Estimating the Net Fiscal Cost of a Child Tax Credit Expansion — NBER Working Paper 29342

Summary

NBER Working Paper 29342, dated October 2021, Revised February 2022, is by Jacob Goldin, Elaine Maag, and Katherine Michelmore. The abstract says the authors study the fiscal cost of three proposals to expand the Child Tax Credit (CTC) that would expand refundability to low-income children, increase the maximum credit amount, and/or eliminate the income phase-out to make the credit universal. They use the Current Population Survey to estimate three components of net fiscal cost, direct cost, revenue changes due to labor supply responses, and long-term changes in tax revenue due to changes in children's future earnings. The abstract says direct costs are by far the most important component and that long-term earning changes offset about 20% of the direct costs of making the credit fully refundable.

Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used

Full text

                              NBER WORKING PAPER SERIES




     ESTIMATING THE NET FISCAL COST OF A CHILD TAX CREDIT EXPANSION

                                        Jacob Goldin
                                        Elaine Maag
                                    Katherine Michelmore

                                      Working Paper 29342
                              http://www.nber.org/papers/w29342


                    NATIONAL BUREAU OF ECONOMIC RESEARCH
                             1050 Massachusetts Avenue
                               Cambridge, MA 02138
                         October 2021, Revised February 2022




We are grateful to Daniel Feenberg, Hilary Hoynes, David Splinter, and Robert Moffit for helpful
suggestions and to Arun Ramesh for outstanding research assistance. The views expressed herein
are those of the authors and do not necessarily reflect the views of the National Bureau of
Economic Research.

NBER working papers are circulated for discussion and comment purposes. They have not been
peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies
official NBER publications.

© 2021 by Jacob Goldin, Elaine Maag, and Katherine Michelmore. All rights reserved. Short
sections of text, not to exceed two paragraphs, may be quoted without explicit permission
provided that full credit, including © notice, is given to the source.
Estimating the Net Fiscal Cost of a Child Tax Credit Expansion
Jacob Goldin, Elaine Maag, and Katherine Michelmore
NBER Working Paper No. 29342
October 2021, Revised February 2022
JEL No. H2,H24

                                         ABSTRACT

Recent proposals to expand the Child Tax Credit (CTC) are at the center of current policy
discussions in the United States. We study the fiscal cost of three such proposals that would
expand refundability of the credit to low-income children, increase the maximum credit amount,
and/or eliminate the income phase-out to make the credit universal. For each proposal, we use the
Current Population Survey to estimate three components of the net fiscal cost: the direct cost
(additional tax refunds or lower tax liability), revenue changes due to taxpayers’ labor supply
responses, and long-term changes in tax revenue due to changes in children’s future earnings. We
find that direct costs are by far the most important component but that long-term earning changes
also play an important role, offsetting 20% of the direct costs of making the credit fully
refundable. In contrast, labor supply responses modestly contribute to the fiscal cost of the CTC
expansions we model.

Jacob Goldin                                    Katherine Michelmore
Stanford Law School                             735 S State St
559 Nathan Abbott Way                           Ford School of Public Policy
Stanford, CA 94305                              University of Michigan
and NBER                                        Ann Arbor, MI 48109
jsgoldin@law.stanford.edu                       kmichelm@umich.edu

Elaine Maag
The Urban Institute
emaag@urban.org
Estimating the Net Fiscal Cost of a Child Tax Credit Expansion

Jacob Goldin, Elaine Maag, and Katherine Michelmore1

February 2, 2022


Abstract

Recent proposals to expand the Child Tax Credit (CTC) are at the center of current policy
discussions in the United States. We study the fiscal cost of three such proposals that would
expand refundability of the credit to low-income children, increase the maximum credit amount,
and/or eliminate the income phase-out to make the credit universal. For each proposal, we use
the Current Population Survey to estimate three components of the net fiscal cost: the direct cost
(additional tax refunds or lower tax liability), revenue changes due to taxpayers’ labor supply
responses, and long-term changes in tax revenue due to changes in children’s future earnings.
We find that direct costs are by far the most important component but that long-term earning
changes also play an important role, offsetting 20% of the direct costs of making the credit fully
refundable. In contrast, labor supply responses modestly contribute to the fiscal cost of the CTC
expansions we model.


    I.         Introduction

The Child Tax Credit (CTC) is a federal income tax benefit for families with children. First
enacted in 1997, the credit has been the focus of increasing political attention and debate in
recent years. In particular, some researchers and advocates have criticized its design, pointing out
that – before 2021 - children living in the lowest-income households received no or limited
benefits and children of color received the lowest average benefits. Recent proposals to reform
child benefits to remedy these concerns have been embraced by President Biden, members of
Congress, and in influential reports authored by academics. Proponents of these reforms point to
research suggesting significant short- and long-term benefits of increased financial assistance to
low-income households in which children are being raised.

In this paper we study the cost of three proposals to expand the CTC relative to 2020 law:
making the credit fully refundable allowing even very low-income families to receive the
maximum benefit, making the 2021 American Rescue Plan Act (ARPA) reform permanent
(which makes the credit fully refundable and increases the generosity of the credit), and creating
a universal child allowance that would provide a flat dollar amount to all children regardless of
income. We focus on the fiscal cost of these proposals, and specifically, on three factors that
shape the net cost of the policy. First, using the Current Population Survey (CPS), we estimate
the direct fiscal cost – i.e., the additional tax refund or reduced tax liability that taxpayers would

1
 Goldin: Stanford Law School and NBER. Maag: Urban-Brookings Tax Policy Center. Michelmore:
Ford School of Public Policy, University of Michigan. We are grateful to Daniel Feenberg, Hilary
Hoynes, Greg Leiserson, David Splinter, and Robert Moffit for helpful suggestions and to Arun Ramesh
for outstanding research assistance.
                                                  1
receive under the reform, relative to 2020 tax law. Second, we study how an expanded CTC
would shape individuals’ labor supply decisions through both income effects and changes in
marginal tax rate schedules. Third, drawing on a literature estimating effects of financial support
on children’s earnings in adulthood, we estimate how an expanded CTC would affect tax
revenue in later years. We quantify each of these cost components and explore their sensitivity to
alternative assumptions.

We find that making the CTC fully refundable for low- and middle-income taxpayers with
children while maintaining the pre-2021 maximum benefit amount would cost approximately
$21 billion annually. We estimate the reform would very slightly reduce labor supply among
low-income taxpayers in the short-term but actually reduce the net fiscal cost of the policy by
about 20% by increasing longer-term federal revenue collection. The primary mechanism for this
increase stems from the effect of the expanded credit on future earnings of children in low-
income households. We also estimate the net fiscal cost of a permanent extension of the CTC in
place during 2021 (“the ARPA design”), which was fully refundable and which contained a
higher maximum CTC benefit amount. The larger maximum credit amount substantially raises
the net fiscal cost of this policy relative to full refundability alone – we estimate it to be
approximately $85 billion annually. Finally, we estimate that a universal CTC, delivering the
2021 maximum benefit as a flat dollar amount per child regardless of income, would cost
somewhat more, approximately $95 billion annually. For all three of these policies, we estimate
that the direct costs dwarf changes in revenue due to labor supply effects but we find that longer-
term changes in revenue due to increased earnings by the children who benefit lead to reduced
net fiscal costs in the range of 15 to 20%.

A number of previous studies have analyzed the costs of the components of the CTC expansion
proposals upon which we focus. In particular, a number of researchers have modeled the
distributional effects of proposed CTC expansions, which yields a direct fiscal cost for the policy
(e.g., Tax Policy Center 2021, Acs and Werner 2021, Garfinkel et al. 2021). Like us, Garfinkel et
al. (2021) also investigated the long-term effects of the policy on tax revenue due to increased
earnings (as well as other mechanisms in their case). Finally, both Garfinkel et al. (2021) and a
National Academies of Sciences report (2019) estimated the revenue effects of labor supply
changes induced by the CTC expansion, although both focused only on the associated income
effects of the policy. In contrast, we also model labor-supply changes resulting from changes in
the marginal tax rate schedule that taxpayers would face under an expanded CTC.
Contemporaneously with our study, Corinth et al. (2021) also investigate the labor supply effects
of the CTC expansion, although their modeling approach differs from ours and they do not
consider effects stemming from increased financial assistance to low-income children.2

The remainder of the paper proceeds as follows. Section II provides background and a brief
history of the CTC. Section III discusses the large literature from which our analysis draws the
parameters we assume. In Section IV we estimate the costs of full refundability. Section V we
estimate the cost of alternative policies that would expand the CTC. In Section VI we consider
the sensitivity of our results to alternative behavioral assumptions and modeling choices. Section
VII concludes.

2
 We consider an approach to estimating labor supply responses similar to Corinth et al. (2021) in Section
VI, below.
                                                    2
    II.     Background and History of the CTC

The CTC was enacted as part of the Taxpayer Relief Act of 1997. The credit was introduced as a
$400 per child credit under age 17 that was largely non-refundable. It primarily benefited middle
and upper-middle income taxpayers. The credit has never been indexed to inflation, but has been
expanded several times since its initial enactment. In 1999, the maximum benefit increased to
$500 per child. In 2001, the credit increased to $1,000 as part of the Economic Growth and Tax
Relief Reconciliation Act, which phased in fully in 2003. Finally, in 2018, the credit increased to
$2,000 as part of the 2017 Tax Cuts and Jobs Act – though for many families this represented
more of a consolidation of family benefits than a true reduction in tax liability because of the
simultaneous elimination of the exemption for dependents (Maag 2019). Starting in 2001, a
portion of the CTC was made broadly available as a tax refund and thereby benefited taxpayers
with zero or negative tax liability.3 After 2017, families could receive up to $1,400 per child as a
tax refund.4

Until 2009, there was a minimum earnings threshold of $10,000 to receive a tax refund from the
CTC; this threshold was reduced to $3,000 as part of the American Recovery and Reinvestment
Act in 2009, and subsequently lowered to $2,500 as part of the 2017 tax bill. Families first offset
taxes owed with the CTC and then could receive up to 15 cents of credit for each dollar earned in
excess of the minimum earnings threshold until the combined value of the nonrefundable and
refundable portions of the credit was equal to the maximum credit a family qualified for or until
the maximum refundable credit had been reached. The minimum earnings required to claim the
full credit depended on the number of qualifying children in the household and the filers’ tax
liability.

The CTC also has an upper income threshold, above which the credit phases out. In 1997, the
phase-out region began at $55,000 for married filing separate returns, $75,000 for unmarried
filers, and $110,000 for married joint return filers. Beginning in 2018, the phase-out threshold
was increased substantially to $200,000 for unmarried filers, and $400,000 for married joint
return filers when other tax benefits were consolidated with the CTC. Benefits phased out at a
rate of $50 for every $1,000 of earnings above the upper income threshold until 2021. In 2021,
the formula was changed to accommodate two phase out ranges: the expanded credit amount
began phasing out at $75,000 for single filers, $112,500 for head of household filers and
$150,000 for married couples. Second, the remaining $2,000 credit phased out according to the
2018 phase-out schedule described above.

The requirements for a child to qualify a taxpayer for the CTC are similar, but not identical, to
other child-linked tax provisions. First, the taxpayer generally must reside with the child for
more than half of the year. Second, the taxpayer must be older than the child and one of a

3
  Prior to this, families with at least three children could receive a refundable CTC calculated as payroll
taxes in excess of the EITC.
4
  Beginning in 2018, families with dependents who do not qualify for the CTC, including children ages 18
or full-time students 19 through 24, older dependents, and some children who are not citizens can qualify
for a smaller, nonrefundable credit referred to as the Other Dependent Tax Credit. This credit phases out
concurrent with the higher phase-out of the CTC.
                                                    3
specified set of relatives to the child, such as the child’s parent, grandparent, sibling, aunt, uncle,
or step-parent. Third, the child must not provide more than half of his or her own support.
Fourth, when multiple taxpayers would qualify to claim the same child, a series of tie-breaker
rules specify which tax unit may do so. Finally, since 2018, only children with Social Security
Numbers can be claimed for the credit.5

The CTC has long been criticized for not providing benefits to the lowest-income households.
Estimates suggest that approximately 27 million children under age 17 (one-third of all children
in the U.S.) were living in households that were not eligible for the full benefit under the pre-
2021 law (Burman and Wheaton 2005; Greenstein et. al 2018; Collyer et al. 2019; Goldin and
Michelmore 2020). Of those 27 million, approximately 6.5 million children were completely
ineligible for the credit, either due to insufficient household earnings or because they did not
reside with a qualifying relative. Approximately half of Black and Hispanic children were not
eligible for the full CTC under the pre-2021 law compared to about 25% of white and Asian
children (Goldin and Michelmore 2020).

The American Rescue Plan Act (ARPA) of 2021 introduced several temporary changes to the
CTC to broaden access to the benefit for tax year 2021. The maximum credit was increased to
$3,000 per child aged 6 to 17 and $3,600 per child under age 6. Families can receive up to half of
that amount as monthly payments which started as early as July 2021 and will continue through
December 2021 unless families opt out. The remainder of the credit can be claimed on a 2021 tax
return, no earlier than February 15, 2022. Families that opt out of advanced, monthly payments
can claim their full credit on their 2021 tax return. The credit is fully refundable – low-income
families qualify for the full benefit, regardless of how much they earn. The credit phases out in
two steps. First, the credit begins to decrease at $112,500 of income for single parents filing as
head of household ($150,000 for married couples), declining in value at a rate of 5 percent of
adjusted gross income over that amount. It cannot go below the value of the credit in 2020,
which did not begin phasing out until adjusted gross income reached $200,000 for single parents
($400,000 for married couples). The credit applies to all children who qualify; all else equal,
larger families receive higher benefits. As a result, the income level at which the credit is fully
phased out varies based on how many children are in the family.

After 2021, the CTC is scheduled to return to its pre-2021 design, with a maximum value of up
to $2,000 per child and limited refundability. In his budget proposal, President Biden proposed
extending the 2021 CTC design through 2025 and extending full refundability and eligibility for
17-year-olds (but not the higher maximum credit amounts) thereafter.



    III.     Previous Literature

The recent expansion of the credit will increase after-tax income for the many families who were
previously ineligible, or received only a portion of the CTC. Many families that were already

5
 We do not observe in our data whether children have a Social Security Number (SSN), but estimates from the
Urban Institute tax model suggest that only about 75,000 tax filing units would be fully excluded from the CTC
because their children lack SSNs.

                                                        4
receiving the full benefit will also experience an increase in after-tax income, as their benefits
will increase from $2,000 per child up to $3,000 for children ages 6 to 17 and to $3,600 per child
under age 6. A long line of research informs our understanding of how these expansions may
affect families and children in both the short- and long-term.

Effects of Financial Assistance on Children’s Well-Being

A large body of evidence suggests that increasing income, particularly among low-income
families, has a positive causal impact on a number of child outcomes, ranging from early
childhood up through early adulthood.

For instance, research on the Earned Income Tax Credit (EITC), a refundable tax credit for low-
income workers and one of the largest anti-poverty programs in the United States, links the credit
with higher infant birth weight (Hoynes, Miller, and Simon 2015), higher child test scores (Dahl
and Lochner 2012; 2017), improved access to private health insurance (Baughman 2012), and
improved maternal health (Averett and Wang 2013; Evans and Garthwaite 2014). More recently,
there is also evidence of long-run improvements in child outcomes due to the EITC, from
increases in educational attainment and earnings (Bastian and Michelmore 2018), to improved
health and reductions in obesity in early adulthood (Braga et al 2020).

Another component of the proposed CTC expansions is that they provide larger benefits for
younger children. The benefits of early childhood interventions are well-established in the
literature (e.g. Cunha and Heckman 2007; Duncan et al. 2007; Duncan and Sojourner 2013;
Bailey et al. 2020a). Providing a more generous child benefit for families with children under
age 6 is likely to have greater payoffs for both children and society in the form of improved
human capital, better health, and reductions in poverty, crime, and incarceration rates.

In estimating the costs of an expanded CTC, we will consider behavioral responses in both the
short- and long-run. If an expanded CTC creates substantial negative labor supply effects, for
instance, this would add to the cost of the program through reductions in tax revenue. On the
other hand, if an expanded CTC leads to substantial improvements in child outcomes, this could
reduce the long-run costs of the program through reductions in crime, improvements in education
and labor market outcomes, and reductions in dependence on social welfare programs. We
consider both labor supply and earnings responses in greater detail below.


   A. Prior Research on Labor Supply Responses

Expanding the CTC could affect labor supply decisions of taxpayers subject to the reformed
credit. The direction of the effect is theoretically ambiguous, and the effect likely differs by
income, marital status, and gender. On the one hand, removing the earnings requirement from the
child tax credit will reduce the incentive for some parents to work (Rachidi 2021). On the other
hand, the increased income from the benefit could allow parents to pay for child care or access to
reliable transportation, which may increase their labor supply, as well as incur the expenses and
foregone part-time income necessary to transition to full-time work (West et. al 2021).



                                                5
Responses may also differ depending on whether they were eligible for a full or partial child tax
credit under the previous law. Because of the earnings requirements to claim the child tax credit
under previous law, approximately 27 million children under age 17 lived in families that were
not eligible for the full credit, including about 6.5 million children who were completely
ineligible (Greenstein et al. 2018; Goldin and Michelmore 2020). Those who were completely
ineligible for the credit will face an income effect from the increased generosity of the benefit,
while those who were eligible for a partial credit will face both an income as well as a
substitution effect generated by changes in their marginal tax rates and incentive to maintain
employment. We discuss each of these responses, and the literature on labor supply elasticities in
more detail below.

   i.      Income effects

Expanding the CTC could generate income effects in two ways. First, for filers whose earnings
were too low to qualify for the full pre-2021 CTC benefit, making the credit fully refundable will
raise their after-tax incomes (in some cases, substantially). Second, by increasing the maximum
credit amount, the expansion would generate income effects for all taxpayers, including those
already benefiting by the full credit amount. By increasing after-tax income, either of these
income effects could in theory increase demand for leisure and reduce incentives to work.

While we lack empirical evidence on labor supply responses to this particular type of shock to
household income, smaller policy changes in the U.S. and in other countries offer some insights.
For instance, in Alaska, all residents receive an annual dividend from the profits generated from
the states’ oil reserves, known as the Alaska Permanent Fund. The benefit amount fluctuates
depending on the investments in any given year, but averages about $2,000 per person per year.
Recent research on the effect of this universal benefit finds no negative effect on employment,
and some evidence of a positive part-time employment effect among women (Jones and
Marinescu 2018).

In North Carolina, the opening of a casino on a Cherokee reservation created a natural
experiment where adult members of the Cherokee tribe were suddenly eligible for annual
benefits (averaging about $4,000 per adult per year) from the profits of the casino. Previous
research finds no significant effect of the benefit on labor force participation of men or women,
though they do find suggestive evidence of declines in full-time work among women (Akee et al
2010). In Canada, the introduction of a child allowance program had no negative effect on the
labor supply of single mothers (Baker, Messacar, and Stabile 2021), though an earlier study of a
child benefit policy in Manitoba, Canada did find that increasing the generosity of the benefit led
to reductions in earned income among low-educated households (Milligan and Stabile 2009).

More broadly speaking, there is a large literature in labor economics that estimates labor supply
elasticities with respect to changes in household income. An older paper by Blundell and
Macurdy (1999) provides a review of this literature, and estimates extensive-margin labor supply
elasticities with respect to changes in household income of -0.05 for married men, -0.12 for
married women, and -0.085 for single women. On the intensive margin, they estimate hours of
work elasticities of -0.05 for married men, -0.09 for married women, and -0.07 for single
women. Notably, because these elasticities are estimated from policy reforms for which the

                                                 6
additional income was not concentrated among low-income parents, they may not capture any
positive effects of income on finding and maintaining employment for that group.

    ii.     Substitution effects

For certain taxpayers eligible for a partial benefit under the pre-2021 CTC, eliminating the
refundability phase-in rate by earnings would increase marginal tax rates, thus reducing the
return to work. Similarly, eliminating the earnings phase-in would increase the relative appeal of
not working at all relative to working, which could in principle induce even some middle and
higher income earners to exit the labor force (see Corinth et al. 2021 for an extended discussion
of this point).

The literature on the labor supply effects of the Earned Income Tax Credit (EITC) offers some
insights into the labor supply responses among this group. The EITC has a similar benefit
structure as the previous CTC structure, in that benefits phase-in up to a certain level, reach a
plateau, and then phase out at higher levels of income. The phase-out threshold for the EITC is
substantially lower than that of the CTC, with the EITC completely phasing out for incomes
above roughly 230 percent of the federal poverty line. A long line of research on the EITC has
evaluated both the extensive and intensive margin effects of expanding the EITC over the last
several decades, with most evidence pointing towards large, positive extensive margin effects of
the EITC on single mothers (see Hotz and Scholz 2003; Chetty et al. 2013; and Nichols and
Rothstein 2016 for reviews), slight negative extensive margin effects for married mothers (Eissa
and Hoynes 2004), and no effect of the policy on the labor supply of men.6 Extensive margin
elasticities with respect to changes in the net-of-tax return to work, calculated based on changes
in the EITC tax schedule, range from 0.30 to 0.45 (see Chetty et al. 2013).7

There is not much empirical evidence that individuals respond to intensive margin labor supply
incentives embedded in the EITC benefit structure. Nichols and Rothstein (2016) summarize this
literature, and estimate intensive margin labor supply elasticities ranging from about 0.07 to
0.10, smaller than the extensive margin elasticities found in this literature. This lack of an
intensive margin effect could be explained by a couple of factors: individuals may lack
information on their precise positioning on the EITC benefit schedule and therefore do not know

6
  There is scant evidence on how the previous structure of the CTC affects labor supply, though one study
finds negative labor supply effects associated with children aging out of the qualifying child age range
(Lippold 2019).
7
  Methods for calculating these elasticities vary. Those cited here are calculated by measuring the percent
change in labor supply divided by the percent change in after-tax labor income (i.e. the “return to work”)
before and after reforms to the EITC, as advocated by Chetty et al. (2013). Other, larger elasticities
reported elsewhere (e.g. Hotz and Scholz 2003) are based on changes in gross labor income rather than
net labor income, and suffer from other measurement issues. For instance, Hotz and Scholz (2003)
calculate an elasticity of 1.16 from resaerch done by Eissa and Liebman (1996). This calculation
overestimates the return to work (i.e. the denominator in the elasticity formula) by relying on average
earnings among working single mothers between 1984 and 1990, rather than using earnings from the pre-
reform period. They also do not take into account the change in return to work by comparing after-tax and
transfer income among working and non-working single mothers between 1984 and 1990. Chetty et al.
(2013) adjust for these measurement concerns and calculate an elasticity of 0.30 based on the same
empirical results in Eissa and Liebman (1996).
                                                    7
how their labor supply decisions may affect their eligibility, or they are unable to meaningfully
adjust their employment and earnings throughout the year so as to maximize their benefits.

There has been at least some support for the second hypothesis, by comparing the behavior of
wage and salary workers to self-employed individuals, who typically have more control over
their hours worked and reporting of earnings to the IRS. Previous research has shown that self-
employed individuals are more likely to report earnings right around the first kink point in the
EITC benefit schedule. Most of this literature argues that this response likely reflects a reporting
issue rather than a meaningful change in labor supply (e.g. LaLumia 2009), since self-employed
individuals at all income levels tend to self-report earnings less accurately than wages reported
by employers on behalf of wage and salary workers.

More broadly speaking, there is also a line of research that uses changes in tax rates to estimate
extensive and intensive margin labor supply elasticities with respect to marginal tax rates.
McClelland and Mok (2012) conduct a literature review and estimate a range of extensive
margin elasticities with respect to marginal tax rates: 0 to 0.10 for men and single women8, and 0
to 0.30 for married women. On the intensive margin, they estimate elasticities ranging from -0.10
to 0.20 for men and single women, and 0.10 to 0.30 for married women. Notably, these ranges
include estimates from studies examining behavioral responses to policies enacted as long ago as
the 1980s and 1990s; more recent studies have documented lower elasticities as married
women’s attachment to the labor force has strengthened (McClelland and Mok 2012). For
instance, Lin and Tong (2017) use changes in the tax code during the 2000s and estimate
extensive margin elasticities for married men ranging from 0 to 0.03, and 0.07 to 0.11 for
married women. On the intensive margin, the authors separately estimate income effect
elasticities with respect to the net-of-tax price of income (-0.04 to 0.0 for married men -0.12 to -
0.04 for married women) and substitution effect elasticities (-0.01 to 0.11 for married men 0.08
to 0.22 for married women).

Beyond the income effect of the CTC, there are theoretical reasons to expect that the expansion
may increase parental labor supply. A monthly benefit could provide parents with the means to
purchase more reliable child care, which could allow them to work more. Previous research has
found that access to the Head Start preschool program increased labor supply among single
mothers, providing evidence on the importance of access to affordable child care in labor supply
decisions (Wikle and Wilson 2020). A monthly payment of $500 to some randomly selected
residents of Stockton, California was associated with an increase in full-time employment as
recipients of the payments were able to complete internships, training, and coursework that
improved employment (West et. al 2021).

Overall, there is not much empirical evidence that moderately-sized shocks to household income
lead to substantial declines in labor supply. However, nothing on the scale of the contemplated
CTC expansions has been implemented in recent years, which makes predictions of this type
difficult. There is also a dearth of research on how these income effects affect different groups –
men versus women, higher-income households versus lower-income households, married
couples versus single parents, and by age of child, race, and ethnicity.


8
    Including both those with, and without children.

                                                       8
      B. Prior Evidence on Long-Term Revenue effects

While there is not much empirical evidence of negative labor supply effects of shocks to
household income, the long-run effects of the child tax credit expansion could be quite different,
particularly among the children of child tax credit recipients. With the growing availability of
long-run panel data sources such as the Panel Study of Income Dynamics (PSID) and
longitudinal administrative data, there is a growing body of research on the long-run effects of
childhood exposure to social safety net programs. Previous work finds positive effects of early
life exposure to food stamps, Head Start, and the EITC on a variety of long-run outcomes such as
educational attainment, earnings and employment in early adulthood, adulthood health, and
longevity. These long-run positive effects, if generalizable to the expansion of the child tax
credit, would substantially reduce the long-term costs of the program.

The literature on the long-run effects of exposure to the EITC in childhood on outcomes in
adulthood suggests that a $1,000 increase in EITC exposure in childhood increases the likelihood
of completing a college degree by 4% (Bastian and Michelmore 2018), increases the likelihood
of working and earnings in early adulthood by about 1-2% (Bastian and Michelmore 2018), and
increases the likelihood of reporting excellent or very good health in adulthood by about 3%
(Braga et al 2020). Research also indicates that the EITC reduces early childbearing and
marriage by age 21 by about 3% (Michelmore and Lopoo forthcoming).

There is also a growing literature on the effects of early life exposure to other social safety net
programs on adulthood outcomes, such as the Supplemental Nutrition Assistance Program
(SNAP, formerly food stamps), and the Head Start preschool program. Recent evidence suggests
that being exposed to food stamps in early childhood (before age 6) is associated with increases
in human capital and economic self-sufficiency in adulthood (ages 25-54), as well as increases in
neighborhood quality (Bailey et al 2020a). Similarly, research indicates that children who were
exposed to the Head Start preschool program were more likely to complete high school and earn
a college degree (Bailey et al. 2020b). All of these positive effects suggest long-term cost
savings through increases in tax revenue and potential decreases in dependence on other social
welfare programs. If these effects generalize to exposure to an expanded CTC in childhood, we
would expect similar long-term cost savings associated with an expanded CTC.

      IV.     Simulating the Cost of Full Refundability

In this section we focus on estimating the fiscal cost of expanding the CTC to be fully
refundable. Specifically, the policy we consider eliminates the income-based limits on
refundability that characterized the pre-2021 CTC while maintaining the pre-2021 maximum
credit amount of $2000 per child. In the next section, we consider alternative policies that would
increase the maximum credit amount (as in the ARPA) as well as a universal child allowance
that would be available to all tax units regardless of income.9




9
    We do not model the proposed expansion of the credit to 17 year-olds.
                                                     9
We distinguish three sources of these reforms’ net fiscal cost: (1) direct costs from increased
benefit amounts; (2) tax revenue changes from labor supply responses; and (3) long-term tax
revenue changes driven by effects of the reform on children’s longer-term earnings.

To estimate these costs, we rely on the 2018 ASEC supplement to the CPS. We use the
relationship and household information in the CPS to construct tax units and use the income
information to estimate tax liability for the tax unit for a single tax year. All analyses employ the
CPS survey weights. We limit our analysis to tax units with one or more children below the age
of 17.

   A. Direct Costs

Figure 1 shows CTC benefits by income for a single filer with one child (top panel) and two
children (bottom panel) under alternative CTC designs. For simplicity, we assume the children
are over the age of five. Comparing the pre-2021 CTC with a fully refundable CTC shows that
the benefits of the reform are concentrated among families with low earnings, who, because of
the limits on refundability, do not benefit from the pre-2021 CTC by the full credit amount or in
some cases by any amount. One-child households with incomes below $24,350 would see a
higher CTC under full refundability as would two-child households with income below $30,350.
Households with income above this threshold would not be affected by making the CTC fully
refundable.

The dashed orange line indicates how the CTC benefit structure would change if the ARPA
reforms were made permanent. For single filers with one dependent, this expansion would raise
the CTC benefit amount for those with income below $132,500, but households with income
above this amount would receive the same benefit as in the pre-2021 benefit structure. For single
filers with two dependents, the expansion would increase benefits for all those with income
below $152,500. Finally, the red dashed line indicates the benefit structure of implementing a
universal child allowance, which would raise the benefit amount for all households, regardless of
income.

Although Figure 1 shows variation in the benefits of full refundability by income, it does so
holding fixed several factors on which CTC eligibility depends – marital status and the number
and ages of children in the tax unit. To capture the average benefit gain from a fully refundable
CTC for all households by income, Figure 2 uses the CPS to estimate the average change in CTC
benefits by making the CTC fully refundable for all households with children. We use NBER’s
TaxSim to calculate CTC benefits under the pre-2021 law. To calculate benefits under a fully
refundable CTC, we replace the pre-2021 CTC with the maximum credit amount ($2,000) for
each child in tax units below the phase-out range.

We find that approximately 30% of tax units with children would benefit from full refundability.
On average, the reform benefits this group by $686, and among tax units that benefit, the average
CTC increase is $2,255. Notably, 80% of the tax units that benefit from the reform have positive
earnings. To estimate the total direct cost of making the pre-2021 CTC fully refundable, we sum
the change in benefits for each tax unit in the CPS. The total direct cost of the policy is
approximately $25.49 billion, reported in Column 1 of Table 1.

                                                 10
   B. Labor Supply Effects

Apart from the direct cost of the policy, expanding the CTC could affect revenue by altering
individuals’ decisions about whether and how much to work. This change could operate through
either an income effect or a substitution effect.

       i.      Income Effect

With more income from an expanded CTC, individuals may choose to consume more leisure and
therefore work less. If they work less, they may generate less tax revenue, effectively increasing
the fiscal cost of the program. On the other hand, the additional income could increase work
incentives by reducing common barriers to work, such as affording childcare, transportation, or
job training. All considered, the magnitude and even direction of the effect of income on labor
supply could vary by income and other circumstances.

To simulate the income effect of a fully refundable CTC, we abstract from much of this potential
heterogeneity and use the estimated income elasticities from recent labor economics work on the
topic, applying the same elasticity to the entire income distribution, but allowing for different
elasticities by gender and marital status. In particular, for each tax unit in the CPS, we calculate
the percent change in income that the additional benefit represents and obtain a predicted change
in labor supply from the product of that quantity and the relevant elasticity for the household in
question.

Slightly complicating this approach, however, is the fact that additional income may lead
individuals to adjust their labor supply on either the intensive margin or extensive margin – that
is, they may continue working but work fewer hours or they may stop working entirely. We
account for each of these possibilities. First, we focus on the intensive margin response to
additional income. We calculate the change in labor supply by scaling the percent change in
income that the new CTC benefit represents for the household by the relevant intensive-margin
elasticity from the literature. For married couples, we calculate labor supply changes separately
for each spouse, using household adjusted gross income to calculate the percent change in
income associated with the new CTC. We assume a constant wage rate, so that percent changes
in labor supply translate into equal-sized percent changes in labor income.

To estimate the change in labor supply because of the new income from the expanded credit on
the extensive margin, we again start from the percentage change in income associated with the
new benefit and scale it by the relevant extensive margin income elasticity from the literature.
This yields a percent change in the probability of employment. To model this effect on earnings,
we assume that each individual in the CPS stops working with this derived probability, and has
no earnings for the calendar year.

Following the National Academy of Sciences report (2019), we take the midpoint of the range of
income elasticities from Blundell and Macurdy (1999) for our simulations. On the extensive
margin, we use income elasticities of -0.05 for fathers, -0.12 for married mothers, and -0.085 for

                                                11
single mothers. On the intensive margin, we use hours of work elasticities of -0.05 for fathers, -
0.09 for married mothers, and -0.07 for single mothers.10

To estimate the total change in annual earnings from the fully refundable CTC’s income effect,
we combine the estimated intensive and extensive margin effects. This yields, for each
individual, a change in annual earnings based on the percent change in their income that the
benefit represents.11 Using this predicted change in earnings, we next calculate the implied
change in tax liability for each household using TaxSim. Finally, we estimate the total change in
tax revenue for a year due to the expansion’s income effect by summing the change in tax
liability for each individual.

Panel A of Figure 3 shows the average change in annual tax liability arising from the income
effect of the expanded CTC by adjusted gross income. Because the elasticities we draw from the
literature are positive, our analysis implies the income effect causes a reduction in earnings for
everyone who receives an expanded benefit. However, as the Figure shows, for low-earning
individuals the average effect of the expansion on tax revenue is actually positive. The
explanation for this is that the decline in earnings causes individuals to qualify for a lower EITC
benefit,12 thus reducing their overall refund amount and increasing the net revenue collected by
the government. Thus, the income effect causes individuals to modestly reduce earnings but does
not increase (and in fact reduces) the net fiscal cost of the CTC expansion.

        ii.     Substitution Effects

Next we turn to the change in labor supply from a fully refundable CTC induced by changes in
the marginal tax rate schedule. By eliminating the phase-in of refundability by income, a fully
refundable CTC would eliminate a provision pushing marginal tax rates negative, thereby
increasing the marginal tax rate that individuals face. Whereas the pre-2021 CTC generated a
negative marginal tax rate on earned income, the fully refundable CTC would generate a zero
marginal tax rate. In this way, replacing the pre-2021 CTC with the fully refundable CTC would
effectively raise the marginal tax rate on earned incomes among low-income earners.

To assess the substitution effect generated by the expansion, we simulate how changes in
marginal tax rates will affect earnings, and as a result, tax revenue. To do this, for each tax filing
unit, we first use NBER’s TaxSim to calculate marginal tax rates under the pre-2021 CTC. We
next calculate how marginal tax rates would change under the fully refundable CTC.

Figure 4 plots the change in marginal tax rate by earnings associated with the policy change for a
single tax filer with one dependent, assuming all income comes from labor earnings. Since the
fully refundable CTC eliminates the phase-in of the credit, tax filing units that were previously
on the phase-in portion of the benefit schedule would experience an increase in their marginal

10
   We abstract from heterogeneity in the elasticity of labor supply response by income or race although it
is of course possible that labor supply response elasticities could differ along these margins.
11
   We assume throughout that individuals do not reduce their incomes below zero.
12
   The EITC begins to phase out at approximately $20,000 of earnings, and completely phases out for
households earning more than about $50,000, depending on the number of EITC qualifying children a
household claims.
                                                    12
tax rates, which would reduce their net-of-tax return from working. For a single filer with one
dependent and earnings between $2,500 and $11,800, the net-of-tax rate decreases by about
15%. For those with earnings between $18,650 and $24,600, the net-of-tax rate decreases by
about 10%. A single filer with one dependent and earnings above $24,650 does not face a change
in their net-of-tax rate, since they were receiving the full child tax credit under the pre-2021 law.
Similarly, single filers with one dependent and earnings between $11,800 and $18,650 also
experience no change in their net-of-tax rate of income because they received the full refundable
portion of the pre-2021 CTC ($1400), but were not eligible for the full credit because they lacked
positive tax liability.

To map these changes in marginal tax rates to changes in labor supply, we again turn to the
empirical literature and use the midpoint of the elasticity ranges by gender and marital status
reported by McClelland and Mok (2012): 0.2 for men (married and single) and single women,
and 0.3 for married women.13 These ranges include both individuals with and without children;
we assume these elasticities apply to all individuals in our sample, which is exclusively parents.
Multiplying the elasticities by the percent change in marginal tax rates yields a percent change in
labor supply for each individual. Note that the total tax unit MTR change is the same for both
spouses, but the elasticities could be different (e.g., by gender) so each individual might change
their labor supply by a different percent. As above, we map the change in earnings into a change
in tax revenue by recalculating tax liability for each tax unit assuming the new earnings level and
compare it to tax liability before the labor supply adjustment.

The estimated change in tax revenue from the fully refundable CTC’s substitution effect is
presented in Panel B of Figure 3. As with the income effect, the reduction in earnings for low-
income tax units under the pre-2021 CTC results in an average increase in tax revenue collected,
since many of these filers are in the EITC phase-in range. Hence, even though it slightly reduces
earnings, the net effect on fiscal cost is positive for very low-earning households. For individuals
outside of the EITC phase-in range, the higher marginal tax rates generated from the fully
refundable CTC do cause a reduction in tax revenue.

         iii.     Combined Income and Substitution Effects

To calculate the total change in tax revenue from labor supply responses to the fully refundable
CTC, we sum the estimated labor supply responses from the income and substitution effects. For
each individual, we first calculate the total change in labor supply from these two effects. Next,
we calculate federal tax liability at the new labor supply. Finally, the sum of the difference
between the old and new tax liability is the overall change in annual revenue from the labor
supply effects of the policy. Figure 5 summarizes this combined effect on labor supply. As
described above, and detailed in Figure 6, for some income ranges, a reduction in labor supply
actually translates into increased federal revenue by reducing EITC refunds. Overall though, we
estimate that switching from the pre-2021 CTC to the fully refundable CTC would decrease
federal income and payroll taxes by approximately $0.74 billion because of labor supply
responses.


13
  This elasticity range reflects substitution effects pooled over both the intensive and extensive margins; see
discussion in McClelland and Mok (2012).

                                                          13
     C. Long-Term Earnings Effect

In addition to shaping labor supply incentives for the taxpayers receiving the benefit, expanding
the CTC could affect tax revenue in the longer-term by shaping the later-life earnings of children
who receive benefits. Establishing a causal link between childhood family income and
employment in adulthood is difficult, but quasi-experimental evidence based on social safety net
expansions suggests that increases in family income in childhood lead to employment and
earnings gains in adulthood (Aizer et al. 2016; Hoynes et al. 2016; Bastian and Michelmore
2018). In a review of this literature, Garfinkel et al. (2021) conclude that, on average, a $1,000
increase in household income in childhood is linked with an average increase in earnings in
adulthood by a present discounted value of about $1,060. This increase in earnings could reduce
the long-term costs of the CTC through higher tax revenue.

To estimate how these long-term earning changes would affect the net fiscal cost of a fully
refundable CTC, we take the average of the long-term earnings effects summarized in Garfinkel
et al. (2021) and assume that a $1,000 increase in household income leads to a 0.63% increase in
annual earnings in adulthood. Because household income gains are likely to have a larger effect
on children growing up in lower income families, we follow Garfinkel et al. (2021) and assign
the full multiplier to families with income below $50,000, half of the multiplier to families with
income between $50,001 and $100,000, and no multiplier to families with income above
$100,000. With these assumptions in place, we predict an average increase in children’s future
annual earnings of 0.43% and about 1.02% for children in households with income less than
$50,000.

Translating this increase in future income into a change in tax liability requires knowing the
child’s future earnings absent the expansion. To predict this quantity, we draw on Chetty et al.
(2014), which calculates intergenerational income correlations based on tax return data of
children born in the late 1970s and early 1980s. We use these correlations to derive the expected
future earnings of children based on the income percentile of their parents (or other taxpayer
with whom they live). We then compare the child’s tax liability using TaxSim with and without
the additional income attributable to the CTC expansion.14 Because the earning changes –and
hence the additional tax revenue – do not occur until future years when children enter the labor
market, we follow Garfinkel et al. and discount the additional revenue.15 The results of this
exercise suggest that full refundability generates approximately $4.99 billion of additional tax
revenue per year by increasing the future earnings of children living in households that receive
the expanded credit, reducing the net fiscal cost of the reform by approximately 20% (to around
$21.25 billion per year).

14
   The child’s future tax liability depends not just on income but also on factors such as marital status and
family size. To account for these factors, we estimate the joint probabilities of marital status and number
of children among households in the CPS and randomly assign children in the future to be married or
single and to have 0, 1, 2, or 3 children using these joint probabilities. Note that this approach implicitly
assumes the CTC expansion would not affect fertility or marital patterns.
15
   Specifically, we assume (again following Garfinkel et al. 2021) that children’s higher future earnings
are not reflected for an average of 11 years. We assume a discount rate of 2% per year in our baseline
analysis. We consider a discount rate of 3% per year in Table 2, below.
                                                     14
   V.      Alternative CTC Expansion Policies

This section repeats the analyses in Section IV for two other CTC expansions. The first
corresponds to the one-year policy enacted by the ARPA for 2021, which includes a higher
maximum credit amount. The second corresponds to a universal child allowance, in that it would
provide the same credit amount for all children, regardless of income.

   A. ARPA Design

This subsection repeats the analysis in the previous section for a CTC expansion modeled after
the ARPA design, i.e., the version of the CTC that is in place for 2021. Like the full refundability
policy studied above, the ARPA CTC provides the full credit amount to low- and middle-income
children – i.e., there is no minimum earnings requirement and the benefit amount does not phase-
in by income. However, in addition to making the benefit fully refundable, the reform also
increases the maximum credit amount – to $3,000 per child aged 6 through 16 and $3,600 per
child under the age of 6. Like the pre-2021 CTC, the ARPA CTC gradually phases out for higher
income tax units. In particular, the expanded credit amount begins to phase out starting at
$112,500 for head of household filers ($150,000 for married couples) at a rate of 5%. The
remainder of the credit amount (the $2000 per child in place under the pre-2021 CTC), phases
out at the same rate once a tax unit’s AGI crosses $200,000 for unmarried parents and $400,000
for married couples. The design of the ARPA CTC is summarized in Figure 1.

Because the ARPA CTC makes the credit fully refundable, it generates the same income and
substitution effects described above for the low- and middle-income tax units whose incomes
placed them in the CTC phase-in range under prior law. Similarly, it also generates the same
long-term revenue effects due to increasing transfers to children in these households. However,
the magnitude of these effects are in some cases larger with the ARPA CTC than with full
refundability due to the increase in credit amount. In addition, the ARPA’s phase-out of the
increased credit amount generates additional labor supply effects among higher-income filers not
present with full refundability.

To calculate the direct cost of the ARPA CTC, we manually calculate the amount of CTC each
household would qualify for under the ARPA design and compare it to the amount of pre-2021
CTC the household qualifies for, as calculated by TaxSim. Because of the higher maximum
credit amount, the direct cost of the ARPA CTC is substantially larger than the full refundability
expansion: approximately $96 billion per year (as shown in Column 1, Row 2 in Table 1).

Turning to the labor supply effects of the reform, the increase in the maximum credit amount
does not alter the substitution effect generated from full refundability; under either reform, the
CTC amount does not change based on income for low-earnings tax units. Hence, the ARPA
CTC expansion would generate the same substitution effect for low-income households as
depicted in Figure 3. However, unlike the full refundability reform, the ARPA CTC increases
marginal tax rates on upper-middle income taxpayers due to the phase-out of the maximum
credit amount. By reducing incentives to work for taxpayers in this phase-out range, this reform
would modestly reduce tax revenue collected from that group.


                                                15
Turning to the ARPA CTC’s income effect, the larger credit amount induces a larger change in
labor supply than the full refundability expansion. Still, the magnitudes of this effect on tax
revenue are quite modest, again partly due to the fact that reductions in labor supply for
taxpayers in the EITC phase-in range actually result in a net increase in tax revenue. Combining
the income and substitution effects yields a net increase in the annual costs of the reform due to
labor supply effects of $4.37 billion.16

Finally, to assess the effects on revenue from long-term changes in labor-supply due to the
reform, we employ the same methodology described in the prior section. Because the ARPA
CTC increases the benefit for low-income children by more than full refundability, it generates
larger average increases in children’s annual long-term earnings: 1.20% among all children and
2.42% among children in households with income less than $50,000. Accordingly, the ARPA
CTC also generates larger increases in future tax revenue (approximately $15.50 billion in
present discounted value per year). Combining the direct cost with the labor supply and long-
term earnings effects described above yields a net fiscal cost of the ARPA CTC of approximately
$85.23 billion (see Row 2 of Table 1)

     B. Universal Child Credit

In this subsection we consider a universal child credit, by which we mean a flat per-child dollar
amount regardless of income. In practice, implementing a universal CTC amounts to expanding
the CTC by eliminating the refundability phase-in (like the other two proposals we consider) as
well as the credit phase-out for high-income tax units. Thus, under the universal CTC we
consider, each tax unit would qualify for a fully refundable CTC of $3,600 per child between the
ages of 0 and 5 and $3,000 per child between the ages of 6 and 16.

Because it does not phase-out by income, the direct cost of the universal CTC exceeds that of the
ARPA design; following the same methodology as above, we estimate it to be $108.86 billion
(see Column 1, Row 3 of Table 1). For low- and middle-income taxpayers, the income and
substitution effects of the universal CTC are the same as those of the ARPA CTC. Higher
income taxpayers, however, face lower marginal tax rates under the universal CTC than under
the ARPA design due to the lack of an income phase-out, implying that the former generates a
smaller substitution effect than the latter. At the same time, high-income taxpayers who would be
subject to the phase-out under the ARPA design face a larger income effect under the universal
CTC, since they receive the full benefit amount. This income effect does not greatly shape labor
supply, however, since even the expanded benefit amount constitutes a relatively small percent
change in income for the high-income taxpayers at issue. We estimate that the net fiscal costs
from labor supply changes induced by the universal CTC reform amount to approximately $1.73
billion per year.

Finally, because the universal CTC provides the same benefits to low- and middle-income
children as the ARPA design, we estimate that it generates similar long-term effects on
children’s future labor supply and hence on tax revenue. Summing these components, we


16
  Our estimates of these effect imply approximately 386,000 parents exiting the workforce because of
this reform.
                                                  16
estimate that the net fiscal cost of a universal CTC would be approximately $95.09 billion per
year.

   VI.     Alternative Assumptions for Cost Simulations

This section will replicate the analyses in Section IV and V using alternative modeling choices
and behavioral parameters.

   A. Alternative Labor Supply Elasticities

In our baseline model, we chose the mid-point of ranges of income and substitution elasticities
from the literature. Here, we illustrate how our cost estimates would change if we use the upper
and lower bound elasticity estimates instead.

Column 1 of Table 2 presents revised estimates of fiscal cost using the lower bound elasticity
estimates from the literature (substitution effect elasticities of 0.1 for fathers and single mothers,
and 0.2 for married mothers; income effect elasticities of 0 on the extensive margin for all, and -
0.01 for fathers and single mothers and -0.05 for married mothers on the intensive margin) to
estimate labor supply responses. That is, we now assume parents will be less responsive to
changes in income and marginal tax rates. For the fully refundable CTC, this reduces the
expected decrease in tax revenue from $0.74 billion to $0.61 billion, slightly decreasing the net
fiscal cost of the fully refundable CTC. For the ARPA and universal CTC simulations, using the
lower bound labor supply elasticities reduces the loss in tax revenue to $2.33 billion and $0.48
billion, respectively. This compares to the $4.37 billion and $1.73 billion in expected losses of
tax revenue generated from our main models for the ARPA and universal child benefit
simulation, respectively.

Column 2 of Table 2 shows results using the upper bound elasticities for labor supply responses
found in the literature (substitution effect elasticities of 0.3 for fathers and single mothers, and
0.4 for married mothers; income effect elasticities of -0.10 on the extensive margin for fathers
and single mothers, and -0.15 for married mothers; and -0.29 for fathers and single mothers and
0.40 for married mothers on the intensive margin). Accordingly, this yields larger labor supply
responses, which increases the costs of the fully refundable CTC slightly, but substantially
increases the costs of the ARPA and universal CTC. For the fully refundable scenario, using the
upper bound elasticities for labor supply responses decreases tax revenue by about $0.80 billion,
compared to a $0.74 billion reduction in revenue in our main estimates. On the other hand, since
the ARPA and universal CTC simulations affect the labor supply of higher-income tax units,
assuming larger labor supply responses results in larger reductions in tax revenue than in our
main specifications. In particular, we predict that the ARPA expansion would reduce annual tax
revenue by about $9.53 billion (compared to $4.37 billion in the main model), while the
universal child benefit would reduce annual tax revenue by $6.69 billion (compared to $1.73
billion in the main model).



   B. Separately Modeling the Extensive and Intensive Margin Substitution Effects.



                                                  17
Our baseline labor supply analysis incorporates income and substitution effects. The income
effect is modeled separately using distinct elasticities for the intensive and extensive margins.
The substitution effect also incorporates both intensive and extensive margin responses, in that
the substitution elasticities we draw from (summarized by McClelland and Mok (2012)), are
estimated based on individuals’ earning responses on both the intensive and extensive margins.
Although our approach should yield a similar overall reduction in work as modeling the intensive
and extensive margins separately, it could imply different results for the distribution of earnings
effects (and hence changes in tax liability) as well as the other outcomes such as the change in
employment.

To assess the robustness of our findings to this difference in modeling approach, we estimate the
percent change in the return to work for each individual under each of the reforms we consider,
using the approach described in Corinth et al. (2021). As described above, each of the CTC
expansions reduces the return to work (i.e., the difference in after-tax income between working
and non-working) by eliminating the current phase-in structure of the CTC, which may generate
reductions in labor supply on both the intensive and extensive margins. In the follow simulations,
we separately estimate the intensive and extensive margin responses due to the substitution
effect. Unlike Corinth et al., we continue to model the intensive margin of the substitution effect
as well as the extensive margin.17

To fill in the required elasticities for this exercise, we consider two sources. First, we use the
midpoints of the ranges described in McClelland and Mok (2012), matching our methodology for
selecting elasticities in our baseline analysis. On the intensive margin, this implies an elasticity
of 0.05 for fathers and single mothers, and 0.2 for married mothers. On the extensive margin, this
implies an elasticity of 0.05 for fathers and single mothers, and 0.15 for married mothers. We
also conduct additional simulations using the lower and upper bound estimates for the extensive
margin, implying lower bound elasticities of zero for all groups, and upper bound elasticities of
0.10 for fathers and single mothers, and 0.30 for married mothers.

Our second approach is motivated by the observation in the literature, described above, that
behavioral elasticities appear to have declined in magnitude in recent years as women have
become more attached to the labor force. Consequently, applying the elasticities reported in
papers studying responses from 30 years ago or more may overstate the magnitude of the labor
supply response to the CTC expansion. To assess this possibility, we consider the most recent
high-quality estimates of labor supply substitution elasticities of which we are aware, reported in
Lin and Tong (2017), which studies behavioral responses to taxation from 2000-2009.18
Although Lin and Tong study only married couples, we assume that single women and men have
the same elasticities as married men – a consistent finding in the literature (McClelland and Mok

17
   Another difference from Corinth et al. (2021) is that we apply population average elasticities by gender
and marital status, whereas they apply different elasticities for single mothers claiming the EITC versus
everyone else. In principle, applying separate estimates for the two groups can yield more accurate
results, but in practice there are few elasticity estimates available in the empirical literature for non-EITC
claimants. And applying an EITC-specific estimate to EITC claimants while applying an overall
population estimate to non-EITC claimants would bias upwards the behavioral response by effectively
double-counting the contribution of EITC claimants.
18
   Specifically, we take the midpoint of the estimated elasticity range described in Lin and Ton (2017),
Table 4.
                                                      18
(2012)). This approach yields intensive margin elasticities of 0.145 for married mothers and
0.055 for others (fathers and single mothers), and extensive margin elasticities of 0.09 for
married mothers and 0.015 for others (fathers and single mothers).

The results of these analyses are presented in Columns 3 through 6 of Table 2. Both approaches
lead to slightly larger estimates of the fiscal cost of the expansions, driven by larger cost
contributions from labor supply adjustments. They also suggest a slightly larger reduction in
employment than our baseline estimates; for example, under the ARPA expansion, these
estimates suggest that 386,000 to 675,000 parents would choose to leave the labor force
(approximately 0.7% to 1.2% of currently working parents).19



     C. Alternative Future Earning Assumptions

Column 7 of Table 2 reports the effect of adopting more conservative assumptions about the
income ranges for which additional household income increases children’s future earnings.
Recall that in our baseline approach, following Garfinkel et al. (2021), we applied the full effect
from the literature for households earning less than $50,000, 50% of the effect for households
earning between $50,000 and $100,000, and no effect for households earning above $100,000. In
contrast, Column 7 reports the results of assuming the full effect is limited to households earning
less than $25,000, that households earning $25,000-$50,000 receive 50% of the effect, and that
there is no effect on future earnings for children living in households with income above
$50,000. This change in assumptions reduces the revenue effect from the change in long-term
earnings across all three policies, but only modestly for the fully refundable CTC, since the
additional income from that reform is concentrated among the lowest income households.

Finally, Column 8 of Table 2 applies a discount rate of 3% rather than 2% to the revenue
generated from children’s future earnings. As expected, this change modestly reduces the present
discounted value of the future revenue, increasing the net fiscal cost of the reforms.



     VII.    Administering an Expanded CTC

Questions remain about the best way to deliver the CTC (monthly, quarterly, or annually).
Providing the benefit on a monthly basis would theoretically reduce income instability, which is
associated with worse outcomes on a variety of dimensions for children including undermining
health and food security (Wolf and Morrissey 2017) and lowered education attainment,
particularly among moderate income families (Hardy 2014). Delivering the credit throughout the
year – rather than at tax time, could allow the credit to be used to smooth income over the year,
reducing extreme hardship. The IRS has limited experience delivering monthly tax credit
payments.



19
  For the mid-point estimate of the extensive margin substitution elasticities from McClelland and Mok (Column 4
of Table 2), we predict that about 528,000 parents will exit the labor force.

                                                       19
More challenging than delivering the payments on a monthly (or even quarterly) basis would be
ensuring that the benefit is paid to the adult(s) actively caring for the child. Traditionally, the tax
system has operated on an annual reconciliation process after the tax year has ended and only
one person may claim the CTC for a given child, even if multiple people care for that child.20
Reconciling the correct caregiver on a monthly (or quarterly) basis would better ensure that the
credit is delivered to the adult who is actively caring for the child – rather than the adult who
cared for the child for the majority of the year—and presumably would be more likely to directly
benefit the child. Splitting the credit across multiple tax units could affect a substantial number
of families with children. In a recent survey of families with children with incomes less than
$150,000, approximately 80 percent reported claiming a CTC (IPSOS 2021). The most common
reason reported for not claiming a CTC (42 percent) was that another parent had claimed the
payment instead. As children move between homes, particularly those who have low income, a
way to split benefits among caregivers may need to be developed to make the CTC most
effective.

Although family benefits are not generally split between multiple caregivers, there are other
examples of splitting tax benefits among multiple households. Mortgage interest, for example,
can be apportioned to multiple people who hold one mortgage. A system for splitting the CTC
might allow the credit to better support children who are moving between households. Prior
research shows that almost 6 percent of children overall change the adult with whom they live
over the course of the year – a share that has been rising. Much higher levels of moving among
households are reported among families with at least one child not biologically related to the
parent(s) in the household (Maag, Peters, and Edelstein 2016). An important challenge for the
IRS in administering a system that accounts for children moving households within a year is that
the agency lacks administrative data documenting where children live or the identity of their
primary caretaker, indeed, no government or private entity has these data.

In addition to creating rules for which taxpayers can claim which children for purposes of the
credit, another administrative challenge for the IRS is to increase take-up of the CTC. No
program has perfect take-up. Estimates of the CTC are not widely available, though a large share
of families benefit from the program (90 percent in a typical year, 92 percent under the 2021
rules). Those least likely to participate are those who are not required to file a tax return
(generally very low-income families who have not benefited much or at all from the CTC in
years prior to 2021). To have the largest effect on poverty and other measures of well-being,
increasing take-up of the CTC is essential.



      VIII.   Conclusion

In this paper we estimated the net fiscal cost of various CTC expansions. Across expansions, the
direct fiscal cost is by far the largest cost component. Labor supply responses of the households
receiving an additional benefit from the expansion constitute a relatively small percentage of the
net fiscal cost of any of the expansions we examined. We estimate larger fiscal savings stemming


20
     IRS Publication 501.
                                                  20
from increased future earnings of low-income children receiving larger CTC benefits because of
the reform.

Estimates such as the ones reported here necessarily rely on a large number of assumptions
relating to behavioral responses and other parameters, and in some cases we’ve relied on some
simplifying assumptions. For example, we’ve assumed full take-up of the credit. Incomplete
take-up is an important feature of transfer programs like the CTC, but we are not aware of take-
up estimates for the CTC itself. We expect take-up to be relatively high under any of the
expansions we consider, similar to the high take-up rates estimated for the EITC among
households with children (e.g. Jones 2013). Along the same lines, we have not modeled the fiscal
costs or savings from changes in the degree to which ineligible taxpayers claim the CTC despite
lacking eligibility for it.

A second simplifying assumption in our analysis is that we have not modeled contributions to net
fiscal costs arising from safety net programs outside of the tax code. Although CTC benefits do
not directly affect eligibility for other social safety net programs, they can affect eligibility and
participation indirectly by shaping labor supply. In particular, reductions in labor supply could
increase fiscal costs associated with means-tested safety net programs whereas the long-term
increases in child earnings could reduce the fiscal cost of such programs. Similarly, by
improving low-income households’ financial well-being, they may reduce participation in such
programs, reducing these programs’ costs.

Finally, in estimating how CTC reforms would affect the fiscal costs, we have focused on
changes in labor supply – those of the taxpayers claiming the child and those of the children
themselves via changes in human capital accumulation and earnings. At the same time, however,
there are reasons to expect CTC expansions to reduce medium- and long-term federal spending
through other mechanisms as well, such as through less crime (Akee et al. 2010), fewer children
entering the foster care system (Berger et al. 2017), and reduced medical spending due to
recipients’ improved health (e.g., Hoynes et al. 2016; Evans & Garthwaite 2014; East 2018).
Garfinkel et al. (2021) undertakes a fuller accounting of these additional channels; their analysis
suggests that incorporating these additional channels would further reduce the net fiscal cost of
expanding the CTC.




                                                 21
References

Aizer, A., Eli, S., Ferrie, J., & Lleras-Muney, A. (2016). The Long-Run impact of cash transfers
       to poor families. American Economic Review, 106(4), 935–971.
       https://doi.org/10.1257/aer.20140529

Akee, R.K., Copeland, W.E., Keeler, G., Angold, A. and Costello, E.J., 2010. Parents' incomes
       and children's outcomes: a quasi-experiment using transfer payments from casino
       profits. American Economic Journal: Applied Economics, 2(1), pp.86-115.

Averett, S. and Wang, Y., 2013. The effects of earned income tax credit payment expansion on
       maternal smoking. Health Economics, 22(11), pp.1344-1359.

Bailey, M.J., Hoynes, H.W., Rossin-Slater, M. and Walker, R., 2020a. Is the social safety net a
        long-term investment? Large-scale evidence from the food stamps program (No.
        w26942). National Bureau of Economic Research.

Bailey, M.J., Timpe, B.D. and Sun, S., 2020b. Prep School for poor kids: The long-run impacts
        of Head Start on Human capital and economic self-sufficiency (No. w28268). National
        Bureau of Economic Research.

Baker, M., Messacar, D. and Stabile, M., 2021. The Effects of Child Tax Benefits on Poverty and
       Labor Supply: Evidence from the Canada Child Benefit and Universal Child Care
       Benefit (No. w28556). National Bureau of Economic Research.

Bastian, J. and Michelmore, K., 2018. The long-term impact of the earned income tax credit on
       children’s education and employment outcomes. Journal of Labor Economics, 36(4),
       pp.1127-1163.

Baughman, R.A., 2012. The effects of state EITC expansion on children’s health. Carsey
      Institute Brief No. 48

Berger, L., Font, S., Slack K., & Waldfogel J., 2016. Income and child maltreatment in
       unmarried families: evidence from the earned income tax credit. Review of Economics of
       the Household, 15, pp.1345-1372.

Blundell, R. and MaCurdy, T., 1999. Labor supply: A review of alternative
      approaches. Handbook of labor economics, 3, pp.1559-1695.

Braga, B., Blavin, F. and Gangopadhyaya, A., 2020. The long-term effects of childhood
       exposure to the earned income tax credit on health outcomes. Journal of Public
       Economics, 190, p.104249.

Burman, Leonard E. and Laura Wheaton. 2005. “Who Gets the Child Tax Credit?”, The Urban
      Institute, Washington, DC. October 3.



                                               22
Burman, Leonard E. 2021. “A Universal EITC: Making Work Pay in the Age of Automation”,
      The Urban Institute, Washington, DC. January 21.
      https://www.taxpolicycenter.org/author/leonard-e-burman/publications

Chetty, R., Guren, A., Manoli, D. and Weber, A., 2013. Does indivisible labor explain the
       difference between micro and macro elasticities? A meta-analysis of extensive margin
       elasticities. NBER macroeconomics Annual, 27(1), pp.1-56.

Collyer, S., Harris, D., & Wimer, C. 2019. Left behind: The one-third of children in families who
       earn too little to get the full child tax credit. Columbia University Poverty and Social
       Policy Brief 3(6).

Corinth, K., Meyer, B., Stadnicki, M., and Wu, Derek. The Anti-Poverty, Targeting, and Labor
       Supply Effects of the Proposed Child Tax Credit Expansion. Working Paper (2021).

Cunha, F. and Heckman, J., 2007. The technology of skill formation. American Economic
       Review, 97(2), pp.31-47.

Duncan, G.J., Ludwig, J. and Magnuson, K.A., 2007. Reducing poverty through preschool
      interventions. The Future of Children, pp.143-160.

Duncan, G.J. and Sojourner, A.J., 2013. Can intensive early childhood intervention programs
      eliminate income-based cognitive and achievement gaps? Journal of Human
      Resources, 48(4), pp.945-968.

East, C., 2018. The Effect of Food Stamps on Children's Health: Evidence from Immigrants'
       Changing Eligibility. Journal of Human Resources, 55(2), pp.387-427.

Eissa, N. and Hoynes, H.W., 2004. Taxes and the labor market participation of married couples:
        the earned income tax credit. Journal of public Economics, 88(9-10), pp.1931-1958.

Evans, W.N. and Garthwaite, C.L., 2014. Giving mom a break: The impact of higher EITC
       payments on maternal health. American Economic Journal: Economic Policy, 6(2),
       pp.258-90.

Garfinkel, I., Sariscany, L., Ananat, E., Collyer, S., Wimer, C. 2021. The costs and benefits of a
       child allowance. Poverty and Social Policy Discussion paper.


Goldin, J. and Michelmore, K., 2020. Who Benefits From the Child Tax Credit? (No. w27940).
       National Bureau of Economic Research.

Greenstein, Robert, Elaine Maag, Chye-Ching Huang, Emily Horton, and Chloe Cho. 2018.
       “Improving the Child Tax Credit for Very Low-Income Families”, The Urban Institute,
       Washington, DC. April. https://www.mobilitypartnership.org/improving-child-tax-credit-
       very-low-income-families

                                                23
Hardy, Bradley L. 2014. “Childhood Income Volatility and Adult Outcomes”, Demography.
       October.

Holt, Steve. 2015. Periodic Payment of the Earned Income Tax Credit Revisited. Metropolitan
        Policy Program. Brookings Institution, Washington, DC. December.

Holtzblatt, Janet. 2021. Will Biden’s Plan to Reduce the Tax Gap by Increasing the IRS’s Budget
       Raise $300 Billion? TaxVox. Tax Policy Center, Washington, DC. May 3.

Hotz, J., & Scholz, J. 2003. The Earned Income Tax Credit. In Means-Tested Transfer Programs
        in the United States. University of Chicago Press.

Hoynes, H., Miller, D. and Simon, D., 2015. Income, the earned income tax credit, and infant
      health. American Economic Journal: Economic Policy, 7(1), pp.172-211.

Hoynes, H., Schanzenbach, D. W., & Almond, D. (2016). Long-Run impacts of childhood access
      to the safety net. American Economic Review, 106(4), 903–934.
      https://doi.org/10.1257/aer.20130375


IRS (Internal Revenue Service). 2014. “Compliance Estimates for the Earned Income Tax Credit
       Claimed on 2006 – 2008 Returns.” Washington, DC: The White House.

Jones, D. and Marinescu, I., 2018. The labor market impacts of universal and permanent cash
       transfers: Evidence from the Alaska Permanent Fund (No. w24312). National Bureau of
       Economic Research.

Jones, M.R., 2013, November. Changes in EITC eligibility and participation: 2005–2009. In Tax
       Administration at the Centennial: An IRS-TPC Research Conference (pp. 45-61).

LaLumia, S., 2009. The earned income tax credit and reported self-employment
     income. National Tax Journal, 62(2), pp.191-217.

Lin, E.Y. and Tong, P.K., 2017. Married couple work participation and earnings elasticities:
       evidence from tax data. International Tax and Public Finance, 24(6), pp.997-1025.

Lippold, Kye, 2019. The Effects of the Child Tax Credit on Labor Supply. Working paper,
       online at https://ssrn.com/abstract=3543751

Maag, Elaine. 2019. “Shifting Child Tax Benefits in the TCJA Left Most Families About the
      Same”. The Urban Institute, Washington, DC. August 16.
      https://www.taxpolicycenter.org/publications/shifting-child-tax-benefits-tcja-left-most-
      families-about-same

Maag, Elaine, James R. Nunns, and Hang Nguyen. 2016. “An Option to Reform the Income Tax

                                               24
       Treatment of Work and Families”, The Urban Institute, Washington, DC. December 5.
       https://www.taxpolicycenter.org/publications/option-reform-income-tax-treatment-
       families-and-work

Maag, Elaine, H. Elizabeth Peters, and Sara Edelstein. 2016. “Increasing Family Complexity and
      Volatility: The Difficulty in Determining Child Tax Benefits”, Urban-Brookings Tax
      Policy Center, Washington, DC. March 3.

McClelland, R. and Mok, S., 2012. A review of recent research on labor supply elasticities.
      Congressional Budget Office Working Paper Series.

Michelmore, K. and Lopoo, L.M., forthcoming. The effects of EITC exposure in childhood on
      marriage and early childbearing. Demography.

Milligan and Stabile. 2009. Child Benefits, Maternal Employment, and Children's Health:
       Evidence from Canadian Child Benefit Expansions. The American Economic Review Vol.
       99, No. 2, Papers and Proceedings of the One Hundred Twenty-First Meeting of the
       American Economic Association (May, 2009), pp. 128-132 (5 pages)

National Academies of Sciences Engineering and Medicine. (2019). A Roadmap to Reducing
       Child Poverty. In A Roadmap to Reducing Child Poverty. The National Academies Press.
       https://doi.org/10.17226/25246

National Taxpayer Advocate, 2016. Annual Report to Congress.
       https://www.taxpayeradvocate.irs.gov/reports/2016-annual-report-to-congress/full-report/

Nichols, A. and Rothstein, J., 2016. 2. The Earned Income Tax Credit (pp. 137-218). University
       of Chicago Press.

Rachidi, A. 2021. “How would a child allowance affect employment?”, American Enterprise
       Institute. https://www.aei.org/poverty-studies/how-would-a-child-allowance-affect-
       employment/

Wamhoff, S., & Gardner, M. (2019). Who pays taxes in America in 2019? Institute on Taxation
     and Economic Policy. https://itep.org/who-pays-taxes-in-america-in-2019/

West, S., Castro Baker, A., Samra, S., and Coltera, E. (2021). “Preliminary Analysis: SEED’s
       First Year”. Stockton Economic Empowerment Demonstration.
       https://static1.squarespace.com/static/6039d612b17d055cac14070f/t/603ef1194c474b329
       f33c329/1614737690661/SEED_Preliminary+Analysis-
       SEEDs+First+Year_Final+Report_Individual+Pages+-2.pdf

Wikle, J. and Wilson, R., 2020. Access to Head Start and Maternal Labor Supply: Experimental
       and Quasi-experimental Evidence. Manuscript. Brigham Young University, Provo, UT.




                                               25
Wolf, Sharon and Morrissey, Taryn, “Economic Instability, Food Insecurity, and Child Health in
       the Wake of the Great Recession”, Social Service Review Vol 91(3).




                                             26
Table 1. Fiscal Cost of Child Tax Credit Expansion Policies

                                                                                                                  Cost Components
                                             Policy Elements
                                                                                                                    ($ Billions)
                                                                                               (1)                 (2)                  (3)              (4)
                                                  Increased Max          No Income                                                Long-Term
                        Fully Refundable                                                     Direct         Labor Supply                               Total
                                                      Credit             Phase-Out                                                 Earning
                                                                                             Costs             Effects                                 Cost
                                                                                                                                    Effects
Fully
                                 X                                                           25.49                0.74                -4.99            21.25
Refundable CTC

ARPA
                                 X                        X                                  96.36                4.37               -15.50            85.23
(2021 CTC)


Universal CTC                    X                        X                    X            108.86                1.73               -15.50            95.09




Notes: The table presents estimated cost components for various expansions of the CTC. Direct costs (Column 1) represent the additional CTC benefit that
taxpayers receive. Labor supply effects (Column 2) represent the effect on tax revenue of labor supply changes induced from a change in marginal tax rate or
the additional income provided under the expansion. Long-term labor effects (Column 3) represent the effect on tax revenue from changes in adulthood
earnings of children growing up in tax units that receive additional benefits because of the expansion. The total cost of the expansion (Column 4) is calculated
by summing Columns 1-3. The universal CTC is a flat benefit amount of $3,600 per child under the age of 6 and $3,000 per older child under the age of 17.
The 2021 CTC is modeled after the CTC design in place for tax year 2021; it differs from the Universal CTC in that the benefits phase out by income. The
Fully Refundable CTC is modeled after the CTC design that President Biden proposed for tax years following 2025; it is fully refundable, phases out by
income, and has the same maximum credit amount as the pre-2021 CTC. All calculations are performed using survey weights from the 2018 ASEC
supplement to the CPS.




                                                                              27
 Table 2. Alternative Assumptions for Estimating Fiscal Cost (in billions) of CTC Expansion


                    Labor Supply Effect
                                                                                                                                         Long-Term
                         (Alternate                                     Labor Supply Effect
                                                                                                                                        Labor Effects
                        Elasticities)                                   (Separate Margins)
                      (1)           (2)              (3)               (4)            (5)                     (6)                 (7)                   (8)

                   Elasticity    Elasticity                McClelland and Mok (2012)                                          Alternate
                                                                                                       Lin and Tong                              Alternate
                    Lower         Upper                                                                                        Benefit
                                                                                                          (2017)                               Discount Rate
                    Bound         Bound                                                                                       Multiplier
                                                   Lower           Midpoint             Upper

 Fully
 Refundable           0.61          0.80            0.03              1.73               3.28                0.70                -4.39              -4.48
 CTC
 ARPA
                      2.33          9.53            2.20              4.26               6.27                2.88                -9.57              -13.92
 (2021 CTC)
 Universal
                      0.48          6.69            1.53              3.11               4.69                2.22                -9.57              -13.92
 CTC

Notes: The table reports the estimated fiscal cost of various CTC expansions under alternative assumptions. The cost components and CTC expansions
considered are described in the notes to Table 1. Columns 1-6 report the change in federal tax revenue due to labor supply changes using elasticities from the
inelastic (Column 1) or elastic (Column 2) range of estimates described in Section VI.A, or from separately modeling the intensive and extensive margins of the
substitution effect based on the methodology described in Section VI.B and the lower bound (Column 3), midpoint (Column 4), and upper bound (Column 5) of
the corresponding elasticity estimates from McClelland and Mok (2012). Column 6 reports the change in tax revenue due to labor supply changes using more
recent estimates of substitution elasticities from 2000-2009 from Lin and Tong (2017). Columns 7 and 8 report the change in federal tax revenue from children's
future earnings, under the assumption that the effect of additional household income on children’s future earnings fades more quickly as a household’s income
rises (Column 7) or using a 3% discount rate for the additional revenue from increases in children’s future earnings (Column 8).




                                                                              28
Figure 1. Comparison of Child Tax Credit Designs




Notes: The figure shows the benefit a tax unit would receive from the pre-2021 CTC (solid blue line), the Fully
Refundable CTC (dashed green line), the ARPA CTC (dashed orange line), and the universal CTC (dashed red line).
Each tax unit is assumed to consist of a single parent with one child (Panel A) or two children (Panel B) between the
ages of 6-17, to have income consisting solely of salary compensation, and to claim the standard deduction.




                                                         29
Figure 2. Average Change in Benefits for Fully Refundable CTC by Income




Notes: The figure shows the average change in CTC benefits by binned tax unit income of a change from the pre-
2021 CTC to the Fully Refundable CTC. Each tax unit is assumed to claim the standard deduction.




                                                       30
Figure 3. Average Tax Change by Income from Fully Refundable CTC

         Panel A. Income Effect




         Panel B. Substitution Effect




Notes: The figure shows the average change in tax liability by binned tax unit income of a change from the pre-2021
CTC to the Fully Refundable CTC due to the income effect (Panel A) and substitution effect (Panel B). Each tax
unit is assumed to claim the standard deduction. In Panel A, each individual in a tax unit is assumed to adjust their
income based on (1) the relevant income elasticity estimate based on the individual’s gender and marital status, and
(2) the percent change in household income that the expanded CTC represents. The intensive and extensive margin
income effects are separately calculated, as described in the text. In Panel B, each individual in a tax unit is assumed
to adjust their earnings based on (1) the relevant wage-elasticity estimate based on the individual’s gender and
marital status, and (2) the percent change in the marginal net-of-tax rate induced from the CTC expansion. In both
panels, the change in tax revenue is calculated by comparing TaxSim’s federal tax liability estimate for the tax unit

                                                          31
using the tax unit’s pre-reform earnings with estimated federal tax liability after imposing the predicted changes in
earnings based on the specified effect.




                                                          32
Figure 4. Marginal Tax Rate Change from Fully Refundable CTC Reform by Income for
Single Filer with One Child




Notes: The figure shows the change in the marginal tax rate a tax unit would face from a change from the pre-2021
CTC to the Fully Refundable CTC, by earnings. Each tax unit is assumed to consist of a single parent with one child
between the ages of 6-17, to have income consisting solely of salary compensation, and to claim the standard
deduction. The marginal tax rate under the pre-2021 CTC is calculated from TaxSim. The marginal tax rate under
the Fully Refundable CTC is calculated as described in the text.




                                                        33
Figure 5. Average Earnings Change by Income from Fully Refundable CTC




Notes: The figure shows the average estimated change in earnings by binned tax unit income of a change from the
pre-2021 CTC to the fully refundable CTC due to the income and substitution effects. In calculating the income
effect, each individual in a tax unit is assumed to adjust their income based on (1) the relevant income elasticity
estimate based on the individual’s gender and marital status, and (2) the percent change in household income that the
expanded CTC represents. The intensive and extensive margin income effects are separately calculated, as described
in the text. In calculating the substitution effect, each individual in a tax unit is assumed to adjust their earnings
based on (1) the relevant wage-elasticity estimate based on the individual’s gender and marital status, and (2) the
percent change in the marginal net-of-tax rate induced from the CTC expansion. For married tax units, the displayed
change in earnings is the sum of the change in earnings of each individual spouse.




                                                         34
Figure 6. Average Change in EITC Benefits by Income from Fully Refundable CTC




Notes: The figure shows the average change in EITC benefits by binned tax unit income from a change from the pre-
2021 CTC to the fully refundable CTC due to the income and substitution effects. In calculating the income effect,
each individual in a tax unit is assumed to adjust their income based on (1) the relevant income elasticity estimate
based on the individual’s gender and marital status, and (2) the percent change in household income that the
expanded CTC represents. The intensive and extensive margin income effects are separately calculated, as described
in the text. In calculating the substitution effect, each individual in a tax unit is assumed to adjust their earnings
based on (1) the relevant wage-elasticity estimate based on the individual’s gender and marital status, and (2) the
percent change in the marginal net-of-tax rate induced from the CTC expansion. The change in EITC is calculated
by comparing TaxSim’s EITC estimate for the tax unit using the tax unit’s pre-reform earnings with estimated EITC
after imposing the predicted changes in earnings.




                                                         35


File and source

File

estimating-the-net-fiscal-cost-of-a-child-tax-credit-expansion.pdf

Size

714,969 bytes

SHA-256

0d9c8fc4e09a0e7e27b10f014b1f5a5f06b470782bc1f408250d62182bf2fe30

Our copy

estimating-the-net-fiscal-cost-of-a-child-tax-credit-expansion.pdf

Original

www.nber.org