10

Recommendations for Improvements in the Measurement and Evaluation of the Impacts of the EITC and CTC on Child Poverty

The committee addressed the four elements of its charge:

1. What are the impacts of the federal Earned Income Tax Credit (EITC) and Child Tax Credit (CTC) in 2021 on the level of poverty for all U.S. children and the level of poverty for specific populations of U.S. children?
2. How was the CTC implemented in 2021? How did the implementation of the program impact participation and therefore its effectiveness for reducing child poverty?
3. Among children in different racial and ethnic origin groups, immigrant families, and other populations of interest (e.g., children in urban and rural areas), how did the implementation of the CTC in 2021 facilitate or reduce program access and therefore its effectiveness for reducing child poverty?
4. What changes to the tax rules and requirements and the procedures for administering the EITC and CTC, if adopted, would further reduce the number of U.S. children in poverty?

In addressing its charge, the committee identified several data and methodological limitations, as well as gaps in the existing research literature, that affected its ability to fully address elements of the statement of task. This chapter summarizes major data and research limitations identified and recommends research and data improvements that would enhance the ability of researchers and policy makers to evaluate and design future EITC and CTC policies to reduce child poverty.

These recommendations are grouped into four areas:

1. Improvements to data sources and measurement of key issues;
2. Research priorities to enhance the monitoring and evaluation of the impact of the EITC and CTC on child poverty and child well-being;
3. Strategies for experimenting with and evaluating changes made to the EITC and CTC to reduce child poverty; and
4. Research on information available about the tax system and the claiming of tax credits, access to the tax credits, and outreach efforts to improve both.

In several instances, the committee recommends methodological approaches to support this work, especially for measuring child poverty and evaluating the implementation and effectiveness of the 2021 reforms under the American Rescue Plan Act of 2021 (ARPA).

IMPROVING DATA SOURCES AND MEASUREMENT FOR EVALUATING AND MONITORING TAX CREDITS AND THEIR IMPACTS ON CHILD POVERTY

Several efforts could improve data sources and the measurement of key phenomena related to the evaluation and monitoring of the effects of the EITC and CTC on child poverty.

Enhance Quality of and Access to Linked Survey and Administrative Data for Analyzing Tax Credits, Income, and Poverty

The committee relied upon Transfer Income Model version 3 (TRIM3), and its use of Current Population Survey Annual Social and Economic Supplement (CPS ASEC) data, for its assessments of the impacts of the 2021 ARPA reforms to the EITC and CTC and its evaluation of potential future changes to these tax credits that might further reduce child poverty relative to current policy. As discussed in Chapter 8 and Appendix H, TRIM3 uses administrative data from many safety net programs to adjust CPS ASEC data based on participation in and benefits received from those programs. However, as noted in Chapter 4, the CPS ASEC does not ask respondents whether they filed a tax return or received either EITC or CTC benefits. As a result, TRIM3 and other tax models assume all households in the CPS ASEC sample that appear eligible for either tax credit file a tax return and receive those credits. These models do not attempt to make adjustments based on administrative data reported by the Internal Revenue Service (IRS) (see Chapter 4 and Appendixes F and I).

Chapter 4 and Appendix E report evidence from studies that used data linking IRS tax return information to CPS ASEC respondents. Such access is limited to U.S. Census Bureau and IRS researchers and also to certain researchers outside these organizations.¹ While the use of linked survey and IRS data for studying take-up and impacts of the EITC and CTC is in its early stages, the committee views linked data as important to provide information on tax return filing (especially by low-income families), families’ eligibility for and receipt of the EITC and CTC, and accuracy of income data and thus poverty status. Accordingly, the committee offers several recommendations concerning the further development of such linked data and its access by researchers in government, universities, and other research settings.

Several areas could benefit from these efforts. First, it is important to use linked data to identify issues of incomplete take-up among tax filing units eligible for the EITC and CTC, as well as the underreporting of benefit receipt in surveys. Second, linked data could be used to identify mismatches in the composition of tax filing units assigned by the Census Bureau’s tax model and those identified using IRS tax return data, including identifying qualifying children. Third, it is critical for data-linkage efforts to produce and regularly disseminate publicly available data products that support research and monitoring of EITC and CTC claiming among all U.S. households—especially by the various demographic and income groups considered in this report.

Finally, it is important for a broad group of researchers to have access to these linked data. With access, researchers could address several important research questions intrinsic to understanding the measurement of income and poverty for various groups, along with other issues critical for assessing the impacts of the EITC and CTC in 2021 on child poverty status and additional measures of child well-being in the United States.

___________________

¹Researchers must obtain special permission and agree to submit their findings to a review designed to ensure that released statistics do not reveal respondents’ identities or otherwise compromise confidentiality.

Some important research using linked data has already been performed, including work undertaken by the University of Chicago’s Comprehensive Income Data Project. Many other key research priorities could be advanced through improved access to linked data, and the following recommendation addresses steps the Census Bureau could take to accomplish this goal.

Improve Poverty Measurement

The committee concluded that the Supplemental Poverty Measure (SPM) is the most suitable tool currently available for evaluating the impacts of the EITC and CTC in 2021 on child poverty (see Key Finding 3-1 in Chapter 3). However, as also noted in Chapter 3, the SPM suffers from data limitations that hinder the committee’s ability to answer the questions posed in the statement of task. Further, the accommodations the committee undertook to partly address data-quality issues are not fully available to the public.

A recent National Academies of Sciences, Engineering, and Medicine (National Academies) report, An Updated Measure of Poverty: (Re)Drawing the Line (National Academies, 2023), provided a thorough assessment of improvements needed to the SPM for measuring poverty in

the United States.² A second National Academies report, A Roadmap to Reducing Child Poverty (2019), also considered the strengths and weaknesses of using the SPM to measure child poverty. Both reports, particularly the 2023 report, offered several recommendations for improving the SPM, many of which would, if implemented, address issues that this committee encountered in its work.

Improve Data to Identify Key Subgroups in Existing Survey and Administrative Data

Accurately evaluating the impact of tax credits such as the EITC and CTC on child poverty requires data systems that capture the full diversity of family arrangements, legal statuses, and demographic characteristics. The committee found that existing survey and administrative datasets have important limitations, including underreporting of noncustodial parents, misalignment between tax and household units, and missing information on citizenship status, which is crucial for assessing tax credit eligibility. These gaps are especially problematic for immigrant and mixed-status families, as well as for children supported by multiple caregivers. Moreover, small sample sizes for certain subgroups, including rural families, children with disabilities, and racial or ethnic minorities, limit the statistical power to assess program impacts equitably.

___________________

²This report built on an earlier report, Measuring Poverty: A New Approach (National Research Council, 1995) that laid the groundwork for the SPM.

Capturing Dynamic Household and Family Structure

A growing number of children have parents who live apart. These children usually spend time with and receive resources from both parents. The CPS ASEC only captures data on the children living in a household at the time of the survey; it does not collect data on which children lived in the household during the previous year or how long they lived in the household. Such information is necessary to determine which children might qualify families for either the EITC or CTC, and which parent is eligible to claim the credit. Without this information, assumptions must be made regarding which children lived in a household in the tax year. Evidence from linked CPS ASEC and IRS tax records indicates that some tax filers in CPS ASEC households do not report any children in the household at the time of the survey, yet they claim dependents on their federal tax returns.³ Whether these are cases of tax noncompliance (e.g., a noncustodial parent claiming children who do not satisfy the requirement of living with that parent for six or more months in a given tax year) or legitimate claiming of a child who did live with a parent during the tax year but was not doing so at the time of the survey is almost impossible to determine with current information on household composition (see Chapter 4).

This problem has no simple solution. However, such data are collected in the Survey of Income and Program Participation (SIPP), which may offer a potential resource for addressing it. It would be useful to explore the reliability of SIPP data on household composition dynamics, especially for low-income families, and the cost of adapting SIPP’s strategy for inclusion in the CPS ASEC. The committee also encourages research on other strategies to improve understanding of complex household dynamics.

Increasing Sample Size of Underrepresented Subgroups

To produce more reliable subgroup estimates—particularly in key datasets like the CPS ASEC—it is important for federal agencies to adopt sampling strategies that oversample small or underserved populations, such as children in immigrant families, children with disabilities, families in rural areas, and families with complex caregiving arrangements. Efforts could include geographic oversampling, improved stratification, or linking surveys with administrative records. Expanding representation of these subgroups is essential for evaluating equity in tax credit receipt and accurately measuring impacts on child poverty.

___________________

³See Bee et al. (2023b) and the discussion of findings for tabulations from Bee and Unrath (2025) in Appendix E.

ENHANCING THE MONITORING AND EVALUATION OF THE IMPACT OF THE EITC AND CTC ON CHILD POVERTY AND WELL-BEING

To guide future tax credit policy, it is important for evaluations to assess broader outcomes for children and families, beyond poverty rates.

Chapter 8 emphasized the importance of understanding the impacts of heterogeneity across family types, subgroups, and over time—highlighting the need for research on long-term effects, family dynamics, and labor market interactions. Evaluations focused on the short- and long-term benefits and costs of tax credits, and on how credits impact child poverty and family well-being, are warranted. Several outcomes are important to evaluate, including fiscal costs and benefits; effects on adult education, earnings, and employment; and time investments in children and children’s education, health, quality of parenting, family stability, and the interaction of job types and work hours with parenting.

Increasingly complex and dynamic family structures require more flexible and comprehensive data collection to accurately assess child and family well-being. This issue is particularly salient for noncustodial parents, especially fathers. These parents currently qualify for the childless EITC, the maximum value of which was only $632 in 2024. Several proposals have been made to expand the childless EITC, to raise employment rates among noncustodial parents (among other childless adults), and perhaps improve their abilities to provide for their children (Holzer, 2023). In 2021, ARPA temporarily raised the maximum value of the childless EITC to over $1,500, though its effects on noncustodial parents and their children have not been evaluated. Other efforts to evaluate the effects of expanding the childless EITC, and especially their impacts on children, have been limited in scope and have not yet led to conclusive results (Miller et al., 2022).

Improve Understanding of Tax Credit Impact on Employment and Consequences for Child Poverty

The committee’s review of existing literature on the impacts of tax credits on employment, discussed in Chapter 7, revealed several knowledge gaps, providing several areas for future research. Evidence on how the EITC and CTC influence employment decisions is substantial, but uneven. While there is general agreement that these credits influence whether individuals work, considerable uncertainty exists regarding the magnitude of these effects and how they vary by demographic subgroup, economic conditions, and policy designs.

Existing literature reveals several additional areas of uncertainty. First, existing studies do not clearly establish whether employment responses to

increased work incentives are symmetric with employment responses to decreases. This has implications for how parental labor supply might respond to changes in credit generosity or the introduction or removal of work requirements. Second, most studies focus on lower-income households, leaving little empirical evidence on how higher-income families—those eligible for the CTC but not the EITC—respond to changes in labor market incentives. Third, while research has documented a decline in labor supply elasticities for married women over time, it remains unclear whether similar trends apply to unmarried mothers, who are often the primary beneficiaries of these credits. Fourth, it remains unclear whether elasticities vary across the business cycle.

To strengthen the evidence base and better inform policy design, the committee recommends the following:

Develop Expanded Employment Elasticities

Current literature shows wide variation in employment elasticities. While estimates of the employment effects associated with expansions to the CTC are relatively consistent across studies, the EITC literature shows a much wider range of labor supply elasticities. Further research is warranted to reconcile these discrepancies, namely by exploring the extent to which they arise from methodological differences, population characteristics, sampling frames, or policy design. It would be advantageous for new analyses to produce elasticity estimates across a broader range of demographic groups (e.g., by income, marital status, gender) and credit types (e.g., refundable versus nonrefundable, work-conditioned versus unconditional).

Improve Employment Elasticity Estimates

Current evidence on the CTC’s employment effects—other than those produced using the temporary changes in the CTC under ARPA—remains sparse and methodologically inconsistent. No consensus has been reached on best practices for evaluating CTC reforms, and the limited pre-ARPA research constrains the understanding of long-term employment responses. More research is needed on the impact of the CTC on employment, generated from policy changes other than those associated with the changes to the CTC in 2021 under ARPA. While several studies were published in the years after the expansion of the CTC in 2021, until that point very little research existed on the employment effects of the CTC aside from three working papers cited above. More research on the employment effects of the CTC is important for several reasons. First, evidence on how labor supply elasticities have changed over time is inconclusive; evaluation of more recent policy changes to the CTC could shed light on this question. Second, the CTC affects a different population than the EITC, so more research on the labor supply effects of the CTC would provide insights into how employment elasticities differ across these populations. Finally, due to the dearth of evidence on the employment effects of the CTC, consensus is lacking regarding the best and most rigorous ways to evaluate the effect of the CTC on employment.

Investigate Employment Effects Across Economic Cycles

The effect of the EITC on work incentives is an important issue when estimating the effects of that credit on poverty. However, the magnitude of such effects is likely to vary over the business cycle. In strong economic periods, jobs are relatively abundant, and individuals have greater ability to take jobs and thus benefit from work-incentivizing tax credits—or to forego employment to take advantage of tax credits available for nonworkers. In contrast, during recessions, jobs opportunities are more limited, reducing individuals’ ability to respond to work incentives. As a result, more individuals may be out of work due to the lack of available jobs, rather than due to the design of tax credits for nonworkers. Evidence in the research literature is limited and uneven when it comes to estimating work incentive effects during the COVID-19 recession. Congress often strengthens safety net protections in recessions, as they did in 2020–2021 and will likely do in future recessions. More evidence on how work incentives function over the business cycle would enable more accurate work-incentive estimates in the future.

ADDITIONAL STRATEGIES FOR EVALUATING THE IMPACTS OF CHANGES TO THE EITC AND CTC ON CHILD POVERTY

This report’s analysis of the impacts of the EITC and CTC in 2021 on child poverty, as well as estimates of the potential impacts of the committee’s alternative policy options on child poverty, relied on microsimulation methods with observational data. While this approach has many advantages, its validity for estimating the causal effects of such policy changes on child poverty relies on difficult-to-substantiate assumptions about the methods and data used. This is especially true when evaluating the impacts of potential future changes to features of these tax credits, as was undertaken in Chapter 9. Ideally, stronger causal evidence is needed to assess the real-world impacts of such reforms. Experimental designs that exhibit random assignment of alternative tax credit “treatments” and include a control group for which current policy is maintained represent the ideal standard for obtaining causal estimates for policy options. It is well known that such an evaluation design is hard to implement and suffers from other potential limitations. As a result, other designs, often referred to as quasi-experimental designs, have been used to evaluate social policies, including the impacts of programs designed to reduce poverty and improve wellbeing. These designs include using state and local policy differences as a strategy for identifying the impacts of such programs.

While mindful of the challenges of implementing randomized controlled trials (RCTs) to evaluate policy options, the committee nonetheless encourages relevant groups to evaluate alternative configurations of the EITC and CTC, such as changes in benefit generosity or eligibility requirements, using RCTs.

Such RCTs have the potential to provide stronger, more valid evidence on the potential impacts of changes in the EITC and CTC on child poverty and the well-being of children and their families.

A related strategy would be to further study state-level EITC and CTC programs that differ in their provisions:

It is important for these studies to include policies and programs that extend eligibility to tax filers with Individual Taxpayer Identification Numbers (ITINs). States have been experimenting with creating their own CTCs and improving their EITCs—for example, to allow families with ITINs and Social Security Numbers to claim the state EITC. It will be important for studies to assess the impact of these reforms on take-up rates, poverty, and family outcomes.

OUTREACH EFFORTS TO IMPROVE RESEARCH ON INFORMATION ABOUT AND ACCESS TO THE TAX SYSTEM

The committee’s analysis identified that the effectiveness of the EITC and CTC is shaped not only by the structure of the policies but also by how these credits are communicated, delivered, and experienced by families. The committee’s review found that take-up of the EITC and CTC remains incomplete, especially among marginalized groups such as immigrant families, gig workers, and those in rural communities. Misinformation, fear of a repayment requirement, and confusion about eligibility deterred some eligible families from participating. Listening sessions and qualitative data presented in Chapter 5 further revealed the critical role of trusted nonprofit organizations in helping families access credits and highlighted a lack of clear, official communication about program changes in 2021. Additionally, the ARPA CTC Expansion introduced monthly payments, which some families found more helpful than lump-sum disbursements—though preferences varied.

To better inform future policy design and ensure equitable access to tax credits, the committee recommends a coordinated research agenda to examine how payment structures, administrative burdens, outreach capacity, and communication strategies affect families’ ability to claim and benefit from the EITC and CTC.

Evidence presented in Chapter 5 shows that some families preferred predictable monthly payments, while others valued lump-sum refunds for major expenses. Thus, it would be useful for future research to compare monthly versus lump-sum disbursements and their effects on financial stability, savings, and financial decision making. Understanding how payment timing shapes outcomes may also inform the design of benefit delivery systems. To guide future design choices, research is also needed to identify which subgroups benefit most from each structure.

Chapters 4 and 5 highlighted the importance of trusted messengers and showed that many families first learned about the expansion of the CTC in 2021 from social media or nonprofit partners. Understanding which strategies increase trust and take-up is essential for effective outreach.

Chapter 4 and the listening sessions described in Chapter 5 revealed that the fear of having to repay benefits and worries about impacts on other benefits deterred participation in the CTC in 2021. Understanding how fears of administrative errors or cross-program eligibility influence take-up would offer valuable insights into program design and administration. It is important for research to examine how these perceptions vary by demographic groups and how program design can mitigate these barriers.

In terms of accessibility and barriers to credit receipt, parents with

irregular earnings face unique concerns and additional challenges when filing taxes or understanding how the CTC would affect their tax liability, as discussed in Chapters 4 and 5. These issues require focused attention to ensure that program structures are inclusive. It is important for research to examine how current tax structures serve or exclude nontraditional earners and what adjustments could improve accessibility and participation for these groups.

As seen in Chapter 5, nonprofits played a pivotal role in CTC awareness and access, especially for hard-to-reach populations. It is important for research to assess which outreach strategies are most effective and to document capacity constraints faced by long-standing community organizations. This analysis could include exploring the reach of chat-a-thons and peer education in rural areas. It would be useful for such investigations to include long-standing community action agencies involved in such outreach efforts—for example, community action councils founded under the Economic Opportunity Act of 1964. In assessing these organizations, it will be important to focus on collecting comparable information on metrics including services delivered, financial capacity constraints, agency locations, and whether information provision varies across regions.