Pathways to Reduce Child Poverty: Impacts of Federal Tax Credits (2026)
Chapter: 4 Take-Up of the EITC and CTC Under ARPA in 2021
4
Take-Up of the EITC and CTC Under ARPA in 2021
For many federally administered programs in the United States—including the Earned Income Tax Credit (EITC) and Child Tax Credit (CTC)—a set of rules regarding income, assets, and family structure, among others, determines eligibility. For many programs, federal and state governments require that people submit forms (and sometimes meet with caseworkers) to certify their eligibility and claim benefits. The delivery of benefits is not automatic: the number of people who receive benefits among those eligible to receive them—a concept referred to as “take-up”—is often not even close to 100%.1
In the case of tax credits, the process of claiming benefits is administered through the tax system and entails filing tax returns to confirm eligibility. Most adults and a sizeable number of minors in the United States file tax returns (or are included on one) every year. Due in part to the tax filing requirement, receipt of tax credits among eligible populations is typically higher than receipt of many other means-tested programs among eligible populations.
This chapter summarizes the evidence on take-up rates for EITC and CTC policies enacted by the American Rescue Plan Act of 2021 (ARPA), as prior versions of these two credits, and how take-up varied across
___________________
1As discussed in more detail throughout this chapter, it is often very difficult to identify with complete accuracy the population eligible for tax credits. While the committee uses the term “take-up” throughout the chapter, note that in many cases the denominator used to construct “take-up” may, in fact, include ineligible individuals. In these cases, it may be more appropriate to conceptualize this term as “participation.”
demographic groups. Eligibility requirements for the ARPA EITC and CTC policies in tax year (TY) 2021 (and surrounding years) are laid out in Chapter 2. The sections that follow consider the strengths and limitations of evidence from various types of data sources, including data from the Internal Revenue Service (IRS) and other government agencies, and self-reported accounts of take-up from surveys (see Chapter 8).
Tax credit take-up is difficult to measure with much accuracy, both because some individuals who are not eligible to receive credits nevertheless receive them (affecting the numerator in the take-up calculation), and because eligibility in the broader population cannot be perfectly observed (affecting the denominator in the take-up calculation). The second part of the chapter discusses plausible reasons why measured take-up of tax credits is incomplete and potentially subject to measurement error. The tax system is complex and filing mistakes may occur—either by individuals or tax preparation services. Even for the IRS, it is sometimes difficult to identify tax credit eligibility or to identify people who may not be eligible for credits but have received them anyway, whether unintentionally or otherwise. Thus, estimates of take-up are typically imperfect.2 The IRS attempts to measure what it calls noncompliance, or overclaiming by those the IRS deems ineligible, but error likely exists in this measurement as well.
Finally, this chapter focuses primarily on evidence drawn from what will be referred to as quantitative studies—data from various surveys and other sources that report on EITC and CTC take-up for what are, or are thought to be, representative samples of the U.S. population or its subgroups. These sources allow the committee to provide numerical estimates of take-up rates among the population and specific subgroups. Another important dimension of take-up can be informed by qualitative research studies examining how individuals and households—especially those that are more disadvantaged—learned about the changes made to the EITC and CTC policies under ARPA in 2021, how they navigated access to these credits, and what barriers they encountered in doing so. Chapter 5 describes findings from qualitative research on these matters and insights gained through listening sessions. These listening sessions utilized many of the strategies employed in previous qualitative studies to enrich understanding of the important dimensions of credit take-up.
Main messages from this chapter are highlighted in Box 4-1.
___________________
2For example, requirements state that filers have a Social Security Number (SSN) that is valid for employment to be eligible for the EITC, and since TY 2018, children were required to have their own SSNs that are valid for employment to qualify for the EITC and CTC while filers need an SSN or Individual Taxpayer Identification Number (ITIN). Many Hispanic children may lack SSNs and, in most survey datasets, individuals do not report their ITIN/SSN holdings. Similarly, while the IRS attempts to flag noncompliance, it does not always flag it correctly and thus may mismeasure some ineligible individuals as valid recipients.
BOX 4-1
Chapter 4 Main Messages
- Not all families who are eligible for tax credits end up receiving them; this has implications for the committee’s calculations of how tax credits affect child poverty.
- Incomplete take-up varies across demographic characteristics—Hispanic families tend to have lower take-up rates compared to Black and non-Hispanic White families, and lower-income families tend to have lower take-up rates than higher-income families.
- Several data issues limit the ability to accurately measure which families are eligible for and receive tax credits. These include the lack of information on how household composition translates into tax filing units, and reliance on imputations from tax calculators to estimate tax credit eligibility, rather than administrative accounts or self-reported data on tax credit participation.
- Several factors contribute to incomplete take-up, including lack of information on eligibility for tax credits, ease of filing tax returns for potentially eligible groups, noncompliance, and Internal Revenue Service capacity issues, some of which might be unique to the changes to the Earned Income Tax Credit and Child Tax Credit policies that occurred under the American Rescue Plan Act in 2021.
EXISTING EVIDENCE ON TAKE-UP OF THE EITC AND CTC
Data Sources Used to Measure Take-Up
Various data sources are used to measure take-up, and, throughout this chapter, information is drawn from sources including aggregate numbers released by the IRS, individual administrative tax data, and various surveys that include self-reported receipt of tax credits or imputed values generated from tax calculators. Each of these data sources has strengths and weaknesses. Survey data, for instance, rely on self-reports that can suffer from recall, social desirability, and related biases that often lead to underreporting of benefit receipt (Meyer & Mittag, 2019). Many surveys fail to ask about tax filing status or receipt of tax credits and instead impute these values, further complicating matters.
Data from the IRS and comparable agencies have overcome many of these challenges by more completely representing the population receiving benefits and, importantly, accounting for benefit receipt more accurately. However, identifying information for certain demographic groups is less reliable or not collected at all, and no information is available for certain populations, such as individuals who do not file tax returns, making it difficult to identify individuals eligible for credits.
More recently, several efforts have been made to link survey data with administrative tax data (e.g., the Current Population Survey Annual Social and Economic Supplement [CPS ASEC] linked with IRS data) to incorporate demographic and geographic information into tax data. These data linkages leverage the strengths of survey and administrative data and present new opportunities for researchers to better understand participation in social welfare programs. As discussed in more detail below and in Appendix E, these linkages allow users to estimate tax credit receipt for various demographic groups and to identify discrepancies between self-reported accounts from survey data and administrative benefit receipt information. The discussion that follows reports take-up rates from the literature based on various types of quantitative data, noting where differences exist across sources.
Evidence on Take-Up from IRS Data
The complete picture of participation in the EITC and CTC in 2021 is only beginning to be understood. At the time of this writing, the IRS reported that approximately $93.6 billion in advance CTC payments were made in 2021 (IRS, 2021), reaching an estimated 61.9 million children. Additionally, approximately $116 billion in payments were made during tax filing season in early 2022 (IRS, 2025).3 While there are official reports regarding the number of individuals who received benefits and how much was spent on benefits, the difficulty in accurately measuring take-up of a program is that it is difficult to identify with certainty which population is eligible for benefits. Since the CTC under ARPA made benefits available to nearly all U.S. citizen children in the United States, one way to determine eligibility is through U.S. Census Bureau reports on the number of children living in the United States in 2021. According to American Community Survey data from the Census Bureau, approximately 73.5 million children under the age of 18 lived in the United States in 2021 (American Community Survey, 2021), which implies that approximately 84% of all U.S. children received a monthly CTC payment in 2021. However, not all U.S. children were eligible for the monthly credits, so this number likely understates benefit receipt among those eligible.4
___________________
3This number reflects the number of unique children who ever received a monthly payment and does not include payments made at the time of tax filing for the 2021 tax year, which would have been distributed in early 2022. A U.S. Government Accountability Office report (Government Accountability Office, 2022b) stated that approximately 2.1 million families opted out of monthly payments in favor of claiming the benefits (presumably) at tax time in early 2022.
4For instance, children without SSNs were not eligible, and neither were children residing in households in the top 1–2% of the income distribution. For a more detailed discussion of eligibility criteria, see Chapter 2.
The IRS reported receiving approximately 32 million tax returns claiming the EITC in TY 2021 and calculated an 80.8% take-up rate among eligible filers,5 representing $65.7 billion in federal payments (IRS, 2025). Take-up of the EITC in 2021 was similar to rates reported in previous years: between 2013 and 2021, the IRS estimated that EITC take-up ranged between 76% and 81%.6 No parallel analysis of CTC take-up has been reported by the IRS in recent years.
Evidence on Take-Up from Survey Data
Survey data reports of receipt of monthly CTC payments in 2021 tend to be lower than rates reported using administrative data, with several studies finding a self-reported take-up rate of approximately 66% using data from the Census Bureau’s nationally representative Household Pulse Survey (HPS; Chiang et al., 2023; Parolin et al., 2021). Similarly, Michelmore and Pilkauskas (2023) found a take-up rate of 67% among a population of low-income families using a survey administered by Propel, the makers of an app to help food stamp users track their benefits. This lower take-up could reflect the disadvantaged nature of the population; take-up rates tend to be lower among populations within incomes below the tax filing requirement level. However, the authors also found that approximately 80% reported ever receiving a monthly payment, which is closer to IRS estimates.7 Using smaller surveys, other studies have reported take-up rates ranging between 75% and 80% (Burnside et al., 2022; Hamilton et al., 2022b; Karpman & Maag, 2022). Prior evidence suggests that individuals may underreport receipt of social benefits in survey data, which may partially explain why take-up rates in these various surveys lags that of the IRS estimates (Meyer & Mittag, 2019).
Variation by Subgroups
Overall take-up rates for the EITC and CTC are high relative to other social safety net programs; nevertheless, wide variation exists across demographic and economic groups. This section discusses differential take-up rates by race and ethnicity, income, educational attainment, family
___________________
5https://www.eitc.irs.gov/eitc-central/participation-rate-by-state/eitc-participation-rate-by-states
6Using CPS ASEC data matched to IRS data, Jones (2014) also estimated EITC take-up of 75–80% between TYs 2005 and 2009.
7An additional 8% of families reported that they did not receive the benefit for some legal or “valid” reason, such as another person in the household claimed the child on their tax filing, or that the child only recently moved into their household and thus did not meet qualifications for receiving the Child Tax Credit.
structure, and across the business cycle.8 Issues related to subgroup variation are discussed in greater detail in Appendix D.
Variation by Race, Ethnicity, and Immigration Status
Several studies show differences in take-up of the EITC and CTC by race and ethnicity. Nearly all these studies find lower take-up rates of both credits among Hispanic families, but similar or slightly higher take-up rates among Black households relative to White households (e.g., Chiang et al., 2023; Collyer et al., 2023; Jones, 2014; Parolin et al., 2021; Thomson et al., 2022). Some of the best evidence on EITC take-up across demographic groups comes from Jones (2014), who used CPS ASEC data linked with IRS data between 2005 and 2009 and estimated that approximately 78% of eligible White individuals took up the EITC in 2009, compared to 82% of Black individuals and only 72% of Hispanic individuals.9 Using self-reported data from the Survey of Income and Program Participation10 from 2014, as well as the National Bureau of Economic Research’s TAXSIM program,11 Thomson et al. (2022) also documented gaps in reported EITC receipt between Hispanic and non-Hispanic households.
For the CTC in 2021, Chiang et al. (2023), Collyer et al. (2023), and Parolin et al. (2021) all found lower rates of monthly CTC receipt in 2021 among Hispanic households compared to White non-Hispanic households. Using HPS data from 2021, Parolin et al. (2021) found that 67% of eligible White families, 70% of eligible Black families, and only 61% of eligible Hispanic families reported receiving the monthly credit. These findings are corroborated by two other studies that used HPS data to examine differences in take-up by race and ethnicity (Chiang et al., 2023; Jahnke & Williams, 2023; Karpman et al., 2021).
Differences in EITC and CTC take-up by Hispanic ethnicity might reflect several factors, including differential eligibility rates due to SSN requirements (see Chapter 2), lack of information or misinformation about
___________________
8The committee was also interested in reporting take-up rates according to geography, but there is not much evidence on this from either the EITC or the CTC literature, aside from official IRS reports of EITC take-up by state: https://www.eitc.irs.gov/eitc-central/participation-rate-by-state/eitc-participation-rate-by-states. This report indicates take-up rates range between 72.9% in Washington, DC, and 85.5% in Vermont in 2021, with no clear geographic pattern. Geographic variation in participation in the 2021 monthly CTC is discussed in Chapter 5.
9This gap was reversed in 2005, when 76% of non-Hispanic individuals claimed the EITC, compared to 82% of Hispanic individuals. Because Jones (2014) also limited analysis to individuals who can be linked to IRS data and are therefore more likely to have an SSN, her reported rate of take-up among Hispanic individuals is likely higher than survey-based estimates, which often cannot distinguish between those with and without SSNs.
eligibility, language barriers, and concerns about anti-immigrant sentiments and policies. These issues might be compounded by the differing eligibility criteria of the EITC and CTC, which may render some Hispanic families eligible for one credit but not the other. Using administrative records data on the Supplemental Nutrition Assistance Program (SNAP) in California, Fischer et al. (2022) estimated that take-up of the monthly CTC among eligible mixed-status families was approximately 67%, compared to nearly 80% among families in which both parents and children were likely to have SSNs.
Variation by Income and Education
Income and socioeconomic status are other important dimensions of EITC and CTC take-up. For the EITC, Jones (2014) estimated higher take-up rates among families with incomes on the plateau (81% take-up rate in 2009) or phase-out range (84% take-up rate in 2009) of the benefit schedule relative to those with incomes on the phase-in region (68% take-up rate in 2009). Families with incomes on the phase-in range of the EITC benefit schedule are likely not required to file tax returns, which may explain lower take-up rates. Similarly, Jones (2014) also documented lower EITC take-up rates among less-educated households: 72% of those with less than a high school diploma received the EITC compared to 82% of eligible households with some college.12
For the CTC under ARPA in 2021, several studies documented income gradients in credit take-up. For instance, Parolin et al. (2021) found that only 57% of eligible child-containing families with incomes below $25,000 reported receiving the first monthly CTC payment, compared to 73% of families with incomes between $75,000 and $100,000.13 Using SNAP data from California, Fischer et al. (2022) estimated that only 58% of households with no earnings likely received the monthly CTC, compared to 93% of households with at least $35,000 in annual earnings.14
Lower take-up rates of the monthly CTC payment option in 2021 among lower-income households might reflect differences in tax filing. Michelmore and Pilkauskas (2023) found that the single greatest predictor for receipt of monthly CTC payments was whether the household filed taxes
___________________
12Jones (2014) also found lower take-up rates among college-educated households (73%), but this is likely driven by lower eligibility rates among college-educated households compared to less-educated households.
13Based on a sample of low-income households, Michelmore and Pilkauskas (2023) similarly documented a steep income gradient in reported receipt, even among low-income households.
14The sample in Fischer et al. (2022) was restricted to households receiving SNAP, so the overall income of these households was low, similar to the population in Michelmore and Pilkauskas (2023).
in the prior year—a finding corroborated by other work demonstrating tax filing as a key barrier to tax credit receipt among low-income families (Goldin et al., 2022b).
Variation by Family Structure
Family structure is another demographic dimension that exhibits systematic variation in tax credit receipt. For the EITC, Jones (2014) documented much higher take-up rates among families with children (83–85%) compared to families without children (65%). This was likely driven in part by the fact that the childless EITC (i.e., the EITC for individuals without qualifying child dependents) has historically phased out at a very low income level, such that most individuals eligible for the childless EITC are not required to file tax returns. For tax filers with children, Jones (2014) found comparable EITC take-up rates among married filers (79%) and unmarried filers (78%) in 2009.
Less evidence exists on variation in CTC take-up in 2021 along family structure dimensions, but, using HPS data, Chiang et al. (2023) reported higher take-up rates among families with at least two children. They also found higher take-up rates among married filers relative to single filers (Chiang et al., 2023).
Relatively less is understood about EITC or CTC receipt among young adults who are parents (many of whom may be claimed as dependents on their parents’ tax returns) and families with complex custodial and related structures. Some evidence, however, indicates duplicate claiming, which could be due to confusion around shared custody and residential status of children (Leibel et al., 2020). Additionally, families who experienced the birth of a child in 2021 could not claim monthly advance CTC payments for those newborns; instead, these children had to be claimed in 2022, when those families filed taxes for TY 2021 (Congressional Research Service, 2022c).
Variation Across the Business Cycle
The committee is not aware of any data describing take-up across the business cycle. Bitler et al. (2023) showed that participation of married parents in SNAP increased when the unemployment rate rose, but they also found that being a tax filer with earned income in the range to be eligible for the EITC went up. Since participation in both SNAP and the EITC went up, and the paper did not estimate take-up, it is not possible to determine whether take-up increased. For single parents, while there are no statistically significant results, the point estimates for both SNAP participation and having filers in the eligible income range go in the same direction.
Key Finding 4-1: In 2021, the Internal Revenue Service reported that approximately 81% of those eligible for the Earned Income Tax Credit received the credit and 61.9 million families with children received at least one monthly Child Tax Credit (CTC) payment in the second half of the 2021 calendar year. While the number of CTC-eligible children in 2021 is unknown, U.S. Census Bureau estimates imply that CTC take-up was likely over 80%, though self-reported receipt rates were often lower than 80%. Reported CTC receipt also varied along several dimensions, although some may reflect eligibility differences. For the American Rescue Plan Act of 2021 CTC in 2021,
- White and Black families were more likely to report receiving the monthly CTC compared to Hispanic families.
- Families with higher earnings were more likely to report receiving the CTC relative to lower-income families and families with less education.
- One of the strongest predictors of whether families reported receiving the monthly CTC was whether they had filed taxes in the previous year.
Estimates of EITC take-up by these subgroups are not available for 2021; however, earlier work matching administrative data to survey data documented similar take-up patterns as seen in survey evidence for the CTC under ARPA, along racial, education, and income dimensions. Less concrete evidence exists on how take-up varied along geographic dimensions and dimensions of family complexity.
EXPLANATIONS FOR INCOMPLETE TAKE-UP
There are many plausible reasons why measured take-up of the EITC and CTC policies in 2021 may be less than 100% and why take-up might not reflect receipt among all those who are eligible. The sections that follow discuss several reasons for incomplete take-up, including challenges with accurately calculating take-up, including accurately measuring tax credit eligibility, issues surrounding fraud or noncompliance, and reasons related to the design and implementation of tax credits as delivered through the tax system.
Data from Linked CPS ASEC and IRS Tax Records
One reason tax credit take-up may be mismeasured—particularly in survey data—is the lack of adequate information on eligibility. The rules governing eligibility for both the EITC and CTC are complicated, involving
information on annual earnings, household living arrangements, and the relationships among individuals who live together. Furthermore, as discussed in Chapter 3, the CPS ASEC, a household survey commonly used by researchers to assess tax credit eligibility, measures income over the calendar year (which corresponds to a given tax year), but captures living arrangements in March of the following calendar year. Additionally, the CPS ASEC does not ask respondents to self-report receipt of credits, but instead relies on the CPS ASEC tax model, which employs a set of rules to determine tax filing units (TFUs) and uses federal tax rules to determine tax liability and credit eligibility. This model, like other tax calculators, assumes 100% take-up of credits among those deemed eligible. All these factors may lead to measurement issues in assessing tax credit take-up.
In recent years, there have been efforts to link survey data with administrative tax data, and several researchers have leveraged these linkages to assess the accuracy of tax models in determining tax credit eligibility.15 Findings from several such studies, described in detail in Appendix E, are briefly summarized here.16
One such study (Bee et al., 2023b) examined take-up of the ARPA-expanded CTC. While the study found that, in aggregate, the CPS ASEC tax model estimates corresponded closely to IRS control totals, the model significantly underpredicted IRS CTC control totals for TY 2021 for TFUs below and near the Supplemental Poverty Measure’s poverty thresholds. Furthermore, the study found that the CPS ASEC tax model tended to misallocate qualifying children across TFUs compared with information TFUs reported in their federal tax returns.
The committee’s summary of the tabulations in Bee and Unrath (2025) for TY 2018 also revealed shortcomings in the CPS ASEC tax model. In particular, compared to linked tax returns, the CPS ASEC tax model tended to overpredict the incidence of those who did not file tax returns, especially for those with low earnings (less than $25,000) and to underpredict the incidence of tax filers who file as single heads of households with dependents. In addition, the CPS ASEC tax model tended to slightly underpredict the number of dependents under the age of 18 in CPS ASEC TFUs compared to data from their linked tax returns, with this underprediction driven by
___________________
15See, for example, studies by Bee et al. (2023a,b, 2025), Bee and Unrath (2025), Coleman et al. (2024a,b), Jones and Ziliak (2022), Meyer et al. (2020, 2022), Meyer and Mittag (2019), and Unrath (2022, 2024) using CPS ASEC–IRS data to assess income reporting and poverty status, as well as estimated EITC and CTC take-up for various tax years.
16The committee is unaware of studies that can directly determine the tax credit receipt from linked data, as IRS tax return data elements that would determine credits received are not linked to the CPS ASEC, under agreements between the Census Bureau and the IRS governing elements of IRS tax records included in linked data made available to Census Bureau researchers and the research projects that have approved access to them.
sizeable underprediction of dependents among Hispanic people, non-Hispanic Black people, and Asian people/other race groups. The latter findings are consistent with the evidence in Bee et al. (2023a) demonstrating that a sizeable number of children claimed on tax returns in TFUs formed by the CPS ASEC tax model were not present in CPS ASEC households at the time this survey was conducted (spring 2019). Finally, the tabulations produced by Bee and Unrath (2025) indicated that the CPS ASEC tax model tended to underpredict CTC eligibility for TFUs with low earnings (less than $25,000) and for minorities (i.e., Hispanic people, non-Hispanic Black people, and Asian people/other racial groups).
Taken together, the results of Bee et al. (2023a) and tabulations from Bee and Unrath (2025) suggest that model-based estimates—in this case the CPS ASEC tax model—that rely solely on CPS ASEC survey data for capturing the formation of TFUs and predicting CTC eligibility have some notable shortcomings, especially compared to data from linked tax returns and IRS control totals. While the model used to produce estimates of the impacts of the EITC and CTC policies in 2021 on child poverty (Chapter 8) and of the potential impacts of a range of policy options on child poverty (Chapter 9) rely on a different model—namely the Transfer Income Model version 3 (TRIM3)—the findings from these previous studies suggest potential limitations of existing models based solely on survey data.
In addition, the findings from Bee et al. (2023b) and the Bee and Unrath (2025) tabulations suggest that TRIM3 and other tax models likely underestimate the number of CPS ASEC household members filing tax returns that claim either the EITC or CTC, possibly due to some members claiming children not in the CPS ASEC household at the time of the survey (in the spring following a given tax year). However, the committee finds the evidence in these and other studies far from conclusive on this matter. Furthermore, other studies using linked data (see footnote 15) indicate that some tax filers may misreport their incomes to the IRS, and this misreporting may indicate that some of these tax filers may be ineligible to receive the EITC or CTC. TRIM3 adjusts for TFU income in CPS ASEC data, primarily for cash benefits received from certain safety net programs (see Appendix H). As a result, TRIM3 tends to produce reduced estimates of the number of tax filers claiming children under either credit compared to IRS aggregates.
These findings from linked survey–IRS data clearly suggest that tax models based solely on survey data, such as the CPS ASEC, mismeasure both the eligibility and take-up of both the EITC and CTC. But, the committee concluded that they did not provide clear guidance on how to better align the TRIM3 or other tax models for estimating credit take-up and the impact of these credits on child poverty. This issue is revisited in Chapter 10, which presents the committee’s recommendations for further research
to improve existing models used in analyzing tax credits and their impacts on poverty.
Noncompliance in Receipt of the EITC and CTC
In calculating the share of individuals who receive credits, take-up represents the number of individuals who report receiving benefits among those who appear to qualify for such benefits. This interpretation is complicated by “tax noncompliance” (i.e., nonpayment of taxes, or in this case claiming credits one is not eligible for). The IRS, which administers tax credits, has diverse and competing goals. The IRS aims to count credits toward taxes owed or pay out refunds to eligible claimants, while also flagging instances of noncompliance.17
In the context of the EITC and CTC, noncompliance affects the interpretation of take-up in two primary ways: if some individuals who report receiving the benefit (the numerator in the take-up calculation) were not eligible to receive them, or if some individuals who appear eligible for the benefit (the denominator in the take-up calculation) were not eligible to receive benefits.
Estimates of noncompliance with the CTC policy in 2021 have yet to be generated. However, a 2016 U.S. Government Accountability Office (GAO) report states that the pre-ARPA CTC and Additional CTC (ACTC) generated a total of $6.4 billion per year in overclaiming from 2009 to 2011, relative to $18.1 billion per year in overclaiming in the EITC. Additionally, the U.S. Treasury’s Inspector General for Tax Administration (2022) issued statistics describing the accuracy of the advance part of the CTC during 2021. This report pointed out that approximately 98% of payments were correctly sent out, but around 2% or 3.3 million payments totaling $1.1 billion were incorrectly sent, and 8.3 million payments worth $3.7 billion did not make it to eligible taxpayers—a participation rate of about 95% among those with IRS information suggesting they were eligible for the monthly payments.
Relative to previous tax years, 2021 had an additional twist: advance CTC payments were sent to the parent with the most recent tax return filed for the child in 2020 or, failing that, in 2019.18 By relying on potentially
___________________
17In the IRS’s understanding taxes glossary (https://apps.irs.gov/app/understandingTaxes/student/glossary.jsp), voluntary compliance is defined as when “individual citizens report their income freely and voluntarily, calculate their tax liability correctly, and file a tax return on time.” Noncompliance thus is incorrect reporting of income and tax liability (or could be filing late, but this is not discussed), and it can arise both unintentionally and on purpose.
18https://www.irs.gov/credits-deductions/2021-child-tax-credit-and-advance-child-tax-credit-payments-topic-l-commonly-asked-shared-custody-questions, https://www.childrensdefense.org/blog/understanding-the-expanded-ctc/
outdated information regarding children’s living arrangements, this feature might have led to higher rates of noncompliance in 2021. Karpman and Maag (2022) found that 15% of adults reported that someone outside the household claimed the CTC issued for children who lived with them. This type of claim may not be illicit. For instance, if the child lives with each parent half of the year, the parent with the higher adjusted gross income is instructed to claim the child. Because most survey data do not report how long children have lived in their household, it is difficult to verify whether a child lived with a particular parent for half the year or more.
The CTC Policy in 2021 also had a safe harbor provision for tax filers with modified adjusted gross income below $100,000 ($120,000 if married filing jointly): if the TFU qualified for a smaller advance CTC payment than they received when reconciling taxes at tax filing time, no penalty would ensue if the difference between the amount received and eligibility amount was sufficiently low. The extent of repayment protection depended on adjusted gross income, with TFUs with lower incomes receiving more protection from repayment than those with higher incomes. This could have led to lower levels of noncompliance, particularly in 2021.
From the EITC literature, Leibel et al. (2020) looked at IRS audits to explore mis-claiming according to the relationship between qualifying children and the adults claiming the credits. They found that, in many cases of erroneous EITC claims, the children were likely to still be eligible for the EITC if claimed by the adult(s) who met the criteria to claim them as qualifying children. Therefore, published reports of EITC noncompliance might be overstated. Leibel et al (2020) “find evidence of complicated family circumstances, intentional credit-maximizing behavior, and other possible explanations for EITC non-compliance” (p. 3). Relatedly, Tong (2014) found evidence that, across adjacent years, a significant share of children is claimed by different claimants, suggesting that the practice of sending the credit to the most recent taxpayer claiming the child might introduce errors.
Key Finding 4-2: Estimates of the rates of noncompliance for the American Rescue Plan Act of 2021 (ARPA) Earned Income Tax Credit (EITC) and Child Tax Credit (CTC) in 2021 do not exist. Prior reports from the U.S. Government Accountability Office based on data from 2009 to 2011 found overpayments of approximately $6.4 billion for the CTC and Additional CTC, and $18.1 billion for the EITC. In terms of the monthly CTC payments in 2021 under ARPA, the Internal Revenue Service reported that approximately $1.1 billion in incorrect payments were sent, while $3.7 billion in benefits did not reach intended recipients, representing an accuracy rate of approximately 95%.
Conclusion 4-1: Incorrect payments for both the American Rescue Plan Act Earned Income Tax Credit and Child Tax Credit advance month payments in 2021 were often due to challenges related to qualifying child rules as well as complex family and child custody arrangements. Prior research suggests that a significant share of children is claimed by different tax filers across tax years, which could create challenges for identifying the correct recipient of monthly benefits.
Effects of CTC in 2021 Design Under ARPA and Implementation on Take-Up
Design and implementation issues may affect take-up and receipt rates of tax credits, including specific features of the CTC policy under ARPA in 2021 that acted as barriers among eligible families. For instance, lack of internet access and lack of connection to tax filing are also associated with populations described above (e.g., different racial groups, low-income households, etc.) and so may have differential impacts across groups (see Chapter 5).
Some factors that may have prevented families from receiving tax credits may have resulted from the unique, monthly distribution of the CTC in 2021 under ARPA, while other factors were more broadly applicable. For instance, there was only a four-month period between ARPA’s expansion of the CTC and the first disbursal in July 2021. Prior to ARPA, the IRS had little experience disbursing monthly payments to tax filers, as tax refunds are typically sent once a year.19 Some of the implementation issues discussed in subsequent sections may be less applicable when agencies have more time to set up disbursements, or if agencies have the capacity and institutional knowledge to deliver monthly benefits (e.g., the Social Security Administration).
Program Awareness, Knowledge, and Informational Barriers/Complexity
Take-up of tax credits may be incomplete due to informational barriers including lack of eligibility information. A 2015 study showed that 44% of EITC-eligible respondents were unaware of the tax credit and a majority of those who believed they were eligible underestimated the amount for which they were eligible. For example, simplification and salience of benefits communicated in mailings to tax filers who were eligible to receive the credit increased take-up. Those lacking internet access or fluidity with
___________________
19Historically, the EITC had an option for more frequent distributions through paychecks, but this was discontinued in 2010 because very few taxpayers chose this distribution method.
online platforms and related technology faced greater challenges navigating the tax system (Bhargava & Manoli, 2015).
Hispanic families were less likely to be aware of the EITC (Maag, 2005; Ross Phillips, 2001). Immigrant parents were less likely to know about the CTC than nonimmigrant parents (Acevedo-Garcia et al., 2022; Maag & Brugger, 2021). Awareness challenges were magnified by the need for literacy and English language proficiency. A Spanish translation of the CTC in 2021 under ARPA was slow to become available, which may have contributed to reduced awareness of eligibility and receipt (Augustine et al., 2021; Rodriguez Accounting & Consulting, 2021). Employers can influence information dispersal, but the evidence so far is mixed. For instance, Cranor et al. (2019) found that the EITC employer notification mandate, as currently implemented, has a small impact on EITC take-up. Other work, however, suggested that employer mandates might be more effective for certain groups, such as Hispanic families (Thomson et al., 2022). (For more information on program awareness, see Chapter 5.)
IRS Capacity
The role and capacity of the administering government agency can also affect take-up and people’s experiences. A report by the National Taxpayer Advocate on the rollout of the CTC in 2021 under ARPA found that the IRS had insufficient staff, funding, and technology to manage the influx of new filers, respond to queries, and process and validate returns including CTC advance payments (Taxpayer Advocate Service, 2022).
Additionally, a GAO report (2022a) suggested that the IRS was overwhelmed in TY 2021, partly due to difficulties with paying out refundable tax credits and partly due to office closures due to the COVID-19 pandemic (hereafter, “pandemic”), which delayed processing of 2020 returns. In December 2021, nearly 11 million returns remained to be processed, well more than in typical years. Out of the 150 million returns processed, 35 million were found to have errors, and many refunds were delayed.
In sum, several factors likely contributed to incomplete take-up, including mismeasurement of eligibility, noncompliance, information barriers, and IRS capacity constraints.
Efforts to Increase Take-Up
Several efforts have been made to increase take-up of tax credits, particularly in response to some of the provisions of the CTC enacted under ARPA. Code for America designed a website and phone application (GetCTC.org) to help households claim their CTC benefits, with a particular focus on households with no prior tax filing history. Nearly 200,000
households used these resources to claim approximately $700 million in benefits. The IRS also launched an online portal through which households could update their information to determine eligibility for the monthly advance CTC. Those who had not previously filed taxes in 2019 or 2020 could use the portal to report their information and claim the monthly benefits.
Direct portals launched by the IRS also increased access to the CTC in 2021.20 The first findings of the direct portal pilots in 12 states (CA, TX, FL, NY, WA, MA, TN, AZ, NV, NH, SD, WY) concluded that more than $90 million in tax refunds were issued, with $35 million in tax balances collected.21
A coalition of privately funded nonprofit and local community-based organizations (called the COOP fund) mobilized to spread awareness of the CTC in 2021 and to connect eligible households to the resources and assistance needed to claim the credit. A coalition of nonprofits in Detroit launched broad outreach and engagement campaigns (The Kresge Foundation, 2023). This endeavor especially focused on new filers who had tax filing hesitancy (e.g., fear, distrust, or lack of familiarity); misconceptions about the risk of losing other benefits; or confusion about qualifications and/or purpose, and whether benefits would ultimately need to be paid back. Analyses suggested that COOP helped secure approximately $75 million in tax refunds for low-income families.
COOP results indicated that leveraging highly effective forms of outreach, including health systems, school systems, other community-trusted state and local partners, peer-to-peer influences, and private-sector employers could facilitate knowledge and reduce information-based barriers around the EITC and CTC. (See Chapter 7 for further discussion.) However, a large-scale study of six types of personalized, light-touch interventions among EITC-eligible populations in California showed no detectable impact (Linos et al., 2022). However, community-based approaches demonstrated positive effects on filing and receipt of the EITC, especially among immigrant populations. Studies in Houston and Las Vegas (Bernstein et al., 2020), two cities with histories of immigration but varying in infrastructure and policy context for serving immigrants, suggested that messaging and information about immigration rules dealing with public benefits and their
___________________
20https://www.irs.gov/pub/irs-pdf/p5969.pdf
21In a survey of 15,000 Direct File users (13% response rate, primarily completed in English), 90% ranked their experience as excellent or above average, due to ease, trustworthiness, and being free to use. The average time to file was approximately 30 minutes. The U.S. Department of the Treasury estimated that Direct File users saved $5.6 million in tax preparation fees and filers benefited from quick feedback regarding errors during the filing process, as opposed to waiting for confirmation of whether their returns were accepted or rejected.
interactions with tax credits can be effective strategies (Iraheta & Morey, 2023).
Key Finding 4-3: Several design, delivery, and outreach strategies were deployed to increase information about and take-up of the advance monthly payments option for the Child Tax Credit (CTC) in 2021: the Internal Revenue Service launched an online portal allowing families to update their information and claim monthly benefits; the GetCTC.org website launched by Code for America was also designed to help families receive monthly benefits. Many community-based organizations held outreach campaigns targeted at families who may not have been familiar with the CTC, providing details and resources for claiming the credit.
CONCLUSION
Take-up—the extent to which eligible individuals or families actually receive a benefit—plays a critical role in the evaluation of the EITC and CTC policies in 2021, as well as in the assessment of the credits’ impacts on child poverty. Efforts to increase participation in 2021 included the launch of IRS and nonprofit online portals and targeted community-outreach campaigns. Administrative IRS data suggest that, in 2021, approximately 80% of eligible families received the EITC, and a similarly high share of children benefited from the CTC in 2021. However, survey-based estimates indicated lower take-up rates, particularly among Hispanic families, lower-income households, and families with limited education. Several factors contribute to incomplete take-up, including measurement challenges in survey data, administrative barriers to claiming benefits, informational barriers (especially among immigrant and non-English speaking households), and IRS capacity constraints during the pandemic. These limitations complicate interpretation of the credits’ effects on poverty, particularly because modeling efforts often assume full participation among those eligible.
Conclusion 4-2: Incomplete take-up affects the interpretation of the impact of the Earned Income Tax Credit and Child Tax Credit in 2021 on child poverty in two key ways. First, simulations assume 100% take-up when calculating impacts. As a result, antipoverty effects of these credits may be overestimated if some eligible individuals did not receive them. Conversely, simulations do not account for noncompliance, whereby ineligible individuals receive credits.
Although concrete noncompliance figures for TY 2021 are unavailable, prior work by Jones and Ziliak (2022) found that these two measurement
issues offset one another—at least in the case of the EITC—resulting in similar estimates of antipoverty effects (see Chapter 8 and Appendix E). However, careful interpretation of simulation results is warranted, and continued improvements in measuring eligibility and receipt are critical.
Next, Chapter 5 shifts to perspectives from families and nonprofit organizations that helped facilitate access to these benefits. Drawing on listening sessions and qualitative research, that chapter highlights how families experienced the CTC, how they navigated barriers to participation, and how nonprofits played a key role in outreach and support. These insights complement quantitative findings by revealing how policy implementation was perceived on the ground and by identifying opportunities for improving program delivery.
