The Earned Income Tax Credit, Family Complexity, and Children’s Living Arrangements
=====================================================================================

* Katherine M. Michelmore
* Natasha V. Pilkauskas

## Abstract

Demographic shifts over the last half-century have resulted in dramatic changes in family structure. These changes have implications for the social safety net because public assistance programs define families differently. This article focuses on a critical poverty-alleviation policy, the Earned Income Tax Credit (EITC), to document family complexity in the United States. We find that more than 60 percent of children in lower-income families reside in households with ambiguity in tax filing and thus in claiming valuable credits. Tax filing ambiguity driven by family complexity is especially common among households with Black children, highlighting significant racial inequities in the tax treatment of complex families. We also consider two reforms to reflect the realities of families today: the childless EITC and a noncustodial parent EITC.

*   EITC
*   family structure
*   family complexity

Demographic shifts over the last few decades have changed the nature of children’s living arrangements, especially among those whose parents have low incomes (see, for example, Cherlin 2010). A large literature has documented the increasing diversity in children’s living arrangements (Carlson and Meyer 2014), such as living with unmarried or cohabiting parents (Manning and Stykes 2015); step families, blended families, or social fathers (Furstenberg 2014); or complexity arising from multiple partner fertility (Monte 2019). Beyond parents and their partners, children are also increasingly likely to live with other extended family members (Pilkauskas and Cross 2018). In 2018, more than 15 percent of children lived with adults who were not their parents or their parent’s partners (Harvey, Dunifon, and Pilkauskas 2021). Complexity in children’s living arrangements is also closely linked with socioeconomic status and race-ethnicity: children from low-income and non-White households experience greater complexity than their more affluent peers (Harvey, Dunifon, and Pilkauskas 2021; Cross 2018). The recent COVID-19 pandemic may further increase the share of children in complex living arrangements driven by increased rates of parental death (Kidman et al. 2021) and economic instability.

These demographic shifts have occurred amid critical shifts in the U.S. social safety net and the nature of work. Welfare reform in the mid-1990s marked a shift away from unconditional cash benefits toward in-kind benefits such as Medicaid and food stamps and work-contingent programs such as the Earned Income Tax Credit (EITC). At the same time, polarization in the labor market has led to an increase in low-wage, unstable, and unpredictable work arrangements (see Luhr, Schneider, and Harknett 2022, this issue) that may affect families’ ability to qualify for these work-contingent tax credits, as well as childcare (Pilarz, Sandstrom, and Henly 2022, this issue) and other needed assistance (Randles 2022, this issue). Declines in labor-market opportunities for those without a college degree have also likely contributed to increases in family complexity (see Schneider, Harknett, and Stimpson 2018; Edin et al. 2019).

Family diversity can create complexity in claiming needed assistance as each social safety net program defines families differently. For instance, Food Stamp (or SNAP) benefits are determined based on everyone who lives in the same household and purchases or prepares meals together (spouses and children are automatically considered as one unit; claiming for other relatives is less clear). Medicaid eligibility, on the other hand, is determined by income at the nuclear family structure, which is based on the marital status of parents and children. For benefits provided through the tax code, eligibility is based on tax filing units, which is determined based on marital status as of December 31, who resided in the household for at least half of the year, and the ages of the family members. To further complicate matters, many tax credits and deductions within the tax code have their own set of age, financial support, and residency requirements.

We focus on the interaction between the Earned Income Tax Credit and family complexity. The EITC is a critical U.S. antipoverty policy that lifts at least eight million individuals out of poverty each year. For families with children, it does more to reduce poverty than any other current social policy, perhaps excepting the 2021 Child Tax Credit (Fox 2020). Many studies have linked the EITC with increased labor-force participation among unmarried mothers (Meyer and Rosenbaum 2001; Hoynes and Patel 2018; Michelmore and Pilkauskas 2021). Thus, this work-contingent policy plays a central role in the lives of low-wage workers, serving as a wage subsidy for parents who may face unstable or precarious employment. Yet complexity in family structure and in tax rules may leave working parents without access to this important credit. Although tax rules delineate who is eligible to claim children for EITC purposes, the credit was established in 1975, when the average family looked very different from the average family today. Qualitative studies have documented incidents of confusion in determining tax filing units (Halpern-Meekin et al. 2015) and the challenges arising from family complexity (Edin, Tach, and Halpern-Meekin 2014), but the extent to which EITC rules create complicated and confusing tax filing arrangements, or leave out certain family types, is not well understood at a national level.1

Using the American Community Survey (ACS) and the Current Population Survey (CPS), we document the proportion of children residing in various complex living arrangements. We pay particular attention to the proportion in households where tax filing status might be especially complicated: children living with adult relatives in addition to their parents, and those with only one or no birth parent present.2 EITC tax filing complexity arises in these households because multiple adults can potentially claim the same child. Using information from the National Survey of Family Growth (NSFG), and the Fragile Families and Child Wellbeing Study (FFCWS), we provide more detailed estimates of how many children spend part of their time residing with nonresident or noncustodial parents, and the financial support that they receive from nonresident parents. We then discuss the implications of these complex living arrangements for EITC claiming and potential reforms that could provide more support for complex families. The findings from this study have broader implications for the social safety net, and the 2021 expansion of the Child Tax Credit in particular,3 which requires assessing children’s living arrangements on a monthly basis.

## BACKGROUND

The Earned Income Tax Credit began in 1975 as a temporary credit, made permanent in 1978, for low-income parents, intended to offset payroll tax contributions.4 The credit is fully refundable, meaning that households with no tax liability can still receive the credit in the form of a tax refund. The benefit schedule is trapezoidal: benefits phase in up to a threshold, remain constant over some values of income (plateau), and then phase out for earnings beyond a second threshold. The EITC provides more generous benefits to households with multiple children, but no additional benefit for those with more than three. For households without qualifying children, the maximum benefit for childless filers in 2019 was only $529 and those with earnings above $15,570 (approximately the annual income of an individual working full time at the federal minimum wage) were not eligible.5 In contrast, households with three children and earnings below $50,162 could claim benefits worth up to $6,557. Families with children whose income is up to roughly 230 percent of the federal poverty threshold may be eligible for the EITC, and benefits could be worth up to 45 percent of annual earnings.

Because taxes are levied at the family level, the tax code, and the EITC in particular, is not marriage neutral. Until 2002, married filers faced the same benefit schedule as unmarried filers, creating a substantial “marriage penalty” for EITC claimants. That is, many couples who would be eligible for the EITC if unmarried and filing as separate tax units, would be ineligible for the credit if they were to marry and file their taxes jointly, creating some concern that the EITC discourages marriage (see Dickert-Conlin and Houser 2002; Herbst 2011; Michelmore 2018). Over the last two decades, efforts have been made to reduce the marriage penalty by extending the plateau region of the benefit schedule for married filers. These changes allow married couples to have higher earnings than unmarried filers and still maintain their EITC benefits. In 2019, the threshold for married filers was extended $5,800 beyond that of unmarried filers.

### Qualifying Child Rules

Strict eligibility and low benefit levels for tax filers without qualifying children mean that families with children are the primary beneficiaries of the EITC. Who is considered a child in a household may seem straightforward, but complex relationship and residency tests create uncertainty for some families. Indeed, whether intentional or due to confusion, the vast majority of erroneous claiming of the EITC is due to misclaiming of dependent children (Liebman 2000).

An individual—or qualifying child—must meet three requirements: age, relationship, and residency. We describe the eligibility rules and then highlight areas where the rules are less clear.

The age requirement states that the qualifying child must be under the age of nineteen, or twenty-four if a full-time student at the end of the tax year. The child must also be younger than the tax filer.6

To meet the relationship test, the child must be the biological, adopted, step, or foster child of the parent. A child may also be a descendent of the claimant, such as a grandchild, niece, or nephew. A child could also meet the relationship test if they are the brother or sister of the claimant, including full, half-, or step-sibling.

The child must also reside in the household with the tax filer for at least six months of the year, though exceptions are made if the child was born or died during the tax year.

Although the age requirement is relatively straightforward, both the relationship test and the residency test in complex families can be ambiguous. This can arise when more than one household or person in a household meets all three requirements, such as when a child’s parents are divorced but share custody of the child, when two cohabiting birth parents of the child live together and could both potentially claim the child, or when parents live with other relatives who are also potentially eligible to claim the child.

When who can claim the child is ambiguous, the Internal Revenue Service (IRS) provides a set of tiebreaker rules. These state that if more than one taxpayer can claim the child, the qualification goes first to the child’s parent (in cases when a child lives with other relatives). If more than one parent is a taxpayer (parents who do not file taxes jointly or unmarried parents), the qualification will go to the parent with whom the child has lived the longest in the previous year. However, if a child spends an equal amount of time with both parents (such as in joint custody or cohabiting birth parents), then the qualification goes to the parent with the highest adjusted gross income (AGI). If no parent taxpayer is eligible to claim the child (such as when parents are unemployed), then the taxpayer with the highest AGI who also lives with and is related to the child may claim the child. In practice, whichever parent or caregiver files their taxes first and claims the child as a qualifying child for the purposes of the EITC will receive the refund; it is up to the secondary parent or caregiver (or the IRS auditing process) to contest the claim.

In sum, despite rules to determine who can claim the child, in more complex family arrangements, particularly if the child does not reside in a household with two married, birth parents, the ambiguity in who can claim is significant. This ambiguity is especially important when it might be easy to make mistakes, both parents have significant financial hardships, and because the IRS is far more likely to audit low-income filers than higher-income filers (Kiel and Essinger 2018; Kiel and Fresques 2019; Tax Policy Center 2020).

## DATA

Because no one data source provides a complete picture of family complexity in the early twenty-first century, we use four data sources in this analysis: the March 2019 Annual Social and Economic Supplement (ASEC) of the Current Population Survey, the 2019 American Community Survey, the 2017–2019 National Survey of Family Growth, and the Fragile Families and Child Wellbeing Study.

The CPS ASEC and the ACS are two large, nationally representative data sources, and allow us to document the various living arrangements among children whose families might be eligible for the EITC. In the CPS ASEC and the ACS, we restrict our sample to those under the age of nineteen, given that qualifying children must be younger than nineteen at the end of the calendar year.7 We supplement these data with the NSFG and the FFCWS to ascertain more information about nonresident parents—how often children see their nonresident parents and whether they receive financial support from them and if so how much.

### Current Population Survey

The CPS is a nationally representative annual survey of approximately sixty thousand households. The CPS ASEC contains detailed information on household composition, labor-market status, and annual income from a variety of sources, and is administered in March of each year. Annual income information is based on the 2018 tax year, so we calculate EITC eligibility based on the tax filing rules in 2018. After restricting the sample to those under the age of nineteen, we have a sample size of 49,864 children.

### American Community Survey

The ACS is a nationally representative survey of the U.S. population that samples approximately three million households annually and is collected by the Census Bureau. The ACS data for this study were drawn from extracts made by the Integrated Public Use Microdata Sample, or IPUMS USA (Ruggles et al. 2021). Unlike the CPS, the ACS is conducted on a rolling basis throughout the year, thus income information refers to the prior year rather than the prior tax year. We use tax filing rules for 2018 for the 2019 ACS. We exclude children who live in group quarters because it is not possible to determine their family or household relationships or income and to make more comparable with the CPS. The sample includes 667,326 children.8

### National Survey of Family Growth

The NSFG collects information from a nationally representative sample of men and women of childbearing age, between the ages of fifteen and forty-nine. We use the 2017–2019 survey to obtain estimates about time and financial support provided by nonresident fathers. Men provide information about children who do not reside in their household, including how often their nonresident children sleep over and whether and how much financial support they provide them.9 Of the 2,029 men reporting at least one child, 784 reported having at least one nonresident child.

### The Fragile Families and Child Wellbeing Study

To obtain additional information about nonresident parents, we supplement our analyses with data from the FFCWS, an urban longitudinal birth cohort study of approximately five thousand births, followed at regular intervals between 1999 and 2016.10 Study participants are relatively economically disadvantaged and thus likely eligible for the EITC. Primary caregivers (such as grandparents or other relatives) and both parents were interviewed when the children were nine years old, providing insight into differences in perceptions across caregivers and on nonparental guardianship in general. We draw on data from all survey waves for some questions and focus on the year nine data for comparing across caregiver types.

## MEASURES

The ACS and CPS collect information on all individuals residing in the household and their relationship to the household respondent. We use this information, variables that link children to parents, relationship status, and age to construct the variables for our analyses.

### Household Structure

The ACS and CPS include information on the marital status of every member of the household, as well as the presence of cohabiting partners, which we use to determine whether children reside with married parents, cohabiting parents, or an unmarried, unpartnered parent.11 Using “parent pointers,” or variables that identify coresident parents, we determine whether children have parents in the household. We consider children without parent pointers (or pointers that indicate no parent present) to be residing in households with no parent present.12 Using data on the relationship to the household head and the age of the individuals, we can identify the relationship of other nonparent or nonpartner-of-parent adults in the household.13 We then identify whether other relatives or nonrelatives over the age of eighteen live in the household, distinguishing grandparents from other relatives for sub-analyses.14

Using this information, we categorize children into five living-arrangement groups: married, cohabiting, or single parents with no other adults in the household; no parents but other adult relatives such as a grandparent or uncle in the household; and no related adults in the household. We then create subcategories for each of the parent groups to distinguish living arrangements in which the child lives with only parents from those that include both parents and related adults, and those that include parents and unrelated adults.15 In the CPS, we further distinguish children living with cohabiting parents into those who are living with two birth parents and those who are not as this distinction is important for tax filing purposes.16

### Income and EITC Eligibility

In some analyses, to illustrate the scope of family complexity among this policy-relevant population, we focus on households where some or all of the individuals are likely to qualify for the EITC. To estimate eligibility, we use information on the number of children in the household based on parent reports, marital status of the parents, and parental earnings. We calculate eligibility based on whether parental earnings were below the maximum income threshold in the 2018 tax year given the number of children residing in the household (including those with zero earnings). For married-couple households, we combine both parents’ earnings, because EITC eligibility is based on the earnings of both parents if their filing status is married filing jointly.17 For cohabiting-couple households, we calculate EITC eligibility based on the earnings of the birth parent if only one birth parent is in the household, and based on the earnings of the higher-earning partner if both parents are birth parents, which is consistent with the tiebreaker rules in the EITC.18 We calculate eligibility separately rather than using the CPS imputed variable (which uses NBER’s TAXSIM) because the assumptions made for complicated households (such as cohabiting birth parents) are not clear and because nonparental caregivers (such as a grandparent caregiver) are not considered eligible.19

Because calculating EITC eligibility as described requires many assumptions about the tax filing unit, we also present findings characterizing children by whether their total family income is below 200 percent of the federal poverty line. Per census family definitions, all income of individuals residing in the household related by blood or marriage are considered in this calculation.

### Children’s Household Living Arrangements

Nearly forty-nine million, or a majority of children in our sample, live in married-parent households (see table 1): about 65 percent of children in both samples and the vast majority of children in married households do not live with any other adults beyond the nuclear family (about 90 percent, or 58 percent of all children). Children living with a single parent are the second-largest category—approximately 24 percent (18.3 million) of children, of which two-thirds live with just one parent and no other adults (16 percent of all children). About 7 percent of children (5.6 million) live with cohabiting adults, and according to the CPS, about 40 percent of those children live with both birth parents and 60 percent with one birth parent and one social parent. Very few children in cohabiting households live with other adults (relatives or nonrelatives, less than 1 percent).

View this table:
[Table 1.](http://www.rsfjournal.org/content/8/5/143/T1)

**Table 1.** 
Children’s Living Arrangements

Regardless of the marital status of the parents, about 12 to 13 percent of children live in a household with at least one parent and at least one related adult over the age of eighteen (ten million children). A smaller proportion (1 to 2 percent, or about 1.7 million), live in a household with at least one parent and one unrelated adult over the age of eighteen. Last, 4.5 percent of children under the age of nineteen live in a household without a parent: the majority of those children live with another related adult over the age of 18 (over 3 percent), such as a grandparent (approximately 2 percent). About 1 percent of children live with other adult relatives, such as an aunt or uncle, and the remaining 1 percent live with adult nonrelatives (such as foster care).

When we restrict our sample to children living in poor and near-poor households, we find stark differences in living arrangements relative to the population overall. About 37 percent of children in lower-income households live in a married, two-parent household with no other relatives; they are also nearly twice as likely to reside with cohabiting parents as the full sample of children (12 to 13 percent versus 7 percent among all children). Children in lower-income households are also much more likely to live with a single parent (38 percent versus 24 percent overall) or without any parent than the full population of children (8 percent versus 4.5 percent overall). Interestingly, the share of children living with their parent (or parents) and other relatives or other nonrelatives is similar across the full population and lower-income populations (12 to 14 percent with relatives and 2 to 3 percent with nonrelatives). Results are very similar when we examine households with income below the EITC-eligibility threshold rather than 200 percent of poverty. Because the EITC-eligibility calculation requires several assumptions about tax filing units, and because the findings are quite similar, we rely on the 200 percent of poverty threshold to characterize low-income households for the remaining analyses.

#### Family Complexity by Race-Ethnicity

Table 2 presents the prevalence of each household living arrangement discussed in table 1, by child’s race-ethnicity. For simplicity, we present results using only the CPS data. Results reveal stark differences in the living arrangements of children by race and ethnicity. Whereas almost 84 percent of Asian children and three-quarters of White children reside in married-parent households, only 61 percent of Hispanic children and 37 percent of Black children do. Nearly half of Black children reside in a single-parent household, relative to 26 percent of Hispanic, 16 percent of White, and just 11 percent of Asian children. Cohabitation rates are similar across racial-ethnic groups (6 to 9 percent), with the exception of Asian children, who are substantially less likely to have cohabiting parents (2 percent). Coresidence with other adult relatives is common among all racial-ethnic groups (15 to 18 percent) except among White households (7 percent). Black children are also substantially more likely to live without any parent in the household (7.5 percent) than all others (3 to 4 percent). Much of this is explained by the prevalence of grandparent-headed households among Black children (5 percent).

View this table:
[Table 2.](http://www.rsfjournal.org/content/8/5/143/T2)

**Table 2.** 
Percentage of Children in Each Living Arrangement by Race and Ethnicity

Racial differences remain when restricting the sample to all children residing in lower-income households. Children of all race-ethnicities in lower-income households are less likely to live with married parents: 68 percent of Asian, about half of low-income White and Hispanic, and 20 percent of Black children. White children in low-income households are far less likely to live with a parent in addition to other relatives (10 percent) than their Black, Hispanic, and Asian counterparts (14 to 23 percent).

These differences in family complexity by race and ethnicity have potentially large consequences for EITC claiming. Relative to White children, Black, Hispanic, and especially Asian children are much more likely to have another relative residing in the household who could potentially claim them as a qualifying child. For Black children in low-income households in particular, family complexity due to nonresident parents and living with relatives means that more than 80 percent of such children potentially face complexity in EITC claiming due to a nonresident parent or living with multiple relatives.

### Tax Filing Complexity: Household Type*s*

Nonresident parents, the presence of a relative (with or without a parent present), and living with two, cohabiting birth parents can lead to tax filing complexity. In the next section we consider each household type in turn.

#### Nonresident Parents

The most common intersection between family and tax complexity arises when children have at least one nonresident parent. The complexity stems from the fact that a nonresident parent can potentially claim the child for the EITC if the child spends at least half the year living with that parent. This might arise from a variety of scenarios: joint or shared-custody agreements (postdivorce or when child support orders are set up), midyear moves, or even informal shared-custody agreements. Table 3 presents the proportion of children who have at least one nonresident parent and the common living arrangements among this population. We illustrate this pattern among all children, as well as children from low-income households—overall and by race-ethnicity.20

View this table:
[Table 3.](http://www.rsfjournal.org/content/8/5/143/T3)

**Table 3.** 
Percentage of Children with at Least One Nonresident Parent

Among all children under the age of nineteen, approximately 40 percent have at least one nonresident parent. Living without at least one birth parent is even more common, 58 percent, when we limit the sample to children in households with incomes below 200 percent of the federal poverty line. Differences by race are also significant: nearly 80 percent of Black children in lower-income households have one nonresident parent versus 55 percent of White, 50 percent of Hispanic, and 38 percent of Asian children. Although most children who have at least one nonresident parent live in single-parent-headed households with no others—24 percent of all children and 38 percent of children in low-income households—levels vary dramatically by race-ethnicity. Sixty-five percent of Black children live with a single parent, versus only 22 percent of Asian children.

Among children in lower-income households, 7 percent live with one birth parent and one step-parent and 6 percent live with one birth parent and one “social parent” (the cohabiting nonbirth partner of the parent). Roughly 4 percent of low-income children live with a relative (3 percent with a grandparent and 1 percent with other relatives) with no parent present, but rates for Black children are higher (5 percent grandparent and 2 percent other relatives) than for White (4 percent grandparent, 1 percent other relative), Hispanic (2 percent grandparent, 1 percent other relative), and Asian children (1 percent grandparent, 1 percent other relative) in lower-income households.

Among lower-income households, then, approximately 60 percent of children do not live with two, married, birth parents, creating some ambiguity in tax filing. However, we might overestimate the number of children facing tax filing complexity because of this form of family complexity for several reasons. First, a child may have a deceased parent. According to estimates from the Social Security Administration, approximately 1.6 percent of children receive benefits because of the death of a working parent (Tamborini, Cupito, and Shoffner 2011).21 Second, a child may be born to a parent who used a sperm or egg donor. Estimates of the proportion of children born from donor parents vary wildly, larger estimates showing about 0.5 percent of children in the United States are conceived via sperm donation (Arocho, Lozano, and Halpern 2019), many of whom, about 60 percent (Hertz, Nelson, and Kramer 2013) live in two-parent families, both same and different sex couples. Parents self-report whether they are the biological parent of the child; thus, it is unclear in the survey whether married parents who are not genetically related to the child consider themselves biological parents. Nonetheless, parental death and donor parents may lead us to overestimate the proportion of children with nonresident parents by about 2 percentage points. Third, we assume that nonresident parents even know about their child; however, the extent to which this affects our estimates is unclear.

To be claimed for the EITC, a child must also live with the nonresident parent at least half of the year and that parent must have a relatively low income. To better understand the EITC eligibility of nonresident parents, and thus the number of potentially ambiguous scenarios, we use data from the NSFG on nonresident fathers.22 Fewer than half of nonresident fathers report annual earnings below $40,000 per year, the upper income threshold for EITC eligibility for a single parent with one qualifying child in 2018 (see table 4). Only 26 percent of nonresident fathers report that their child spent any nights with them in the previous four weeks; 10 percent report that the child stayed several times a week or more—an indicator that the child lives with them half-time or more, and thus might be eligible to claim the EITC. Putting these two criteria together, just 5 percent of nonresident fathers in the NSFG meet both the income and residency requirements to claim the child for the EITC.

View this table:
[Table 4.](http://www.rsfjournal.org/content/8/5/143/T4)

**Table 4.** 
Overnight Time with Nonresident Fathers

In table 5, using data from the FFCWS, we further consider ambiguity in EITC eligibility that could arise when children live in more than one household. Among this lower-income sample, 68 percent of nonresident fathers report that the child lives with them at least some of the time; 36 percent report half of the time or more. These estimates are much higher than those reported by nonresident fathers in the NSFG. However, when we turn to mother’s reports on how much of the time the child lives with her, we see large discrepancies between mother’s and father’s perceptions. Ninety-four percent of mothers who do not live with the child’s father report that the child always lives with her; 21 percent of fathers report that the child always lives with him. 23 Although we have no way of assessing which report is more accurate, this finding raises questions about the extent to which nonresident parents may feel entitled to claim the EITC on behalf of their child.

View this table:
[Table 5.](http://www.rsfjournal.org/content/8/5/143/T5)

**Table 5.** 
Time Living with Nonresident Parents, Discrepancies in Reported Time

We see a similar discrepancy when we examine reports of nonresident parents when the child lives with a relative rather than the other parent. The bottom panel of table 5 presents the relative caregivers’ reports of mothers and fathers time spent with the child as well as mothers’ and fathers’ self-reports. 24 As was true of nonresident fathers, nonresident mothers and fathers self-report far higher levels of contact with their child than the relative caregiver’s accounts. These discrepancies could lead parents and relative caregivers to both attempt to claim the child for the EITC. Studies also suggest that challenges around claiming the child for public assistance frequently arise between grandparent caregivers and nonresident parents, given that both may need the financial assistance (Pittman 2015, 2019). The child’s custodial arrangement may not make clear who has the right to claim the child.

In sum, the data from the NSFG suggest that just 5 percent of nonresident fathers meet both the income and residency requirement to claim the child. However, the analyses in the FFCWS demonstrate the complexity of determining how much time children spend with their nonresident parents, given how dramatically opinions differ between parents and between parents and relative caregivers. This high level of disagreement suggests a possibly widespread issue, one that might have real implications for tax filing.

#### Other Relatives in the Household

The second most common living arrangement for children in which some ambiguity in tax filing arises is when children live with other relatives and their parents (13 percent of children). In other words, multiple adults might be eligible to claim the child for the EITC. According to IRS tiebreaker rules, the parent always has priority; but if a relative, such as a grandparent, is also in the household and has a higher AGI, the grandparent may claim the child instead.25 In households where the child’s parents are married and they live with additional relatives, it seems less likely that the relative would claim the child (about 6 percent of children). This situation may be more common, however, among single-parent households who live with other relatives (7 percent of children). The situation is further complicated by the fact that many of these children also have nonresident parents; thus families must negotiate within and across households to ensure that only one individual claims the child. In the rare case that both the parent and the child are under the age of nineteen (less than 1 percent in the CPS sample), the grandparent (or other relative) may claim both the parent and the child for the EITC.

#### Two-Parent Cohabiting Households

Another source of complexity arises for children in cohabiting households with both birth parents (about 2.5 percent of children, and 5 percent of children in lower-income households). In these cases, both parents are eligible to claim the children for the EITC. According to EITC tiebreaker rules, the parent with the highest AGI should claim the child. Yet in households with multiple children (75 percent of all households headed by cohabiting parents), parents are each free to claim some children in the household (as long as they meet income requirements) to minimize tax liability; research suggests that families do so (Halpern-Meekin et al. 2015; Jones and O’Hara 2016). For example, a household with four qualifying children and two income-eligible parents would likely minimize household tax liability if each parent separately claimed two children. That is, families can be strategic in their tax filing to maximize benefits. At the same time, doing so requires a fair amount of tax knowledge. We know little about the extent to which families understand these complexities. Access to tax preparers might increase this likelihood, and 96 percent of taxpayers use software such as Turbo Tax (40 percent) or expert preparers (56 percent) to file their taxes (Goldin 2018). Further complications arise if cohabiting partners have some shared, biological children as well as children from prior relationships.

### Implications of Tax Filing Complexity

Our analysis suggests two main forms of ambiguity in EITC claiming due to family complexity: when a child does not live with two married birth parents, and when multiple related unmarried adults live with the child. Although many children live in households that could potentially raise such ambiguities, our data cannot estimate how many actually do. Recent data from tax audits provide some context. The Internal Revenue Service estimates that between 22 and 26 percent of tax returns that claim the EITC have errors (Greenstein, Wancheck, and Marr 2019). Most of these errors, however, are unintentional. About 70 percent are due to residency and relationship requirements (U.S. Treasury 2014). Thus the interplay between family complexity and the tax system affects many low-income families. That it does is especially important given that the IRS is five times as likely to audit households who receive the EITC than those who do not (TRAC IRS 2022). Families are far less likely to claim the EITC after an audit, and may be less likely to be employed after an audit, especially families with young children (Guyton et al. 2018). Thus these audits and errors can be consequential. That family complexity is especially common among Black families suggests that it may also contribute to racial inequality in income and wealth.

## HOW CAN THE TAX CODE ACCOMMODATE MORE COMPLEX FAMILY STRUCTURES?

The majority of children in lower-income households experience some type of family complexity, posing challenges to determining eligibility for the EITC. Both the Bush and Obama administrations proposed rules to simplify the EITC, and the Treasury Department has also proposed error-reduction policies that have been put forth in various House and Senate bills (Greenstein, Wancheck, and Marr 2019). We focus our discussion on two policy proposals to better accommodate complex family structures: a federal noncustodial parent EITC and a permanent expansion to the childless EITC.

### Noncustodial Parent EITC

Given that the majority of children in low-income households have at least one nonresident parent, our first proposal is to introduce a federal noncustodial parent (NCP) EITC. Proponents of an NCP EITC argue that many noncustodial parents cannot claim their child for the EITC despite providing them financial support.

A noncustodial parent EITC could simultaneously provide tax relief for parents providing financial support for nonresident children and reduce child poverty by encouraging noncustodial parents to further financially support their children. Evidence from New York State, which has its own NCP EITC, suggests that in the first few years after the credit was established, noncustodial fathers were about 1.6 percentage points more likely to be employed and 1 to 2 percentage points more likely to pay their child support in full (Nichols, Sorensen, and Lippold 2012). Providing noncustodial parents their own EITC could also reduce tension and confusion over which parent should claim the EITC when the child spends a portion of the year with each parent, or when children primarily reside with a nonparental relative. Research shows that many low-income, noncustodial parents are taxed into poverty through the tax code (Marr and Huang 2020). An NCP EITC might lift more noncustodial or nonresident parents out of poverty, which should also benefit children.

New York and Washington, D.C., both have noncustodial parent EITCs that could serve as models for a federal NCP EITC. In New York, the maximum NCP EITC available in 2019 was $1,323. A full-time, year-round worker earning the minimum wage in New York at that time would have an annual income of about $23,000 and be eligible for an NCP EITC of about $575, approximately 3 percent of their pretax annual income.26 In both, the NCP EITC is linked with the child support system: to claim the credit, noncustodial parents must have a formal child support order and have paid all of their child support from the previous calendar year.27

Linking the noncustodial parent EITC to the child support system provides a third-party verification system and ensures that only parents who financially support their nonresident children receive the credit. The obvious downside to this structure is that noncustodial parents without formal child support orders are not eligible, and child support orders are less common among low-income parents in particular (Nelson 2004). For example, figure 1 shows that though more than half of nonresident fathers in the NSFG provide financial support to their children, only 28 percent do so through a formal child support order. In the FFCWS, a lower-income sample of fathers, only 23 percent of nonresident fathers provide formal support, but 40 percent provide in-kind support such as toys and clothing. Another study estimated that only about one-third of fathers with incomes under $40,000 had a formal child support order and only one-third of those were fully compliant with the order (Wheaton and Sorensen 2010).

![Figure 1.](http://www.rsfjournal.org/https://www.rsfjournal.org/content/rsfjss/8/5/143/F1.medium.gif)

[Figure 1.](http://www.rsfjournal.org/content/8/5/143/F1)

**Figure 1.** 
Financial Support from Nonresident Parents *Source*: Authors’ calculations from National Survey of Family Growth, 2017–2019 (CDC 2022); Fragile Families and Child Wellbeing Study, 1999–2017 (CRCW and CPRC 2022).

*Note*: Men age fifteen through forty-nine reporting fathering at least one child in the NSFG. In the FFCWS, the sample is restricted to maternal reports of nonresident fathers and sample waves 1–6 are pooled.

A further complication in designing a federal noncustodial credit is that many noncustodial parents reside with some of their biological children. These parents would then potentially be eligible to claim more than one type of EITC. In the NSFG, 35 percent of fathers with nonresident children resided with some of their biological children; the vast majority (85 percent) also provided financial support to their nonresident children.28 Allowing tax filers to simultaneous claim an NCP EITC and the EITC for those with qualifying children would acknowledge the financial support that many fathers provide to both their resident and nonresident children and would likely help reduce the number of children and parents living in poverty.

How much would a federal NCP EITC cost? Estimates vary depending on whether the federal credit would be linked with the child support system. Laura Wheaton and Elaine Sorensen (2010) estimate that a federal NCP EITC that required fathers to be fully compliant with child support orders would cost about $1.1 billion, but only about 11 percent of the noncustodial fathers with income below $40,000 would be eligible. Removing the child support requirement, therefore, would potentially scale up the costs by an order of magnitude and imply a federal cost of about $11 billion.

In sum, although the current NCP EITC policies are closely tied to the child support system, and good reasons support noncustodial parents’ providing financially for their children, a policy not coupled with child support would likely provide greater benefit to children and parents. Although policy approaches that both encourage child support and financial support more generally (say through a bigger credit for those who are also up to date on child support orders) are possible, they would incur additional administrative burdens and costs. Future research and proposals are needed to outline a federal NCP EITC—those that do and do not link it with child support systems.

### Expand the Childless EITC

Rather than create a separate EITC for noncustodial parents, another way to provide benefits to nonresident parents is through an expansion of the EITC for all workers without qualifying dependents, often referred to as the childless EITC.29 Support for expanding the childless EITC has been growing (see, for example, Maag, Werner, and Wheaton 2019; Williams 2019).30 Also, in 2021, the Biden American Rescue Plan Act (ARPA) incorporated a temporary expansion to the credit for the 2021 tax year. The expansion nearly tripled the maximum benefit from the previous year, from $538 in 2020 to $1,502 in 2021. The expansion also reduced the minimum age required to claim from twenty-five to nineteen, and increased the maximum earnings threshold from $15,000 to $21,000. This expansion of the income threshold means that a childless filer working full time, year round at the federal minimum wage would now be eligible for a credit worth about $800; under the previous law they would have been ineligible. For a federal minimum wage earner, the expansion of the childless EITC is more generous than the NCP EITC in New York, for instance, but full-time, year-round workers in states with higher minimum wages (say $12 per hour) would still be ineligible for the childless EITC, even under the ARPA expansion. Although it is too soon to evaluate the full effects of the expansion of the childless EITC, estimates suggest that the 2021 ARPA expansion will increase the costs of the credit from $2 billion in 2018 to $11.9 billion in 2021 (Crandall-Hollick 2021).

Making the ARPA extension permanent for all workers without dependent children has the advantage of simplicity and reduced administrative costs because it avoids the need for a verification system, either child social security numbers or child support payments if formal child support is required. This approach would provide support to a wider range of individuals: nonresident parents with and without formal child support orders, those without children (but who may one day be parents), nonrelative caregivers, and unmarried, cohabiting couples. Despite its name, an expansion to the childless EITC would likely affect many parents who do not live with any of their biological children, but who provide financial support to their children. According to the NSFG, 65 percent of men with nonresident children do not live with any of their children, making these men appear childless, even though they have children. Greater economic stability among this population would also have the potential for spillovers (such as increased education, relationship stability) that may affect children in the future.

## CONCLUSION

Drawing on data from four large-scale U.S. surveys, we consider how different aspects of family complexity might affect EITC claiming among families with children. We focused on three key forms of complexity: nonresident parents, when one or both parents are nonresident; the presence of other relatives in the household; and two unmarried, birth parents. We found that a large share of children experience at least one of these complexities at a given time, variation by household income and race and ethnicity being substantial.

Our first key finding relates to the very high rates of nonresident parents that might complicate EITC claiming. Approximately 40 percent of all children and 60 percent of children in lower-income households have at least one nonresident parent. Although living with a single parent is the most common living arrangement when a child has a nonresident parent (24 percent of all children, 38 percent of low-income children), a nontrivial share of children live with a step- or social parent (14 percent of all children, 13 percent of children in lower-income households), or no parents (4 percent of all children, 7 percent of low-income).

Another key finding relates to the stark differences in rates of nonresident parents by race and ethnicity. As documented in numerous studies (for example, Cross 2018), Black children are far less likely to reside with two, married, birth parents than their White, Hispanic, and Asian counterparts. Among lower-income households in particular, approximately 80 percent of Black children had one nonresident parent, compared to approximately 50 percent of White and Hispanic children, and 25 percent of Asian children.

The existence of a nonresident parent alone does not introduce tax filing ambiguity—the nonresident parent must qualify for the EITC based on income and residency. From the NSFG, we find that only 5 percent of nonresident fathers would meet both the income and residency requirement, suggesting that perhaps only a small share of nonresident parents are truly eligible to claim the credit. Nonetheless, we do not know how many nonresident parents feel entitled to claim it, and our analyses in the FFCWS suggest that mother’s and father’s perceptions can differ substantially. We also find differences in perceptions between nonparental caregivers and nonresident parents, suggesting that these discrepancies might be relatively widespread. Thus, although it is not clear what share of households truly face confusion in which relative is eligible to claim the child for the EITC, it is likely to affect a nontrivial share of children living in lower-income households, particularly Black children.

Consistent with prior studies (see, for example, Harvey, Dunifon, and Pilkauskas 2021), we find that about 15 percent of children live in a shared household (with an additional relative or nonrelative). Although rates of household sharing among children living with married and single parents are similar, ambiguity in tax claiming is likely to be more of a consideration in single-parent households, as the majority of children residing with married parents are likely claimed by their parents. Rates of coresidence with parents and other adult relatives in lower-income households varied by race and ethnicity: 10 percent of White, 13 percent of Black, 17 percent of Hispanic, and 23 percent of Asian children lived with their parents and a relative. Thus, although nonresident parents were uncommon among Asian children relative to other groups, these children are far more likely to coreside with other relatives (usually grandparents).

Last, we estimate that about 5 percent of children in low-income households live with two cohabiting birth parents. Complexity arises in this situation because taxes are levied at the family level, so each parent would file a separate return, creating ambiguity about which parent should claim the child. Differences in cohabitation rates by race and ethnicity are small, Hispanic children being somewhat more likely and Asian children somewhat less likely to live in such households.

How can we reform the EITC to better address the needs of complex families? Our analysis focused on two reforms in particular: establishment of a federal noncustodial parent EITC and expansion to the existing childless EITC (or making the ARPA expansion permanent). Both expansions to the EITC, whether for childless adults or a form of a noncustodial credit, would likely increase employment (Nichols, Sorensen, and Lippold 2012; Miller et al. 2018).31 Offering more generous support to nonresident parents could also reduce conflict and confusion between resident caregivers and nonresident parents, who may feel entitled to claim their children for the EITC even if they fail to meet residency requirements. The appropriate course of action depends on the overall goals—to focus on the economic well-being of children and their parents, or low-income households more broadly. Expansions to either the childless EITC or the NCP EITC could help many children by improving the economic well-being of their parents and caregivers. However, more research on both policies is needed. Research on the spillover benefits of a childless tax credit to children (both born and yet to be born) would be especially useful.

Although we focus on two proposed EITC policy changes, other reforms would help extend the reach of the EITC and lift more children out of poverty. One might allow both coresident unmarried birth parents and other coresident relatives (such as coresident grandparents) to claim an EITC (or a smaller EITC). Others might completely separate child-related tax credits from work-related tax credits in the tax code, effectively eliminating child qualifications from the EITC and providing instead more generous child credits through a child allowance (Burman 2020). Another distinct but related issue is the marriage penalty. When two unmarried partners have earnings in the EITC-eligibility range, it may be advantageous for them to remain unmarried for the EITC. Despite a number of adjustments to reduce the marriage penalty in recent years, the EITC still creates a marriage disincentive, though evidence suggests that the effects on marriage are relatively small (Dickert-Conlin and Houser 2002; Herbst 2011; Michelmore 2018).

Because of data limitations, our analysis focuses on family complexity at a particular time; the share of families affected by ambiguity in tax filing is likely to affect a much larger share of families if we were to follow families over time, as children may be claimed in different tax filing units over the course of childhood (Tong 2014; Splinter, Larrimore, and Mortenson 2017). As noted earlier, we may overestimate the number of children with nonresident parents because of parental death or assisted reproduction. Finally, the data used in this study predate the COVID-19 pandemic, and the effects of the pandemic on employment, and in particular mother’s employment (who are most likely to claim the EITC), are not yet clear. Nor do we know what the effects of the pandemic will be on children’s living arrangements; however, it is likely that complexity, and need, will have both increased.

Our study has implications for the recent expansions to the Child Tax Credit (CTC) as part of the 2021 American Rescue Plan Act. The expansion of the CTC essentially created a child allowance in the United States, providing monthly cash benefits between $250 and $300 per month, per child, depending on the age of the child. Estimates suggest that nearly 90 percent of children under the age of eighteen were eligible for the benefits. Our analysis suggests that many children do not reside with both birth parents, which has implications for determining eligibility not only for the EITC, but for the 2021 CTC as well. Because the CTC was distributed monthly, it likely created additional challenges in determining children’s living arrangements.

The EITC lifts many families out of poverty (Fox 2020) but can be complicated to navigate, especially when families and living arrangements are complex. Reducing the complexity of the EITC would help families understand and better follow the rules. Expanding the reach of the EITC would allow more parents to tap into the resources that might improve their economic stability and the well-being of their children. Given the many studies that show that income improves children’s life trajectories (see, for example, Duncan, Magnuson, and Votruba-Drzal 2017), policies that extend the EITC—acknowledging the complexity of families—are likely to pay off in the long term.

## Appendix

View this table:
[Table A.1.](http://www.rsfjournal.org/content/8/5/143/T6)

**Table A.1.** 
Percentage of Children with at Least One Nonresident Parent by Race/Ethnicity

## FOOTNOTES

*   1. A number of reports have noted the potential complications arising from family complexity and the tax system (Goldin and Kleiman 2021; Landers and Crandall-Hollick 2021; Maag, Peters, and Edelstein 2016).

*   2. The CPS contains information on parent type, distinguishing between biological, step, or adopted parents. Here, we use the term birth parent to refer to a biological or adopted parent, as compared to a step-parent or social parent.

*   3. American Rescue Plan Act of 2021, Pub. L. No. 117-2, 135 Stat. 4 (March 11, 2021).

*   4. As of 2019, twenty-nine states and the District of Columbia also had state-level EITCs.

*   5. The EITC for households without qualifying children is typically referred to as the childless EITC. In fact, many of these filers do have children, but for tax-filing purposes they cannot claim them. In 2021, under the American Rescue Plan Act, the childless EITC was temporarily expanded, tripling the max credit (to approximately $1,500) and extending the income eligibility to $21,000.

*   6. We use the term child throughout even though eighteen-year-olds are generally considered adults by law (they can vote and be drafted). The eligibility age for the EITC is slightly higher than for the Child Tax Credit; until 2020 the Child Tax Credit age was under seventeen, under the 2021 expansion the age was changed to under eighteen.

*   7. Although children under the age of twenty-four who are full-time students are also eligible, we do not include them in these analyses because of data constraints: we are unable to observe the tax filing units of individuals who do not live with their parents while they are enrolled in school.

*   8. Certain differences between the ACS and CPS sampling might lead to slightly different estimates. First, the CPS only includes people who “usually” live in the household, whereas the ACS considers individuals to be part of the household if, at the time of the survey, they have been in the house, or will be in the house for more than two months and do not have a “usual residence elsewhere.” Second, the ACS is mandatory and the CPS is a voluntary survey.

*   9. Although women are asked whether any of their children do not reside with them, women are not asked questions about time spent with nonresident children or about financial support provided to them.

*   10. For more information, see the study website at Princeton University ([https://fragilefamilies.princeton.edu](https://fragilefamilies.princeton.edu)).

*   11. In the CPS, information on the presence of cohabiting partners is available for all household members. The ACS has this information only for the respondent.

*   12. In the ACS, the parent pointer identifies any coresidential romantic partner as the child’s parent. Thus, social parents (unmarried or married step-parents or cohabiting partners) will be identified as the child’s parent in some cases. This may mean that some partners who do not consider themselves a parent figure of the child will be erroneously considered a parent. It is not possible to distinguish biological parents in the ACS.

*   13. If the parent (or grandparent) is the reference person we can identify the relationship to all other household members with great accuracy. However, if a nonrelative is the reference person, the child and all their relatives are likely categorized as nonrelatives and thus the specific relationships cannot be identified, though we can still link children to their parents via the parent pointers. In a supplemental analysis in the ACS, we used the parent pointers to identify the child’s grandparents; in only a handful of cases did children reside with both a parent and grandparent but neither parental figure was the household respondent, suggesting this is unlikely to introduce many errors in identifying children’s relatives.

*   14. Following prior work, we do not consider siblings of the child of any age to be “additional adults” (see Pilkauskas, Garfinkel, and McLanahan 2014).

*   15. When the household includes both additional related adults and unrelated adults, children are categorized as living with related adults because these related adults could potentially claim children for the purposes of the EITC but unrelated adults could not.

*   16. We do not do the same in the ACS because the information is not collected.

*   17. Married couples who file as married filing separately cannot claim the EITC, so we assume that all married couples file jointly.

*   18. In the ACS, we use measures of family income minus public assistance income. This is the measure used by the Census Bureau for estimating income-to-needs in the ACS. We do this because this variable seems to produce more consistent estimates than the sum of the parent’s earnings variables for households with more than one coresident parent. It is not clear why this is the case; however, for households with other relatives (that would be included in the poverty threshold) we just use the earnings of the parents. This approach leads to very similar estimates of eligibility between the ACS and CPS despite different look-back periods for income.

*   19. For example, the CPS imputation does not assign children to grandparents as they are not the child’s “parents,” despite the fact that grandparents can technically claim these children as their primary caregiver. Despite these limitations an analysis using the CPS imputed variable shows similar estimates to those produced using our approach for households where parents are present.

*   20. In table A.1, we show these figures for all children by race-ethnicity.

*   21. About 1 percent of children in the CPS sample reside with a widowed parent, though this is likely an underestimate of the percent of children with a deceased parent, as parents who remarried or were never married would not be considered widows.

*   22. Although both men and women were asked about whether each of their children resides in the home in the NSFG, men were asked more extensively about their support for nonresidential children than women. Among women who had at least one live birth, approximately 5 percent report that at least one of their children lives with their birth father or other relatives. Thus we focus on nonresident fathers.

*   23. Response rates from fathers are much lower than those from mothers. This table only examines data when both parents were interviewed—but it is a somewhat select group as these may be more involved fathers than the general population. However, in supplemental analyses that did not limit to those where both mothers and fathers responded the distribution of time spent was similar.

*   24. These analyses are on very small samples and thus are only suggestive. We restrict the sample to those for which the mother, father, and nonparental caregiver responded. In analyses for which we retained the full sample of nonparental caregivers, the results were similar.

*   25. As is the case with divorced parents who share custody, whichever tax filer claims the child first will receive the EITC benefit; the IRS does not allow multiple filers to claim the same child (as verified by their Social Security Number). However, if multiple filers attempt to claim the same child, this may trigger an investigation by the IRS, and would require filers to substantiate their claims with additional documentation.

*   26. The New York State NCP EITC is calculated as either 20 percent of the federal EITC if the tax filer had one child in the household, or 2.5 times the federal EITC calculated with no qualifying children. In the District of Columbia, noncustodial parents between the ages of eighteen and thirty who paid their child support order in full from the previous year can claim a DC EITC worth 40 percent of the federal credit they would have qualified for under the federal tax rules.

*   27. Noncustodial parents may have child support arrears from previous years, but must have paid child support for the most recent year. A federal NCP EITC would need to consider whether refunds would be intercepted to pay for child support arrears.

*   28. Approximately 30 percent of low-income noncustodial fathers lived with some of their children (Nichols, Sorensen, and Lippold 2012).

*   29. Because noncustodial parents typically cannot claim their nonresident children on their tax returns, the IRS categorizes them as childless filers.

*   30. A few states have also expanded childless EITC benefits by reducing the minimum age required to claim, and increasing the size of the credit; for example, Minnesota doubled the size of the EITC available to childless workers in recent years. California, Maine, Maryland, and Minnesota have all either reduced the age threshold required to claim the childless EITC (twenty-one in Minnesota, eighteen in California and Maine), or eliminated it altogether (Maryland). Washington, D.C., Maine, Minnesota, and California have also increased the maximum benefit available for childless workers. D.C. matches the federal EITC for childless workers dollar for dollar, and increased the income threshold for claiming from roughly $15,000 to $25,477 in 2019. Minnesota also recently doubled the size of the EITC available for childless workers who earn less than $15,570, and Maine increased the match rate for childless workers from 12 to 25 percent of the federal credit.

*   31. This might also help nonresident parents provide financial support to children being raised by relatives.

*   © 2022 Russell Sage Foundation. Michelmore, Katherine M., and Natasha V. Pilkauskas. “The Earned Income Tax Credit, Family Complexity, and Children’s Living Arrangements.” *RSF: The Russell Sage Foundation Journal of the Social Sciences* 8(5): 143–65. DOI: 10.7758/RSF.2022.8.5.07. Research reported in this publication was supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) of the National Institutes of Health under award numbers R01HD036916, R01HD039135, and R01HD040421, as well as a consortium of private foundations. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Direct correspondence to: Katherine Michelmore, at kmichelm{at}umich.edu, 735 S. State St. Ann Arbor, MI 48109, United States; Natasha Pilkauskas, at npilkaus{at}umich.edu, 735 S. State St., Ann Arbor, MI 48109, United States.

Open Access Policy: *RSF: The Russell Sage Foundation Journal of the Social Sciences* is an open access journal. This article is published under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.

## REFERENCES

1.  Arocho, Rachel, Elizabeth B. Lozano, and Carolyn T. Halpern. 2019. “Estimates of Donated Sperm Use in the United States: National Survey of Family Growth 1995–2017.” Fertility and Sterility 112(4): 718–23.
    
    

2.  Burman, Leonard E. 2020. “A Universal EITC: Making Work Pay in the Age of Automation.” National Tax Journal 73(4): 1187–218.
    
    

3.  Carlson, Marcia, J., and Daniel R. Meyer. 2014. “Family Complexity: Setting the Context.” The Annals of the American Academy of Political and Social Science 654(1): 6–11.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1177/0002716214531378&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000342384800001&link_type=ISI) 

4.  Center for Research on Child Wellbeing and Columbia Population Research Center (CRCW and CPRC). 2022. *Fragile Families and Child Wellbeing Study, 1999–2017*. Princeton, N.J.: Princeton University. Accessed June 6, 2022. [https://fragilefamilies.princeton.edu/documentation](https://fragilefamilies.princeton.edu/documentation).
    
    

5.  Centers for Disease Control and Prevention (CDC). 2022. *National Survey of Family Growth, 2017–2019*. Washington: U.S. Department of Health and Human Services. Accessed June 6, 2022. [https://www.cdc.gov/nchs/nsfg/nsfg\_2017\_2019\_puf.htm](https://www.cdc.gov/nchs/nsfg/nsfg_2017_2019_puf.htm).
    
    

6.  Cherlin, Andrew J. 2010. *The Marriage-Go-Round: The State of Marriage and the Family in America Today*. New York: Vintage.
    
    

7.  Crandall-Hollick, Margot. 2021. “The ‘Childless’ EITC: Temporary Expansion for 2021 Under the American Rescue Plan Act of 2021 (ARPA; P.L. 117–2).” *CS Insight* IN11610. Washington, D.C.: Congressional Research Service.
    
    

8.  Cross, Christina J. 2018. “Extended Family Households Among Children in the United States: Differences by Race/Ethnicity and Socio-Economic Status.” Population Studies 72(2): 235–51.
    
    

9.  Dickert-Conlin, Stacy, and Scott Houser. 2002. “EITC and Marriage.” National Tax Journal: 55(1): 25–40.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.17310/ntj.2002.1.02&link_type=DOI) 

10. Duncan, Greg J., Katherine Magnuson, and Elizabeth Votruba-Drzal. 2017. “Moving Beyond Correlations in Assessing the Consequences of Poverty.” Annual Review of Psychology 68(1): 413–34.
    
    [PubMed](http://www.rsfjournal.org/lookup/external-ref?access_num=27648987&link_type=MED&atom=%2Frsfjss%2F8%2F5%2F143.atom) 

11. Edin, Kathryn, Timothy Nelson, Andrew Cherlin, and Robert Francis. 2019. “The Tenuous Attachments of Working-Class Men.” Journal of Economic Perspectives 33(2): 211–28.
    
    

12. Edin, Kathryn., Laura Tach, and Sarah Halpern-Meekin. 2014. “Tax Code Knowledge and Behavioral Responses Among EITC Recipients: Policy Insights from Qualitative Data.” Journal of Policy Analysis and Management 33(2): 413–39.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1002/pam.21739&link_type=DOI) 

13. Fox, Liana. 2020. “The Supplemental Poverty Measure: 2019.” *Current Population Reports*, series P60, no. 272. Washington: U.S. Census Bureau.
    
    

14. Furstenberg, Frank F. 2014. “Fifty Years of Family Change: From Consensus to Complexity.” Annals of the American Academy of Political and Social Science 654(1): 12–30.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1177/0002716214524521&link_type=DOI) 

15. Goldin, Jacob. 2018. “Tax Benefit Complexity and Take-Up: Lessons from the Earned Income Tax Credit.” *Stanford Law and Economics Olin* working paper no. 514. Stanford, Calif.: Stanford Law School.
    
    

16. Goldin, Jacob, and Ariel Jurow Kleinman. 2021. “Designing a Universal Child Allowance: Who can Claim which Kids?” *Up Front* (Brookings Institution blog), April 12. Accessed March 29, 2022. [https://www.brookings.edu/blog/up-front/2021/04/12/designing-a-universal-child-allowance-who-can-claim-which-kids](https://www.brookings.edu/blog/up-front/2021/04/12/designing-a-universal-child-allowance-who-can-claim-which-kids).
    
    

17. Greenstein, Robert, John Wancheck, and Chuck Marr. 2019. “Reducing Overpayments in the Earned Income Tax Credit.” Washington, D.C.: Center for Budget and Policy Priorities. Accessed March 29, 2022. [https://www.cbpp.org/research/federal-tax/reducing-overpayments-in-the-earned-income-tax-credit#\_ftnref9](https://www.cbpp.org/research/federal-tax/reducing-overpayments-in-the-earned-income-tax-credit#_ftnref9).
    
    

18. Guyton, John, Kara Leibel, Dayanand S. Manoli, Ankur Patel, Mark Payne, and Brenda Schafer. 2018. “The Effects of EITC Correspondence Audits on Low-Income Earners.” *NBER* working paper no. W24465. Cambridge, Mass.: National Bureau of Economic Research.
    
    

19. Halpern-Meekin, Sarah, Kathryn Edin, Laura Tach, and Jennifer Sykes. 2015. *It’s Not Like I’m Poor: How Working Families Make Ends Meet in a Post-Welfare World*. Berkeley: University of California Press.
    
    

20. Harvey, Hope, Rachel Dunifon, and Natasha V. Pilkauskas. 2021. “Under Whose Roof? Understanding the Living Arrangements of Children in Doubled-Up Households.” Demography, 58(3): 821–46.
    
    

21. Herbst, Christopher M. 2011. “The Impact of the Earned Income Tax Credit on Marriage and Divorce: Evidence from Flow Data.” Population Research and Policy Review 30(1): 101–28.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1007/s11113-010-9180-3&link_type=DOI) 

22. Hertz, Rosanna, Margaret K. Nelson, and Wendy Kramer. 2013. “Donor Conceived Offspring Conceive of the Donor: The Relevance of Age, Awareness, and Family Form.” Social Science & Medicine 86: 52–65.
    
    

23. Hoynes, Hilary W., and Ankur J. Patel. 2018. “Effective Policy for Reducing Poverty and Inequality? The Earned Income Tax Credit and the Distribution of Income.” Journal of Human Resources 53(4): 859–90.
    
    [Abstract/FREE Full Text](http://www.rsfjournal.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NToid3BqaHIiO3M6NToicmVzaWQiO3M6ODoiNTMvNC84NTkiO3M6NDoiYXRvbSI7czoyMDoiL3JzZmpzcy84LzUvMTQzLmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

24. Jones, Maggie R., and Amy B. O’Hara. 2016. “Do Doubled-Up Families Minimize Household-Level Tax Burden? National Tax Journal 69(3): 613–40.
    
    

25. Kidman, Rachel, Rachel Margolis, Emily Smith-Greenaway, and Ashton M. Verdery. 2021. “Estimates and Projections of COVID-19 and Parental Death in the US.” JAMA Pediatrics 175(7): 745–46.
    
    

26. Kiel, Paul, and Jesse Eisinger. 2018. “Who’s More Likely to Be Audited: A Person Making $20,000— or $400,000?” *Propublica*, December 12. Accessed March 29, 2022. [https://www.propublica.org/article/earned-income-tax-credit-irs-audit-working-poor](https://www.propublica.org/article/earned-income-tax-credit-irs-audit-working-poor).
    
    

27. Kiel, Paul, and Hannah Fresques. 2019. “Where in the U.S. Are You Most Likely to Be Audited by the IRS?” *Propublica*, April 1. Accessed March 29, 2022. [https://projects.propublica.org/graphics/eitc-audit](https://projects.propublica.org/graphics/eitc-audit).
    
    

28. Landers, Patrick A. and Margot L. Crandall-Hollick. 2021. “Child Tax Benefits and Children with Complex or Dynamic Living Arrangements.” *CRS Insight* IN11634. Washington, D.C.: Congressional Research Service.
    
    

29. Liebman, Jeffrey B. 2000. “Who Are the Ineligible EITC Recipients?” National Tax Journal 53(42): 1165–85.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.17310/ntj.2000.4S1.06&link_type=DOI) 

30. Luhr, Sigrid, Daniel Schneider, and Kristen Harknett. 2022. “Parenting Without Predictability: Precarious Schedules, Parental Strain, and Work-Life Conflict.” RSF: The Russell Sage Foundation Journal of the Social Sciences 8(5): 24–44. DOI: [https://doi.org/10.7758/RSF.2022.8.5.02](https://doi.org/10.7758/RSF.2022.8.5.02).
    
    

31. Maag, Elaine, H. Elizabeth, Peters, and Sara Edelstein. 2016. “Increasing Family Complexity and Volatility: The Difficulty in Determining Child Tax Benefits.” Washington, D.C.: Tax Policy Center. Accessed March 29, 2022. [https://www.taxpolicycenter.org/sites/default/files/alfresco/publication-pdfs/2000641-increasing-family-complexity-and-volatility-the-difficulty-in-determining-child-tax-benefits.pdf](https://www.taxpolicycenter.org/sites/default/files/alfresco/publication-pdfs/2000641-increasing-family-complexity-and-volatility-the-difficulty-in-determining-child-tax-benefits.pdf).
    
    

32. Maag, Elaine, Keven Werner, and Laura Wheaton. 2019. “Expanding the EITC for Workers Without Resident Children.” Washington, D.C.: Urban Institute. Accessed March 29, 2022. [https://www.urban.org/research/publication/expanding-eitc-workers-without-resident-children#](https://www.urban.org/research/publication/expanding-eitc-workers-without-resident-children#).
    
    

33. Manning, Wendy D., and Bart Stykes. 2015. “Twenty-Five Years of Change in Cohabitation in the U.S., 1987–2013.” FP-15-01. Bowling Green, Oh.: National Center for Family & Marriage Research.
    
    

34. Marr, Chuck, and Yixuan Huang. 2020. “Childless Adults Are Lone Group Taxed into Poverty.” Washington, D.C.: Center on Budget and Policy Priorities. Accessed March 29, 2022. [https://www.cbpp.org/research/federal-tax/childless-adults-are-lone-group-taxed-into-poverty](https://www.cbpp.org/research/federal-tax/childless-adults-are-lone-group-taxed-into-poverty).
    
    

35. Meyer, Bruce D., and Dan T. Rosenbaum. 2001. “Welfare, the Earned Income Tax Credit, and the Labor Supply of Single Mothers.” Quarterly Journal of Economics 116(3): 1063–14.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1162/00335530152466313&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000170353300009&link_type=ISI) 

36. Michelmore, Katherine M. 2018. “The Earned Income Tax Credit and Union Formation: The Impact of Expected Spouse Earnings.” Review of Economics of the Household, 16(2): 377–406.
    
    

37. Michelmore, Katherine M., and Natasha V. Pilkauskas. 2021. “Tots and Teens: How Does Child’s Age Influence Maternal Labor Supply and Child Care Response to the Earned Income Tax Credit?” Journal of Labor Economics 39(4): 895–929.
    
    

38. Miller, Cynthia, Lawrence F. Katz, Gilda Azurdia, Adam Isen, Caroline B. Schultz, and Kali Aloisi. 2018. “Boosting the Earned Income Tax Credit For Singles: Final Impact Findings from the Paycheck Plus Demonstration in New York City.” New York: MDRC.
    
    

39. Monte, Lindsay M. 2019. “Multiple-Partner Fertility in the United States: A Demographic Portrait.” Demography 56(1): 103–27.
    
    

40. Nelson, Timothy J. 2004. “Low-Income Fathers.” Annual Review of Sociology 30(1): 427–51.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1146/annurev.soc.29.010202.095947&link_type=DOI) 
    
    [Web of Science](http://www.rsfjournal.org/lookup/external-ref?access_num=000223503200020&link_type=ISI) 

41. Nichols, Austin, Elaine Sorensen, and Kye Lippold. 2012. “The New York Noncustodial Parent EITC: Its Impact on Child Support Payments and Employment.” Washington, D.C.: Urban Institute.
    
    

42. Pilarz, Alejandra Ros, Heather Sandstrom, and Julia R. Henly. 2022. “Making Sense of Child Care Instability Among Low-Income Families: (Un)desired and (Un)planned Reasons for Changing Child Care Arrangements.” RSF: The Russell Sage Foundation Journal of the Social Sciences 8(5): 120–42. DOI: [https://doi.org/10.7758/RSF.2022.8.5.06](https://doi.org/10.7758/RSF.2022.8.5.06).
    
    

43. Pilkauskas, Natasha V., and Christina Cross. 2018. “Beyond the Nuclear Family: Trends in Children Living in Shared Households.” Demography 55(6): 2283–97.
    
    

44. Pilkauskas, Natasha V., Irwin Garfinkel, and Sara S. McLanahan. 2014. “The Prevalence and Economic Value of Doubling Up.” Demography 51(5): 1667–76.
    
    

45. Pittman, LaShawnDa. 2015. “How Well Does the ‘Safety Net’ Work for Family Safety Nets? Economic Survival Strategies Among Grandmother Caregivers in Severe Deprivation.” RSF: The Russell Sage Foundation Journal of the Social Sciences 1(1): 78–97. DOI: [https://doi.org/10.7758/RSF.2015.1.1.05](https://doi.org/10.7758/RSF.2015.1.1.05).
    
    

46. Pittman, LaShawnDa. 2019. “How Well Does the “Safety Net” Work for Family Safety Nets?” Paper presented at the APPAM conference “Rising to the Challenge: Engaging Diverse Perspectives on Issues and Evidence,” Denver, Colo. (November 7–9, 2019).
    
    

47. Randles, Jennifer. 2022. “Fixing a Leaky U.S. Social Safety Net: Diapers, Policy, and Low-Income Families.” RSF: The Russell Sage Foundation Journal of the Social Sciences 8(5): 166–83. DOI: [https://doi.org/10.7758/RSF.2022.8.5.08](https://doi.org/10.7758/RSF.2022.8.5.08).
    
    

48. Ruggles, Steven, Sarah Flood, Sophia Foster, Ronald Goeken, Jose Pacas, Megan Schouweiler, and Matthew Sobek. 2021. IPUMS USA: Version 11.0 [dataset]. Minneapolis, Minn.: IPUMS.
    
    

49. Schneider, Daniel, Kristen Harknett, and Matthew Stimpson. 2018. “What Explains the Decline in First Marriage in the United States? Evidence from the Panel Study of Income Dynamics, 1969 to 2013.” Journal of Marriage and Family 80(4): 791–811.
    
    

50. Splinter, David, Jeff Larrimore, and Jacob Mortenson. 2017. “Whose Child Is This? Shifting of Dependents Among EITC Claimants Within the Same Household.” *Finance and Economics Discussion* Series 2017–089. Washington, D.C.: Board of Governors of the Federal Reserve System.
    
    

51. Tamborini, Christopher R., Emily Cupito, and Dave Shoffner. 2011. “A Profile of Social Security Child Beneficiaries and Their Families: Sociodemographic and Economic Characteristics.” Social Security Bulletin 71(1): 1–16. Accessed March 28, 2022. [https://www.ssa.gov/policy/docs/ssb/v71n1/v71n1p1.html](https://www.ssa.gov/policy/docs/ssb/v71n1/v71n1p1.html).
    
    [PubMed](http://www.rsfjournal.org/lookup/external-ref?access_num=22191282&link_type=MED&atom=%2Frsfjss%2F8%2F5%2F143.atom) 

52. Tax Policy Center. 2020. *How Do IRS Audits Affect Low Income Families?* Briefing Book. Washington, D.C.: Tax Policy Center. Accessed March 29, 2022. [https://www.taxpolicycenter.org/briefing-book/how-do-irs-audits-affect-low-income-families](https://www.taxpolicycenter.org/briefing-book/how-do-irs-audits-affect-low-income-families).
    
    

53. Tong, Patricia K. 2014. “Tracking EITC Qualifying Children Over Time.” Presented at the 107th Annual Conference on Taxation. Santa Fe, N.M. (November 13, 2014). Accessed March 29, 2022. [https://ntanet.org/2014/11/107th-annual-conference-proceedings-2014](https://ntanet.org/2014/11/107th-annual-conference-proceedings-2014).
    
    

54. TRAC IRS. 2022. *IRS Audits Poorest Families at Five Times the Rate for Everyone Else*. TRAC, March 8. [https://trac.syr.edu/tracirs/latest/679/](https://trac.syr.edu/tracirs/latest/679/).
    
    

55. U.S. Census Bureau. 2022. *Current Population Survey, 2019*. Washington: U.S. Department of Commerce. [https://www.census.gov/programs-surveys/cps.html](https://www.census.gov/programs-surveys/cps.html).
    
    

56. U.S. Department of the Treasury (U.S. Treasury). 2014. *Agency Financial Report: Fiscal Year 2014*. Washington: Government Printing Office. Accessed March 29, 2022. [http://www.treasury.gov/about/budget-performance/annual-performance-plan/Documents/508%20FY%2014%20AFR%20FINAL.pdf](http://www.treasury.gov/about/budget-performance/annual-performance-plan/Documents/508%20FY%2014%20AFR%20FINAL.pdf).
    
    

57. Wheaton, Laura, and Elaine Sorensen. 2010. “Extending the EITC to Noncustodial Parents: Potential Impacts and Design Considerations.” Journal of Policy Analysis and Management 29(4): 749–68.
    
    [CrossRef](http://www.rsfjournal.org/lookup/external-ref?access_num=10.1002/pam.20533&link_type=DOI) 

58. Williams, Richard. 2019. “Expanding Earned Income Tax Credits for Childless Workers.” NCSL LegisBrief 27(43): 1–2. Accessed March 29, 2022. [https://www.ncsl.org/research/human-services/expanding-earned-income-tax-credits-for-childless-workers.aspx](https://www.ncsl.org/research/human-services/expanding-earned-income-tax-credits-for-childless-workers.aspx).