The Short-Term Labor Supply Response to the Expanded Child Tax Credit (NBER Working Paper 31110)

Summary

NBER Working Paper No. 31110, The Short-Term Labor Supply Response to the Expanded Child Tax Credit, by Brandon Enriquez, Damon Jones and Ernest V. Tedeschi, dated April 2023. The abstract states that the authors estimate the extensive and intensive margin labor supply response to the monthly Child Tax Credit disbursed in 2021 as a part of the American Rescue Plan Act, using Current Population Survey microdata. It states that they do not find strong evidence of a change in labor supply for families receiving the credit and that the results are robust to alternative labor supply models.

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

Issuer: Source documents
Document type: The Short Term Labor Supply Response To The Expanded Ctc
Case: The Short Term Labor Supply Response To The Expanded Ctc

Cite this document

“The Short-Term Labor Supply Response to the Expanded Child Tax Credit (NBER Working Paper 31110).” Original at www.nber.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/the-short-term-labor-supply-response-to-the-expanded-ctc/, SHA-256 cf70a7cdb54e95fcd941d7c68dbcda3a46a220fff6049267a96807303d52600f.

Full text

NBER WORKING PAPER SERIES

THE SHORT-TERM LABOR SUPPLY RESPONSE TO
THE EXPANDED CHILD TAX CREDIT

Brandon Enriquez
Damon Jones
Ernest V. Tedeschi

Working Paper 31110
http://www.nber.org/papers/w31110

NATIONAL BUREAU OF ECONOMIC RESEARCH
1050 Massachusetts Avenue
Cambridge, MA 02138
April 2023

We are thankful for comments from participants at the 2023 ASSA Annual Conference, and for
helpful feedback from Bruce Meyer. 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 They have not been
peer-reviewed or been subject to the review by the NBER Board of Directors that accompanies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.

© 2023 by Brandon Enriquez, Damon Jones, and Ernest V. Tedeschi. 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.
The Short-Term Labor Supply Response to the Expanded Child Tax Credit
Brandon Enriquez, Damon Jones, and Ernest V. Tedeschi
NBER Working Paper No. 31110
April 2023
JEL No. D63,H24,J22

ABSTRACT

We estimate the extensive and intensive margin labor supply response to the monthly Child Tax
Credit disbursed in 2021 as a part of the American Rescue Plan Act. Using Current Population
Survey microdata, we compare labor supply outcomes among households who qualify for varying
relative increases in household income, as a result of their income level and household size. We
do not find strong evidence of a change in labor supply for families receiving the credit. The
results are robust to alternative labor supply models, where households respond mainly to cash on
hand or changes in the annual budget set.

Brandon Enriquez Ernest V. Tedeschi
Massachusetts Institute of Technology ernie.tedeschi@gmail.com
enriquez@mit.edu

Damon Jones
Harris School of Public Policy
University of Chicago
1155 East 60th Street
Chicago, IL 60637
and NBER
damonjones@uchicago.edu
1 Introduction

A general feature of tax and transfer programs is some trade-off between the value of re-
distribution, the fiscal costs of these transfers, and potential economic distortions. These
inputs, along with subjective decisions regarding the relative weights placed on the wellbeing
of different people or groups, and normative preferences over intermediate outcomes, such
as labor supply, and/or the form of transfers, e.g. in-kind or cash, are required to carry out
policy evaluations. In this study, we estimate the short-term labor supply response to cash
transfers for low- and middle-income households, in the form of a fully-refundable child tax
credit (CTC).
The 1990s featured a marked shift from so-called “traditional” welfare programs—such
as AFDC, which featured cash transfers as a part of the safety net and potential work
disincentives—to cash transfers that required recipients to work and have positive earnings,
such as the Earned Income Tax Credit (EITC) and the Child Tax Credit (CTC). The contrast
between these approaches is related to debates about two features of a nonlinear tax and
transfer system: (1) the amount of transfers given to households with zero income, e.g.
a guaranteed income, and (2) whether there should be negative marginal tax rates—also
known as wage subsidies or “phase-ins”—along some portion of the tax schedule (see Saez,
2002; Rothstein, 2010, for example).
Federal policy in 2021 offered an opportunity to study these alternative models, as the
American Rescue Plan temporarily moved the CTC from a policy with a work requirement
and phase-in to a fully-refundable design, extending eligibility to families with no earnings.
We compare labor market outcomes—labor force participation and hours worked—for fami-
lies who qualified for smaller and larger CTC transfers, before and after the credit was paid
out. Using a difference-in-differences and triple difference approach, we do not detect signif-
icant labor supply differences in response to variation in the size of the CTC. When framed
as an increase in cash-on-hand, our confidence intervals rule out a Labor force participation
rate decline of 0.3 percentage points in response to a 10 percentile increase in CTC relative

1
to income, or when cast as a change in the return to entering employment, we rule out an
extensive labor supply elasticity of 0.005.

2 Background

Originally established in 1997, the Child Tax Credit (CTC) reduced tax liability by $500
per child for families with children under 17 years of age. The tax credit was nonrefundable
at its inception—meaning filers had to have federal income tax liability to claim it—but a
portion became refundable after the Economic Growth and Tax Relief Reconciliation Act
of 2001 (EGTRRA). Families could receive a credit for initially 10 percent, and later 15
percent, of earnings above a minimum threshold. In 2019, total spending on the CTC was
$118 billion, comparable to the total federal amount spent on children within Medicaid and
the Child Health Insurance Program, combined (Hahn, Lou and Isaacs, 2020).
The last major, pre-pandemic change to the CTC occured with the Tax Cuts and Jobs
Act of 2018 (TCJA). The TCJA doubled the maximum credit to $2,000 per child and lowered
the minimum earnings threshold to $2,500. Lower-income households with no tax liability
however could receive at most $1,400 per child. The CTC under TCJA law phased out for
single-parent households at an income of $200,000 (which the TCJA raised from $75,000),
and for married households at $400,000 (raised from $110,000). As a result, nearly one-third
of children lived in households that did not receive the CTC, with the share among Black
and Latino children at one-half and virtually no children in the bottom decile of incomes
qualifying for the credit (Goldin and Michelmore, 2020).
In 2021, the CTC was temporarily re-designed as part of the American Rescue Plan
(ARP). The ARP CTC allowed for full refundability of the credit, even for families with
zero earnings. In addition, the maximum benefit per child increased to $3,000 for children
between the ages of 6 and 17, and to $3,600 for children ages 0 to 5, for single filers earning
below $75,000 and married filers earning below $150,000. Above those thresholds, the CTC

2
phased back down to the TCJA-$2,000 level, which then phased-out again per pre-2021 law
at $200,000 fir single filders and $400,000 for married filers. In addition, advance, monthly
payments were made to families beginning in July of 2021, with the remainder applied to
one’s tax balance or refund during early 2022. These new ARP parameters were put in
place for only one year, although there was discussion of potentially instituting them more
permanently. This version of the CTC was more akin to so-called Child Allowance policies
of Australia, Ireland, or Canada.
In a standard model of labor supply, the effect of this policy change on the annual
budget generates two forces. First, the policy removes the phase-in of the credit, which
had previously created an incentive to work through substitution effects, all things equal. In
other words, a negative marginal tax rate, or wage subsidy, was replaced with a more neutral
implicit marginal tax rate of zero for low-income households. In addition, the increase in the
level of the credit generates an income effect, reducing labor supply when leisure is a normal
good. A secondary impact of the policy was to increase effective marginal tax rates over the
range where the credit was phased back down from $3,000 or $3,600 to $2,000.
Because the CTC is applied to annual earnings or income, the preceding incentives op-
erate at the level of the annual budget set. In that case, the timing of the payment of the
advance monthly CTC within the year under the ARP regime is immaterial. However, for
households who are credit-constrained, the more immediate impact of the CTC may have
operated through its effect on monthly income, as the advance began to be paid out in July.
At that frequency, the policy would generally produce an income effect by increasing liq-
uidity for cash-constrained households. We may therefore expect to observe a labor supply
response following the initial advance payments in July. In either case, the standard model
predicts some reduction in labor supply in response to these changes. However, in a search
model, more cash on hand could lead to higher quality (higher wage) matches, as discussed in
Nekoei and Weber (2017), or, as described in Autor, Dube and McGrew (2023), the benefits
could raise reservation wages and increase “job shopping.” In theory, the net effect of this

3
dynamic on employment is ambiguous if workers remain at higher-paying jobs longer.
Prior studies have estimated the impact of the 2021 extension to the CTC on labor supply.
Our study is closest to that of Ananat et al. (2021, 2022), who also used monthly CPS data
in one of their specifications to look at labor supply before and after the onset of advance
payments. Our findings generally replicate theirs of no significant labor supply response. We
extend the analysis by further looking at whether there are labor supply changes when the
policy expires at the end of 2021, and by using data from 2019 to conduct additional placebo
analysis and to flexibly control for group-specific seasonal trends. A number of other studies,
using different sources of data, some admiinstrative and some survey-based, similarly fail to
find significant labor supply responses (Roll, Hamilton and Chun, 2022; Lourie et al., 2022;
Karpman et al., 2022; Pilkauskas et al., 2022). However, Han, Meyer and Sullivan (2022)
find slower employment growth among families with lower levels of education and children,
relative to similar families without children during the period when CTC payments were
made.
A related set of studies ask the more generally how a permanent and fully-refundable
CTC would effect employment. A report by a National Academies of Science, Engineering,
and Medicine panel estimated a reduction of 150,000 jobs, but only considered the income
effects of the CTC. Studies that also consider the substitution effects of the policy change
predict larger employment reductions, ranging from 296,000 (Brill, Pomerleau and Seiter,
2021) to 386,000 (Goldin, Maag and Michelmore, 2022) to between 358,000 and 411,000
(Bastian, 2022) to 1.5 million (Corinth et al., 2022). In contrasts to these studies, we focus
on the response to a temporary CTC change, which is likely to entail a smaller behavioral
response. These studies also rely on simulations, which require the authors to select a labor
supply elasticity, while our estimates use realized outcomes and allow us to remain agnostic
regarding that choice.

4
3 Data and Methods

Our analysis primarily draws on the basic monthly Current Population Survey (CPS), a
monthly survey of about 60,000 U.S. households with detail on demographics and labor
market outcomes. While the monthly CPS includes measures of wages for a subset of re-
spondents, it only contains one measure of total income (HEFAMINC). This family income
variable is coded as a categorical variable rather than a continuous one, with family income
over the prior 12 months classified among 16 ranges. CPS questionnaires show that the
concept of income for this variable is in principle Census money income: that is, CPS inter-
viewers are instructed to exclude tax credits (such as the CTC) and tax liabilities as well as
non-cash transfers when asking about family income.
To allow for a continuous measurement of income, we impute a continuous measure of
family income with the aid of the CPS Annual Social and Economic Supplement (ASEC).
The CPS ASEC is a once-a-year supplement, conducted for most respondents in March,
which asks about prior year components of income. We classify both our basic monthly
and ASEC CPS samples into demographic cells by marriage status (2 categories), number
of children (4 categories), and elderly status (2 categories). We then randomly draw, on a
weighted basis, continuous values of family income from the CPS ASEC conditional on an
individual’s discrete family income category and demographic categories in the basic monthly
CPS (see Han, Meyer and Sullivan, 2022, for example). We also draw the value of adjusted
gross income (AGI) and total earnings from the same CPS ASEC record that provides family
income. For basic monthly CPS samples in calendar years up to and including 2021, we draw
from the CPS ASEC sample corresponding to the same reference year. For example, the
2022 CPS ASEC refers to outcomes in calendar year 2021. For 2022 monthly data, we repeat
our use of the 2022 CPS ASEC, as the 2023 ASEC is not yet available.
We impute CTC eligibility using the drawn continuous values of AGI, reported number
of own-children in the monthly CPS, and the relevant tax parameters of the CTC. We do
not adjust the CTC level for incomplete take-up. Once we have calculated a CTC to income

5
ratio for each household, we order households in their percentile of this ratio, which serves
as our main regressor.
The outcomes of interest are (1) labor force participation, using the official labor force
definition of being employed, actively seeking a job, or being on temporary layoff/furlough;
and (2) actual hours worked unconditional on employment (non-employed individuals and
absent workers are coded as 0 hours worked), which provides a continuous measure of labor
utilization.
We estimate difference-in-differences (DD) specifications using data from before and after
the payment of the advance CTC in July 20211 :

yit = α + β0 CTˆ C i + β1 H2t + βDD CTˆ C i × H2t + βX Xi + εit (1)

where βDD is the coefficient of interest, H2t is an indicator for post-July observations, and
CTˆ C i is the percentile rank of the household’s CTC-to-income ratio. One potential con-
founder is that households who qualify for different levels of CTC may have different trends
in outcomes from the first half to the second half of the year, violating the parallel trends
assumption. We therefore, include data from 2019 and estimate a triple difference regression
(DDD) as follows:

yit = α0 + β0 CTˆ C i + β1 H2t + β3 CTˆ C i × H2t

+ 1{t ∈ 2021} × α1 + β4 CTˆ C i + β5 H2t +βDDD CTˆ C i × H2t + βX Xi + εit (2)

where βDDD is the coefficient of interest, and allows us to address differential seasonal trends
for those with smaller and larger CTC benefits. Finally, we re-estimate these models using
1
We also use regression estimates to calculate the implied labor supply elasticity from changing cash-on-
hand. In this specification, we estimate the main specifications with the change in return-to-work incentives
as the main regressor of interest. We define define change in return-to-work incentives as the negative of
the Tax Cuts and Jobs Act (TCJA) CTC benefit, divided by income. This definition reflects that the ARP
modified the CTC to be refundable, whereas it was non-refundable in the TCJA. See Ananat et al. (2022)
for further discussion.

6
data from before and after the end of 2021, to capture the impact of expiration of the
temporary extension to the CTC.
When households are not cash-constrained, the relevant incentives are the impact of
the change in CTC on the annual budget constraint, as measured by the return to work,
which is the change in after-tax income when moving from not working to earning a positive
amount. Under the TCJA CTC, one can only receive the CTC when earning a positive
amount, incentivizing entry. Under the ARP CTC, the CTC is received at low levels of
income regardless of one’s earnings level, removing this incentive. As explained in Corinth
et al. (2022), the reduction in this incentive for a household considering earning z will be:

[z − TARP (z) − TARP (0)] [z − TT CJA (z) − TT CJA (0)]
△(Return to Work) = −
| z
{z } | z
{z }
Return to work under ARP Return to work under TCJA

=0 =0
z }| { z }| { (3)
[CT CARP (z) − CT CARP (0)] [CT CT CJA (z) − CT CT CJA (0)]
= −
z z

CT CT CJA (z)
=−
z

where T is the tax or transfer under either the ARP or TCJA. In the second line, the level
of earnings and other features of the tax code generally cancel out, leaving just the change
in the CTC. Furthermore, with no phase in, the ARP CTC, at lower levels of earnings, is
unchanged upon entering. We therefore regress labor force participation on this measure,
using either our DD or DDD specifications above, and rescale to interpret as an elasticity.

4 Results

Figure 1 displays a binscatter of labor force participation on our main regressor of interest,
percentiles of CTC-to-income ratios. Panel A displays this for five months pre-CTC extension

7
(February-June 2021) and five months post-CTC extension (August 2021-December 2021).
This figure shows that while labor force participation does decline for higher levels of the
CTC, this pattern is nearly identical pre- and post-CTC extension. The figure suggests that
the CTC did not reduce labor force participation.2
The same conclusion is echoed in Figure 1, Panel B, which displays a binscatter for the
five months of extended CTC in 2021 and three months post-expiration (January-March
2022). Again, this figure shows the relationship between labor force participation and CTC
eligibility is nearly identical pre- and post-CTC expiration, suggesting the extended CTC
did not reduce labor force participation.3
Figure 2 tests for placebo effects. Panel A shows a binscatter of labor force participation
on percentiles of CTC-to-income ratios for five months before early in 2019 (February-June
2019) and five months later in 2019 (August 2019-December 2019), treating July of 2019 as
a placebo CTC rollout. Panel B displays the binscatter for the five months of the placebo
2019 CTC rollout and the three months of 2020 (January-March 2020), which mirror the
expiration of the CTC extension. As in Figure 1, both figures show little difference in
the relationship between LFPR and CTC eligibility, suggesting the estimated relationship
between LFPR and CTC eligibility was similar in 2019 and 2021.
Table 1, Panel A shows regression estimates of the effects of the effect of CTC-to-income
percentiles on labor force participation; we show these estimates for both the rollout and
the expiration.4 Theoretically, this effect is likely to be negative in both main specifications
(double difference and triple difference, columns (1), (2), (4), and (5)); if leisure is a normal
2
In addition, we imprecisely estimate a positive effect on employment and a negative effect on
unemployment—again failing to detect a reduction in labor supply.
3
One caveat to this analysis: the Bureau of Labor Statistics modified the CPS survey weights beginning
with the January 2022 CPS based on the results of the 2020 decennial Census. This population control
resulted in a slight increase in employment and labor force participation rates for some cohorts. To the
extent this affects our analysis, it will shift down the CTC eligibiliy-LFPR gradient in the post-extension
period. Given our null finding, under the assumption that mismeasurement is greatest for lower-income
households, this would imply that the CTC caused increased labor force participation - which again runs
opposite to the hypothesis that CTC disincentivized work.
4
For the expiration regressions, we use the post-expiration period as the base period, showing the coef-
ficient on the interaction term between the CTC-to-income percentiles and an indicator for the expanded
CTC time period, that is, August to December 2021.

8
good, the income effect of an increase in cash-on-hand could likely decrease labor supply.
Yet, across our main specifications, and when using either the rollout or the expiration of the
expanded CTC as the experiment of interest, we fail to detect a significant negative effect of
the expanded CTC on labor force participation. One of the four (double difference, column
(1)) is statistically significant and opposite-signed from our theoretical prediction. The
placebo estimates (columns (3) and (6)) are small and positive, but statistically significant.
That the placebo estimates are significant suggests there may be seasonal patterns in labor
force participation that vary by income level - for example, if seasonal work during the
December holiday season is concentrated among lowest wage workers. Such seasonal patterns
would also confound the double difference specification (columns (1) and (3)), and bias the
results against finding a negative income effect. Yet, when we run our triple difference
specification (columns (2) and (4)) to adjust for baseline seasonal differences, the coefficient
of interest is statistically indistinguishable from zero.
Table 1, Panel B shows regression estimates of the effects of changes in return-to-work
incentives on labor force participation. Theoretically, these coefficients should be positive.
And since the ARP CTC removed the income phase-in from the TCJA CTC, the net income
from earning a positive amount, relative to not working, was diminished. The substitution
effect implies the ARP’s effect on the annual budget set could potentially decrease labor
supply. Yet, across both main specifications (columns (1), (2), (4), and (5))—and, again,
when using either the rollout or the expiration of the expanded CTC as the experiment of
interest—we fail to find significant effects of a temporary change in return-to-work incentives
on labor force participation. Furthermore, the effects are comparable in magnitude to our
placebo estimates in columns (3) and (6), which are, as expected, insignificant.
To evaluate the magnitude of the implied labor supply response, Table 1, Panel C cal-
culates the elasticity of labor supply with respect to return-to-work incentives. To calculate
the largest labor supply elasticity that can be ruled out with our data, we calculate the
elasticity at the 95th percentile of the confidence interval for the regression estimates of the

9
effect of return-to-work incentives on labor force participation (estimated in Panel B). Our
estimates rule out a labor supply elasticity of 0.005; this suggests that a temporary 100
percent decrease in return to work generated by the TCJA CTC reduces labor supply by no
more than 0.5 percent.
Figure 3 reports results from our triple-difference specification (2) for labor force partici-
pation and for total hours worked, broken out by demographic subgroups: gender, education
level, number of kids, and race. The effect sizes are scaled by the standard deviation of the
outcome variable for comparability. Across demographic groups, we fail to find an effect of
the extended CTC on labor force participation.
For total hours worked, we likewise fail to find a significant impact on labor supply,
although the point estimates tend to be more positive than those derived from labor force
participation. Again, this pattern holds across different demographic subgroups.

5 Conclusion

The American Rescue Plan of 2021 extended the Child Tax Credit, by increasing the maxi-
mum benefit per child to $3,000-3,600 for the July 2021-December 2021 period and remov-
ing the work requirement to receive this money. This temporary policy provided a unique
opportunity to study the effect of swithing from a transfer with work requirements to an un-
conditional cash benefit on labor force participation. We fail to find effects of the expanded
benefit, and temporary removal of the work requirement, on labor force participation and a
total hours worked.

References
Ananat, Elizabeth, Benjamin Glasner, Christal Hamilton, and Zachary Parolin.
2021. “Evidence from Real-World Data from April to September 2021.” Center on Poverty
and Social Policy Discussion Paper.

10
Ananat, Elizabeth, Benjamin Glasner, Christal Hamilton, and Zachary Parolin.
2022. “Effects of the expanded Child Tax Credit on employment outcomes: Evidence from
real-world data from April to December 2021.” National Bureau of Economic Research.

Autor, David, Arindrajit Dube, and Annie McGrew. 2023. “The Unexpected Com-
pression: Competition at Work in the Low Wage Labor Market.” National Bureau of
Economic Research.

Bastian, Jacob. 2022. “How Would a Permanent 2021 Child Tax Credit Expansion Aect
Poverty and Employment?” Working Paper.

Brill, Alex, Kyle Pomerleau, and Grant M Seiter. 2021. “Estimating the Labor Supply
Response to a Permanent Child Tax Credit Expansion.” American Enterprise Institute.

Corinth, Kevin, Bruce D Meyer, Matthew Stadnicki, and Derek Wu. 2022. “The
Anti-poverty, Targeting, and Labor Supply Effects of Replacing a Child Tax Credit with
a Child Allowance.” NBER Working Paper, , (29366).

Goldin, Jacob, and Katherine Michelmore. 2020. “Who Benefits from the Child Tax
Credit?” National Bureau of Economic Research.

Goldin, Jacob, Elaine Maag, and Katherine Michelmore. 2022. “Estimating the Net
Fiscal Cost of a Child Tax Credit Expansion.” Tax Policy and the Economy, 36: 159–195.

Hahn, Heather, Cary Lou, and Julia B Isaacs. 2020. “How much does the federal
government spend on programs benefiting children?” Urban Institute.

Han, Jeehoon, Bruce D. Meyer, and James X. Sullivan. 2022. “Real-Time Poverty,
Material Well-Being, and the Child Tax Credit.” National Tax Journal, 75(4): 817–846.

Karpman, Michael, Elaine Maag, Stephen Zuckerman, and Doug Wissoker. 2022.
“Child Tax Credit Recipients Experienced a Larger Decline in Food Insecurity and a
Similar Change in Employment as Nonrecipients between 2020 and 2021.” Urban Institute.

Lourie, Ben, Devin M Shanthikumar, Terry J Shevlin, and Chenqi Zhu. 2022.
“Effects of the 2021 Expanded Child Tax Credit.” Available at SSRN 3990385.

Nekoei, Arash, and Andrea Weber. 2017. “Does extending unemployment benefits
improve job quality?” American Economic Review, 107(2): 527–561.

Pilkauskas, Natasha, Katherine Michelmore, Nicole Kovski, and H Luke Shaefer.
2022. “The effects of income on the economic wellbeing of families with low incomes: Evi-
dence from the 2021 expanded Child Tax Credit.” National Bureau of Economic Research.

Roll, Stephen, Leah Hamilton, and Yung Chun. 2022. “Expanded Child Tax Credit
payments have not reduced employment.” Social Policy Institute.

Rothstein, Jesse. 2010. “Is the EITC as good as an NIT? Conditional cash transfers and
tax incidence.” American economic Journal: economic policy, 2(1): 177–208.

11
Saez, Emmanuel. 2002. “Optimal income transfer programs: intensive versus extensive
labor supply responses.” The Quarterly Journal of Economics, 117(3): 1039–1073.

12
Figure 1: Relationship between ARP Child Tax Credit Eligibility and Labor Force Partici-
pation
Panel A. Before/After Extended CTC Introduction
.9

LFPR

.6
0 20 40 60 80 100
Percentile of ARP CTC Eligibility

2021 Feb-Jun 2021 Aug-Dec

Panel B. Before/After Extended CTC Expiration
.9

LFPR

.6
0 20 40 60 80 100
Percentile of ARP CTC Eligibility

2021 Aug-Dec 2022 Jan-Mar

This figure shows the relationship between ARP CTC eligibility and labor force participation. Panel A
shows the relationship before (Feb-June 2021) versus after (Aug-Dec 2021) the introduction of the ARP
CTC benefit. Panel B shows the relationship before (Aug-Dec 2021) versus after (Jan-Mar 2022) the
expiration of the ARP CTC benefit.

13
Figure 2: Relationship between Placebo ARP Child Tax Credit Eligibility and Labor Force
Participation: 2019
Panel A. Before/After Extended CTC Introduction: Placebo (2019)
.9

LFPR

.6
0 20 40 60 80 100
Percentile of ARP CTC Eligibility

2019 Feb-Jun 2019 Aug-Dec

Panel B. Before/After Extended CTC Expiration: Placebo (2019)
.9

LFPR

.6
0 20 40 60 80 100
Percentile of ARP CTC Eligibility

2019 Aug-Dec 2020 Jan-Mar

This figure shows the relationship between placebo ARP CTC eligibility (in 2019) and labor force
participation. Panel A shows the relationship before (Feb-June 2019) versus after (Aug-Dec 2019) the
placebo equivalent of the introduction of the ARP CTC benefit. Panel B shows the relationship before
(Aug-Dec 2021) versus after (Jan-Mar 2022) the placebo equivalent of the introduction of the ARP CTC
benefit.

14
Figure 3: Effect of ARP Child Tax Credit Extension on Labor Force Participation and Total
Hours Worked

.002

.001

−.001

−.002
ra l
l Men om Deg eg ds Ki d
hi
te
hi
te
ve en ol o
D Ki 1 W on
O W C N 2+ W
N

LFP Hours

This figure shows the effect of the ARP CTC eligibility percentile on labor force participation and hours.
The dependent variable is scaled by its standard deviation. Specification detailed in the text. Standard
errors are clustered by region and household size.

15
Table 1: Effect of ARP Child Tax Credit Extension on Labor Force Participation

(1) (2) (3) (4) (5) (6)
DD DDD DD Placebo DD DDD DD Placebo
Rollout Rollout Rollout Expiration Expiration Expiration

Panel A: Cash on Hand Model

100 × CTC-to-Income Pctile 0.017 -0.010 0.026 0.004 -0.011 0.015
(0.008) (0.013) (0.011) (0.009) (0.012) (0.008)

Panel B: Annual Budget Set Model
16
100 × Return-to-Work Incentive -0.049 -0.025 -0.025 -0.059 -0.046 -0.013
(0.045) (0.061) (0.042) (0.047) (0.068) (0.048)

Panel C: Upper Bound Return-to-Work Elasticity

Elasticity 0.002 0.005 0.003 0.002 0.004 0.004

Notes: Panel A shows regression estimates of the effect of the ARP CTC eligibility percentile on labor force participation. Panel B shows regression
estimates of the effect on labor force participation of ARP CTC return-to-work incentives (defined as the negative of the CTC benefit in the Tax
Cuts and Jobs Act - see text for details). For the expiration regressions in Panels A and B, we use the post-expiration period as the base period,
showing the coefficient on the interaction term between the CTC-to-income percentiles and an indicator for the expanded CTC time period. Panel C
shows the labor supply elasticity with respect to return-to-work incentives; we calculate this implied elasticity at the 95th percentile of the estimated
confidence intervals of the Panel B estimates. Specification detailed in the text. Standard errors are clustered by region and household size.

File and source

File: the-short-term-labor-supply-response-to-the-expanded-ctc.pdf
Size: 318,523 bytes
SHA-256: cf70a7cdb54e95fcd941d7c68dbcda3a46a220fff6049267a96807303d52600f
Our copy: the-short-term-labor-supply-response-to-the-expanded-ctc.pdf
Original: www.nber.org