit, to accommodate different family structures at all income levels.  There are legislative fixes to these issues.

Advocates also argue that presumptive eligibility eliminates the need for a large safe-harbor, which, they believe, encourages improper claims and has reputational risk.  To me, this makes no sense.  Presumptive eligibility encourages a race to be the first to claim the checkbox on the return – the first-to-file get the credit, at least for a few months, until the second-to-file’s claim is “adjudicated.”  It is not enough to say that there will be a claw back for fraud – how is the IRS going to prove that at the time of checking the box on the return, you did not intend for the child to live with you?  And here, to me, is the fatal flaw – because payments to the actually eligible taxpayer are dependent on when that taxpayer applies for the monthly credit, the windfall to the presumptively eligible person comes at the expense of the actually eligible person (i.e., the person who actually is the main carer of the child).  That just seems to me to be a violation of the right to a fair and just tax system.

This is not to say we can’t do monthly payments for children.  But if we want to administer them through the tax system, then we have to understand how they fit into the greater scheme of the tax system, and we can’t just impose a system that was created for a completely different purpose onto the tax system.  In tomorrow’s post I will explore about how we could administer monthly advanced child tax credit payments through the tax system we have, and what changes would be needed to the tax system to make it work.  I believe there is a way to administer a monthly advanced CTC that is paid to the taxpayer/household(s) that actually care for and have responsibility for the child in any given month.  I may not have the complete solution, but I will take a stab at coming up with some ideas.  And I look forward to everyone’s thoughts and comments.