Unconditional cash transfers and mental health symptoms among parents with low incomes: Evidence from the 2021 child tax credit

Abstract

The COVID-19 pandemic increased anxiety and depression in the U.S. population, particularly among low-income households, parents, and Black and Hispanic adults. To address the negative impacts of the pandemic, Congress temporarily expanded the Child Tax Credit (CTC) in 2021, providing a near-universal, unconditional cash transfer to families with children. Using a quasi-experimental, parameterized difference-in-differences research design, we examine the effects of the 2021 monthly CTC on symptoms of anxiety and depression in a large, national sample of parents with low incomes (N∼15,000). We study potential differences in the associations by race/ethnicity and consider whether CTC effects were stronger after a longer treatment period (for instance, due to greater dosage or delayed effects). We find some evidence that the monthly credit reduced parental anxiety and depression symptoms, although the results were not robust throughout all model specifications. Analyses stratified by race/ethnicity show stronger associations for non-Hispanic Black parents than for non-Hispanic White parents or Hispanic parents, although differences were small. We also find the credit reduced anxiety (but not depression) symptoms after three months of payments, suggesting that it took some time for the CTC to affect mental health symptoms. Overall, this study suggests that recurring cash transfers to families in poverty in the U.S. may have small beneficial effects on parental mental health.

Highlights

• •The monthly 2021 expanded Child Tax Credit reduced anxiety and depression symptoms among parents with low incomes in the U.S.
• •Analyses by race/ethnicity show somewhat stronger links with mental health for non-Hispanic Black parents.
• •Reduced anxiety symptoms emerged after three months of payments; the credit took a few months to affect anxiety symptoms.
• •Recurring cash transfers to families with low incomes may have small beneficial effects on parental mental health symptoms.

1. Introduction

Symptoms of anxiety and depression increased in the U.S. population during the COVID-19 pandemic, with larger increases for parents, Black and Hispanic adults, and those with lower incomes (Czeisler et al., 2021; Ettman et al., 2020; McGinty et al., 2020; McKnight-Eily et al., 2021; Patrick et al., 2020). Many studies demonstrate an association between income, poverty, and poorer mental health (Heflin & Iceland, 2009; Sareen et al., 2011). This well-documented association suggests that cash transfer policies might have spillover effects on mental health outcomes. However, despite quasi-experimental evidence from low- and middle-income countries that cash transfers can improve the mental wellbeing of recipients (McGuire et al., 2022), both quasi-experimental and experimental studies on cash transfers in the U.S. context have found more mixed evidence (e.g., Gennetian et al., 2022; Glasner et al., 2022; Pilkauskas et al., 2023).

In March 2021, Congress passed the American Rescue Plan Act to help mitigate the negative effects of the pandemic on households’ economic circumstances, including a temporary expansion to the Child Tax Credit (CTC). The reforms to the CTC were substantial: the benefit size was increased from $2,000 per child to $3,000 per child ages 6–17 and to $3,600 per child under age 6, eligibility was extended to households with little or no earnings, and the credit was made fully refundable. Payments, issued to families by the IRS, were also converted from a single annual lump-sum to more periodic transfers, with half the credit delivered monthly and the other half delivered at tax time. Estimates suggest that the 2021 CTC increased access to the credit for approximately 26 million children whose families were previously either partially or completely ineligible for the benefit, especially Black and Hispanic children (Collyer et al., 2019, Goldin and Michelmore, 2022). Research shows that the 2021 CTC reduced childhood poverty and improved material wellbeing (Parolin et al., 2021; Pilkauskas et al., 2022), but also that the credit had no effect on parental mental health (Collyer et al., 2022; Glasner et al., 2022; although Batra et al., [2023] find lower anxiety and depression symptoms among families with earnings below $35,000 as compared to families with higher earnings and families without children).

Using a large, racially diverse, national sample of families with low incomes, we build on these earlier studies to examine the effects of an unconditional cash transfer policy, the 2021 CTC, on parental depression and anxiety symptoms and differences in effects by race/ethnicity. Unlike earlier work, we focus exclusively on families with very low incomes (average annual earnings of around $10,000), those families who were most strongly impacted by the 2021 CTC reforms (across the income distribution, credit size increased the most for the lowest income families; Collyer et al., 2019, Goldin and Michelmore, 2022) and whose mental health was most negatively impacted by the pandemic (Czeisler et al., 2021; Ettman et al., 2020). We also examine differences by race/ethnicity because Black and Hispanic households were especially likely to gain access to the CTC under the 2021 reforms (Collyer et al., 2019, Goldin and Michelmore, 2022) and because research shows that there are critical racial inequalities in the tax system (Brown, 2022). Our study is also the first to consider whether effects varied as a function of additional months of payments, either because of a time delay in response to the credit or due to a greater dosage of monthly payments.

2. Why might the 2021 monthly CTC payments affect the mental health symptoms of parents with low incomes?

Prior evidence outside the U.S. context generally suggests that unconditional cash transfers improve mental health symptoms (Haushofer & Shapiro, 2016, 2018). There are several plausible ways in which increased income from the 2021 CTC payments could have improved anxiety and depressive symptoms among parents with low incomes. First, the cash payment may have allowed recipients better access to health inputs. For instance, families might have used the additional income from the CTC to purchase food, housing, and medical care, which in turn could have improved mental health. Recent research on the 2021 CTC indicates that the credit reduced material hardship and food insecurity, especially among families in poverty (Collyer et al., 2022; Parolin et al., forthcoming; Pilkauskas et al., 2022; Shafer et al., 2022). By alleviating some types of hardship, the CTC could have reduced parental stress, thus leading to improvements in mental health (Gundersen & Ziliak, 2015; Heflin & Iceland, 2009).

Furthermore, the design features of the 2021 CTC may have improved mental health symptoms among families with low incomes. By creating a recurring monthly payment, the 2021 CTC likely helped reduce income volatility (Maag, 2022), which is more pronounced for families in poverty (Hardy & Ziliak, 2014) and is linked with greater psychological distress (Hill et al., 2013; Sykes et al., 2015). Because the 2021 CTC was extended to parents without earnings, it was nearly universally available to families. Thus, the universal design of the 2021 CTC, which differentiates it from most other cash assistance programs in the U.S., may have removed any stigmatizing barriers to accessing benefits, potentially enhancing its effects on mental health (Mak et al., 2007; Shaefer et al., 2018; Stuber & Schlesinger, 2006). Additionally, unlike most cash assistance programs in the U.S. that provide in-kind or conditional assistance, the CTC provided unconditional cash to families. Distributing cash benefits empowers families to allocate benefits to their needs (Shaefer et al., 2018) and may provide families with greater dignity (Sykes et al., 2015), which in turn may improve mental health symptoms.

Although we expect that, on average, the monthly credit improved the mental health symptoms of parents, some recent studies find null effects of the 2021 CTC on indicators of mental wellbeing (Collyer et al., 2022; Glasner et al., 2022). Furthermore, other experimental studies of unconditional cash transfers have found null or negative mental health impacts (Gennetian et al., 2022; Jacob et al., 2022; Jaroszewicz et al., 2022; Liebman et al., 2022; Magnuson et al., 2022; Pilkauskas et al., 2023). Jaroszewicz et al. (2022) propose that receipt of insufficient cash support (some, but not enough) may increase the salience of unmet needs and therefore produce feelings of distress. In the case of the 2021 CTC, uncertainty concerning the policy's renewal may have also increased beneficiaries' worries about meeting ongoing needs if the new benefits were to cease. Additionally, the 2021 CTC occurred during the COVID-19 pandemic, when parents were facing many other challenges, such as rising inflation rates and closures of child care centers and schools, which could have dampened any positive effects of the cash transfer on symptoms of anxiety and depression.

In contrast to other tax credits, which are delivered in a lump sum following annual tax filing, half of the 2021 CTC was distributed as a monthly benefit for six months (the other half continued to be distributed as a lump sum at tax time). This distribution mechanism could have impacted the temporal pattern of any mental health symptoms. We might expect, for instance, to find larger impacts in later rounds of the monthly payments for several reasons. First, the impacts of any purchased health inputs (such as better housing or medical care) likely take time to affect wellbeing, and thus would not be apparent in mental health symptoms immediately after receipt of the first monthly credit. Second, if lower income volatility were to improve mental health, several monthly CTC payments would be necessary for income to stabilize and for any effects to emerge. Third, it may be the case that several consistent monthly payments are necessary to activate reductions in parental stress that can undermine mental health. For all of these reasons, we might find a slight delay in the effect of the monthly credit on mental health symptoms. Additionally, if CTC payments have an additive or dose-response effects on mental health symptoms, we might find differential associations by payment month, whereby larger associations are observed in later months of the credit.

3. Prior literature

Although there is an extensive correlational literature examining the association between poverty or income and mental health (e.g., Sareen et al., 2011), far less is known about the causal link between unconditional cash transfers and mental health in the U.S. Much of the existing causal evidence on the relationship between income and mental health has been drawn from quasi-experimental studies of the Earned Income Tax Credit ([EITC]; Boyd-Swan et al., 2016; Collin et al., 2021; Dow et al., 2020; Evans & Garthwaite, 2014; Gangopadhyaya et al., 2020; Jones et al., 2022; Morgan et al., 2022; Quian & Wehby, 2021; Shields-Zeeman et al., 2021). These studies have found both null and beneficial effects of the EITC on mental health. One experimental study of the EITC found it improved mental health, especially among women and non-custodial parents (Courtin et al., 2021). However, because EITC benefits are conditioned on employment, it is not clear whether these improvements are driven by employment, income, or both. Another study capturing the broader social safety net found that a $1,000 increase in both food and cash benefits reduced psychological distress in single mothers by 8.4 percent (Schmidt et al., 2023).

Several randomized evaluations of one-time unconditional cash transfers to low-income households have examined effects on mental health and found largely null results (Jacob et al., 2022; Jaroszewicz et al., 2022), although one study found improvements in mental health among very low-income families (Pilkauskas et al., 2023). Studies of recurring unconditional cash transfers to families with low incomes have similarly found little evidence of impacts on parental mental health (Gennetian et al., 2022; Liebman et al., 2022; Magnuson et al., 2022). Outside of the U.S., evidence suggests that income transfers lead to modest improvements in mental health among adults, but income source, payment size, payment timing, and country-setting influence the strength and magnitude of any relationship (see McGuire et al., 2022; Ridley et al., 2020; Thomson et al., 2022; Wilson & McDaid, 2021 for reviews). Furthermore, socioeconomic position of the recipient moderates the influence of cash on mental health, with larger effects among those with the lowest incomes (Thomson et al., 2022).

Although a number of studies of the 2021 CTC have found that the monthly payments improved food insecurity and some measures of material well-being (e.g., Hamilton et al., 2022, p. 173; Parolin et al., 2021; Pilkauskas et al., 2022; Shafer et al., 2022), only three studies have examined whether the cash transfer improved mental health. One study found that among adults in New York City, CTC benefits were associated with less psychological distress, but the association did not reach statistical significance (Collyer et al., 2022). A second study based on a nationally representative sample of adults also found no significant association between the CTC and mental health (Glasner et al., 2022). The final study used a nationally representative sample of adults and a triple interaction model to examine the link between the CTC and anxiety and depression symptoms (Batra et al., 2023). The authors found that the CTC significantly reduced anxiety and depression symptoms for parents with incomes below $35,000 as compared to parents with higher income and childless adults. A related study of the Canadian Child Tax Credit found that regular cash payments improved the mental health of parents and children (Milligan & Stabile, 2011).

The current study differs from earlier CTC studies by focusing on families with very low incomes, those whose CTC benefits were most affected by the 2021 reforms. We also focus on families with children, comparing families of different sizes before and after the expansion, rather than using childless adults as a comparison group (see the Methods section for a discussion of why we prefer this approach). Lastly, we also consider whether the CTC took some time to influence anxiety and depression symptoms (i.e., whether effects grew larger with a longer time receiving monthly payments).

4. Method

Data and Sample.

We draw on unique data collected in partnership with Propel, a nationwide provider of a free mobile application (app) that assists over 5 million U.S. families in managing their SNAP benefits. Approximately 25% of all SNAP beneficiaries in the U.S. use the Propel app to track their monthly SNAP benefits. Propel has achieved a wide reach through referrals from SNAP caseworkers as well as through online advertisement. This partnership provided us with an opportunity to survey many families living in poverty, a population that is typically under-represented in national surveys.

Each month, Propel invites a random sample of their users to participate in a survey on a range of topics related to household well-being and financial circumstances. Following passage of the 2021 CTC reforms, we partnered with Propel to add questions about CTC receipt, economic hardship, and other dimensions of well-being to their survey. We fielded two surveys (June 1-14th and July 1-14th) before the first CTC payment went out on the 15^th of July 2021 and collected data on all six months of the monthly CTC payment ending in January 2022. We restrict the sample to respondents with co-residential children under 18 years old with non-missing responses to mental health questions, leaving an analysis sample of 15,593 individuals.1 In Appendix Table 1, we compare the characteristics of the Propel analytic sample to two nationally representative samples of households with children: the 2019 American Community Survey (ACS) and the 2019 Current Population Survey (CPS) Annual Social and Economic Supplement (ASEC). Because SNAP receipt is often under-reported in national surveys (Meyer et al., 2009), we also compare the characteristics of our sample to the 2019 SNAP Quality Control (SNAP QC) data. We find that the respondents in our Propel analytic sample look quite similar to both households with children below the poverty line and those who reported receiving SNAP in the prior year, with some minor differences (the Propel sample is slightly younger and has more Black respondents than the ACS or CPS but appears more similar to the SNAP QC data).2

Table 1.

Descriptive statistics for the analysis sample.

	Overall	Pre-CTC	Post-CTC
	(n = 15,593)	(n = 3,951)	(n = 11,642)
Age (%)
18–24	7	7	8
25–34	42	40	43
35–44	36	37	36
45–54	11	12	11
55+	4	4	3
Race/Ethnicity (%)
Black non-Hispanic	36	31	37
White non-Hispanic	35	39	34
Hispanic	20	21	20
Other non-Hispanic	8	9	8
Education (%)
Less than High School	21	20	22
High School	40	37	40
Some College	28	30	27
Associates Degree +	11	13	10
Gender (%)
Female	95	95	95
Urbanicity (%)
Urban	46	43	47
Rural	32	32	32
Suburban	22	24	21
Partner/spouse (%)	30	31	29
Household size, mean	4.29	4.26	4.3
Monthly household earnings (%)
$0	22	23	22
$0-500	17	14	17
$500-$999	19	18	19
$1000-$1999	25	26	25
$2000+	13	15	13

Notes: Providers data collected from June 2021–January 2022. Sample is restricted to mental health questionnaire respondents with children under 18. Pre-CTC months include June 2021 and July 2021. Post-CTC months include August 2021–January 2022. Participants were asked about their CTC in the previous month (e.g., in August, participants were asked about CTC receipt in July).

We focus our study on families with children. Some prior CTC research has used childless individuals as a comparison group (e.g., Batra et al., 2023); however, we believe that childless individuals make a poor counterfactual in our study. First, in order to conduct a difference-in-differences analysis comparing families with children to those without children, it is necessary to satisfy the parallel trends assumption. Although we have limited pre-CTC data, trends between families with and without children do not appear parallel in our study sample. Second, individuals with and without children had starkly different experiences during our study period due to the ongoing COVID-19 pandemic and its associated economic disruptions, policy responses, and school re-openings. All of these factors could lead to differential trends in mental health symptoms among families with and without children around the same time that the monthly CTC was implemented. Thus, using childless individuals as a counterfactual could lead to biased estimates if other factors that differentially affect families with and without children were changing at the same time as the CTC. For these reasons, we rely on variation in the amount of CTC benefits that parents received according to the number and ages of children residing in their household. However, if there were differential impacts of policies or contextual factors by the age of the child, our estimates may still be biased (although the direction of this potential bias is unclear). In a supplemental analysis, we estimate a traditional difference-in-differences model using families without children as a comparison for families with children (see Appendix Table 4).

Table 4.

Estimated "dosage" or time effects of monthly 2021 Child Tax Credit (CTC) on anxiety and depression symptoms.

	Anxiety								Depression
	Intent-to-treat				Treatment-on-the-treated				Intent-to-treat				Treatment-on-the-treated
	(1)		(2)		(1)		(2)		(1)		(2)		(1)		(2)
Payment Month 1	−0.015		−0.042		−0.006		−0.014		0.009		−0.014		0.035		−0.014
	(0.028)		(0.030)		(0.048)		(0.055)		(0.028)		(0.030)		(0.048)		(0.055)
Payment Month 2	0.006		−0.022		0.037		−0.014		0.026		0.004		0.058		0.014
	(0.029)		(0.031)		(0.039)		(0.047)		(0.029)		(0.031)		(0.039)		(0.047)
Payment Month 3	0.008		−0.020		0.044		−0.005		0.016		−0.006		0.041		−0.005
	(0.032)		(0.034)		(0.049)		(0.056)		(0.032)		(0.034)		(0.049)		(0.056)
Payment Month 4	0.051	+	0.023		0.111	*	0.075		0.068	*	0.045		0.124	**	0.075
	(0.030)		(0.031)		(0.044)		(0.052)		(0.030)		(0.031)		(0.044)		(0.052)
Payment Month 5	0.023		−0.004		0.071		−0.025		0.002		−0.021		0.020		−0.025
	(0.030)		(0.031)		(0.043)		(0.049)		(0.030)		(0.031)		(0.043)		(0.049)
Payment Month 6	0.051		0.025		0.105	*	0.025		0.040		0.018		0.065		0.025
	(0.033)		(0.035)		(0.045)		(0.051)		(0.033)		(0.035)		(0.045)		(0.051)
CTC Benefits	0.001		0.005		0.000		0.019		−0.002		0.001		−0.013		0.002
	(0.003)		(0.004)		(0.014)		(0.016)		(0.003)		(0.004)		(0.014)		(0.016)
CTC Benefits x Payment 2	−0.004		−0.004		−0.014		−0.016		−0.002		−0.002		−0.005		−0.007
	(0.004)		(0.004)		(0.015)		(0.015)		(0.004)		(0.004)		(0.015)		(0.015)
CTC Benefits x Payment 3	−0.004		−0.004		−0.017		−0.017		0.000		0.000		0.001		0.000
	(0.004)		(0.004)		(0.018)		(0.018)		(0.004)		(0.004)		(0.018)		(0.018)
CTC Benefits x Payment 4	−0.008	*	−0.008	*	−0.033	*	−0.034	*	−0.005		−0.005		−0.019		−0.020
	(0.004)		(0.004)		(0.016)		(0.016)		(0.004)		(0.004)		(0.016)		(0.016)
CTC Benefits x Payment 5	−0.007	+	−0.007	+	−0.025		−0.027	+	0.001		0.001		0.006		0.004
	(0.004)		(0.004)		(0.016)		(0.016)		(0.004)		(0.004)		(0.016)		(0.016)
CTC Benefits x Payment 6	−0.010	*	−0.010	*	−0.032	+	−0.036	*	−0.001		−0.001		−0.001		−0.005
	(0.004)		(0.004)		(0.016)		(0.017)		(0.004)		(0.004)		(0.017)		(0.017)
Household size controls	x				x				x				x
Number of children controls			x				x				x				x
N	15,593

Notes: Coefficients represent the effect of an additional $100 in monthly 2021 CTC benefits on the likelihood of mental health symptoms. Models include the full set of controls.

+ p < .10; *p < .05; **p < .01.

In Table 1, we present descriptive statistics for our main analysis sample and include information presented separately for parents sampled in the pre- and post-CTC monthly survey periods. Thirty-six percent of respondents identified as non-Hispanic Black, 35% as non-Hispanic White, and 20% as Hispanic. The majority of respondents (95%) identified as female, thus, in many ways this study examines the link between the CTC and mental health among mothers rather than parents overall. The sample is also composed of parents living in very low-income households: 22% reported having no household earnings in the previous month, and another 17% reported earning less than $500. Extrapolating from monthly to annual earnings, average household earnings are around $10,000 in the sample. Under the pre-2021 CTC structure, many of these families would not have had sufficient earnings to qualify for the full CTC benefit, or in many cases, any credit at all. Table 1 also shows that characteristics of parents surveyed in the pre-vs. post-monthly CTC periods are similar.

4.1. Mental health measures

To assess parent's mental health, we use questions from the Patient Health Questionnaire-4 (PHQ-4), a widely used and validated instrument for measuring adverse mental health symptoms (Kroenke et al., 2009). The PHQ-4 consists of two subscales that evaluate symptoms of depression (PHQ-2) and anxiety (GAD-2), two of the most common mental health issues among adults (Kroenke et al., 2009). Respondents are asked how often in the past two weeks they (1) felt little interest or pleasure in doing things; (2) felt down, depressed, or hopeless; (3) felt nervous, anxious, or on edge; or (4) felt unable to stop or control worrying. Responses to each question are scored on a 0 through 3 scale (0 = “not at all”, 1 = “several days”, 2 = “more than half the days”, 3 = “nearly every day”). Using these questions, we separately examine depression and anxiety symptoms. Following standard practice on cut points to determine each outcome, we use a score of 3 or higher on the GAD-2 scale (the first two questions above) to categorize depression symptoms and a score of 3 or higher on PHQ -2 scale (the last two questions above) to categorize anxiety symptoms. In Appendix Table 2, we present the results using alternative cut-offs and continuous scores.

Table 2.

Means on the 2021 monthly CTC and mental health symptoms, overall and by race/ethnicity.

		Race/Ethnicity
	Full sample	Non-Hispanic Black	Non-Hispanic White	Hispanic
Child Tax Credit
Self-reported CTC receipt (%)	67	69	67	63
Self-reported monthly CTC payment ($), mean (SD)	332 (257)	339 (244)	347 (266)	311 (257)
Self-reported monthly CTC payment among receivers ($), mean (SD)	496 (314)	489 (304)	495 (319)	498 (315)
Predicted monthly CTC payment ($), mean (SD)	705 (411)	742 (435)	640 (366)	732 (417)
Mental Health Symptoms
Anxiety (%)	48	43	55	41
Depression (%)	47	46	51	42
N	15,593	5,509	5,443	3,159

Notes: Sample is restricted to mental health questionnaire respondents with children under 18. CTC variables measured in August 2021–January 2022. Participants were asked about their CTC in the previous month (e.g., in August, participants were asked about CTC receipt in July). Mental health outcomes were evaluated using the Patient Health Questionnaire-4 (PHQ-4).

4.2. Analytical strategy

To examine whether the 2021 monthly CTC payments affected the mental health symptoms of parents with low income, we use a parameterized difference-in-differences approach, incorporating policy-induced variation in the size of CTC benefits into our models. Because self-reported CTC benefits are endogenous to family characteristics (Pilkauskas & Michelmore, 2021), we simulate monthly CTC benefits based on the number and ages of children in the household and the month of the year. Specifically, we multiply per-child payments ($300 per child under 6 years and $250 per child 6–17 years) by the number of children in each age group (under 6 years and 6–17 years) and sum the amounts. In months before the CTC payments went into effect, respondents are assigned $0 in CTC benefits. We then fit linear probability regression models with our mental health measures as the outcomes and predicted CTC benefits as the primary association of interest. This approach is displayed in the following equation:

$$Y_{it}={\beta }_0+{\beta }_1{CTC}_{itc}+{\beta }_2{X}_{it}+{\gamma }_{st}+{\delta }_s+{\alpha }_t+{{\theta }_c+\epsilon }_{it}$$ (1)

Where Y_it represents the mental health outcome of interest: a binary indicator of depression or anxiety symptoms. The subscripts i, t, s, and c indicate individual, month, state, and number of children, respectively. ${CTC}_{itc}$ is our main explanatory variable, the measure of predicted CTC benefits described above. This CTC variable is scaled such that the corresponding coefficient (${\beta }_1)$ represents an additional $100 in predicted monthly CTC benefits. In addition, we control for individual-level covariates ($X_{it}$): the respondent's age, race/ethnicity, gender, education (less than high school, high school, some college, or associate's degree or higher), place of residence (urban, rural, or suburban), and an indicator for their partnership status (living with a married/cohabiting partner). Our model also controls for a set of state- and month-specific policy variables (${\gamma }_{st}$) that might be associated with both CTC receipt and parental mental well-being: the presence of SNAP emergency allotments (i.e. waivers to provide eligible households with maximum SNAP benefits), the presence of Pandemic Electronic Benefit Transfers (i.e. additional food assistance for school aged-children), and the presence of extended federal Unemployment Insurance. We also include controls (also known as fixed effects) for state of residence (${\delta }_s)$ and for each survey month (${\alpha }_t)$ to control for other unmeasured state-specific economic conditions or national level policies that might confound our results.

Lastly, we include controls for household size to address concerns that respondents in larger households may experience anxiety and depression symptoms at different rates than respondents in smaller households, which would bias our estimates of the impact of the CTC on mental health. It might be the case, for instance, that larger households experience lower rates of anxiety and depression symptoms, and also receive larger CTC benefits, thus creating a spurious, negative correlation between CTC benefits and anxiety and depression symptoms. To address these concerns, in two separate model specifications, we adjust for either 1) household size or 2) number of children (${\theta }_c$ represents either set of variables). Models that include household size controls identify the impact of the CTC by comparing households with the same number of individuals, but different numbers of children under the age of 18, as well as different age compositions of children. Models that include number-of-children controls are more stringent and are identified by comparing households with the same number of children but different age compositions of those children. While both sets of controls address the concern that the various household size groups experience different rates of mental health symptoms (Helbig et al., 2006; Myrskylä & Margolis, 2014), the number-of-child controls are directly correlated with CTC benefits and thus, absorb the majority of the variation in our predicted CTC measure (81%; by contrast, household size controls explain 60% of this variation). Therefore, inclusion of number-of-children controls are a conservative adjustment. We present estimates from both models, noting any differences between the two. Last, we estimate separate subgroup regressions to examine potential heterogeneity in any effects of CTC payments on mental health by race/ethnicity.

In addition to effects that reflect the introduction of monthly CTC payments, we are also interested in effects of CTC “dosage” or time (specifically, the effects of additional months of CTC payments, or whether there is a “critical” dosage needed before change is observed) on parent's mental health. To examine this possibility, we extend Equation (1) by interacting predicted CTC benefits with a set of binary variables indicating month (each month from the second CTC payment to the sixth CTC payment, with the first month as the reference category). This model captures how the effects of the CTC on parent's anxiety and depression symptoms unfolded as monthly payments progressed.

The linear probability models that we have described thus far provide intent-to-treat (ITT) estimates, which represent the effect of the CTC on all respondents regardless of CTC receipt. However, because not all families with children received the monthly CTC payment, we are also interested in the association between actual CTC benefits received and parent's mental health (i.e., the treatment-on-the-treated [TOT] effect, also known as the Local Average Treatment Effect [LATE]). Thus, we estimate a TOT response using predicted monthly CTC benefits as an instrumental variable (IV) for self-reported monthly CTC benefits in our parameterized difference-in-differences model. In particular, we estimate a two-stage least squares regression model in which the first stage regresses self-reported CTC benefits (the endogenous variable) on our predicted measure of CTC benefits (the exogenous variable) This method essentially rescales our ITT estimates to those who self-reported receipt of monthly CTC benefits. We note, however, that if the CTC benefit receipt is under-reported in our survey (as it is in other national surveys), then the TOT estimates may be inflated.

4.3. Study results

Table 2 provides descriptive statistics on the CTC and mental health measures used in our analyses. In our analytic sample, 67% of respondents reported receiving the monthly CTC payment in any given month. The average monthly CTC payment amount was $332, or $496 among those who reported receiving a payment (note, the higher predicted CTC amount is largely driven by many parents reporting they did not get the CTC). Overall, rates of CTC receipt were lower among Hispanic respondents (63%) than among non-Hispanic Black respondents (69%) and non-Hispanic White (67%) respondents. Among those who received the credit, the average payment amount was similar across racial/ethnic groups.

Many parents exhibit anxiety and depressive symptoms. In the full sample, 48% of respondents were considered high risk for anxiety and 47% were high risk for depression. Co-occurrence of depression and anxiety symptoms was common, with 39% of parents exhibiting both symptoms. Non-Hispanic White respondents reported higher rates of anxiety symptoms (55%) than non-Hispanic Black (43%) or Hispanic respondents (41%). This pattern was similar for rates of depression. These rates of anxiety and depression symptoms generally align with recent estimates that around 39% of U.S. adults experienced depression and/or anxiety in 2020–2021, and that after adjusting for socioeconomic resources, 43% of White adults, 38% of Black adults, and 39% of Hispanic adults experienced depression and/or anxiety in 2020–2021 (Thomeer et al., 2022).

Our primary results are presented in Table 3. We show coefficient estimates and standard errors from both the intent-to-treat (ITT) and the treatment-on-the-treated (TOT) analyses for models including either 1) household size or 2) number-of-children controls. In all models, the coefficient for CTC exposure indicated a negative association between CTC benefits and the likelihood that parents exhibit anxiety or depression symptoms. The overall pattern in Table 3 – that TOT estimates from our IV analyses were larger in magnitude than the ITT estimates – is consistent with expectations since TOT estimates represent the effect of the CTC among those who received the credit. In models including household size controls, the ITT (TOT) estimates suggest that an additional $100 in monthly CTC benefits was associated with a 0.5 (2.1) percentage point (p < .01) decrease in anxiety symptoms and a 0.4 (1.7) percentage point (p < .01) decrease in depression symptoms. However, once we adjusted for number of children, the CTC-associated effects were substantially smaller and no longer statistically significant, but the coefficients remained negatively signed. One concern may be that the results were underestimated (biased downwards) if parents with more children, who received larger CTC benefits, also had worse mental health outcomes (e.g., Myrskylä & Margolis, 2014). In a supplemental analysis, we found no evidence to suggest that this was the case; number of children was not significantly associated with anxiety or depression symptoms and rates of these symptoms were similar across number of children.

Table 3.

Estimated effects of monthly 2021 CTC benefits on anxiety and depression symptoms.

	Intent-to-treat				Treatment-on-the-treated
	(1)		(2)		(1)		(2)
Outcome
Anxiety	−0.005	**	−0.001		−0.021	**	−0.006
	(0.001)		(0.002)		(0.006)		(0.010)
Depression	−0.004	**	−0.001		−0.017	**	−0.003
	(0.001)		(0.002)		(0.006)		(0.010)
Household size controls	x				x
Number of children controls			x				x
N	15,593

Notes: "Intent-to-treat" estimates show the effect of monthly 2021 CTC benefits regardless of CTC receipt. "Treatment-on-the-treated" estimates show the effect of actual receipt of monthly 2021 CTC benefits among those who received the CTC. Coefficients represent the effect of an additional $100 in monthly CTC benefits on the likelihood of mental health symptoms. Mental health symptoms were evaluated using the GAD-2 and PHQ-2 scales. Models include the full set of controls.

**p < .01.

Fig. 1 plots the coefficients and 95% confidence intervals for both our ITT and TOT (LATE) estimates by subgroups based on race and ethnicity. The estimated association of the monthly CTC was more pronounced for non-Hispanic Black parents than for non-Hispanic White or Hispanic parents. The associations between the monthly CTC and all mental health symptoms were only statistically significant for non-Hispanic Black parents, but the estimates were negatively signed for both non-Hispanic White and Hispanic parents. Furthermore, the confidence intervals overlapped, indicating that we cannot rule out that CTC effects on mental health symptoms were the same regardless of racial/ethnic identity. Among non-Hispanic Black parents, the ITT (TOT) estimates suggest that an additional $100 in monthly CTC payments was associated with a 0.8 (4.1) percentage point (p < .01) decrease in anxiety symptoms and a 0.6 (3.4) percentage point (p < .01) decrease in depression symptoms. The estimates remain similar with the inclusion of number-of-children controls, but were less precise.

Fig. 1.

Estimated effects of monthly 2021 CTC benefits on anxiety and depression symptoms, overall and by race/ethnicity

Notes: Full sample: N = 15,593; Black non-Hispanic: N = 5,509; White non-Hispanic, N = 5,443; Hispanic, N = 3,159. Coefficients represent the effect of an additional $100 in monthly CTC benefits on the likelihood of mental health symptoms. Circles and black lines represent point estimates and 95% confidence intervals for models including controls for household size. Triangles and grey lines represent point estimates and 95% confidence intervals for models including controls for number of children. ITT="intent-to-treat", LATE="local average treatment effect" or treatment-on-the-treated.

The analyses in Table 4 provide insights into whether effects of the CTC on mental health symptoms grew as additional payments were issued to families. We found some evidence of such an association for anxiety symptoms. Coefficients for anxiety, which were negatively signed for all months, increased in magnitude and statistical significance with additional monthly payments in both the ITT and TOT estimates. Associations were strongest in the final three months of payments, when an additional $100 in monthly CTC benefits was associated with a 0.7–1.0 (2.7–3.76) percentage point decline in anxiety symptoms. Furthermore, these associations were robust across our two model specifications.

4.4. Sensitivity analyses

In Table 2, Table 3 of the Appendix, we conducted several sensitivity analyses. As noted earlier, we tested alternative approaches to measuring depression and anxiety symptoms: as continuous scores and with a stricter cut off of the GAD-2 and PHQ-2 scores to identify each symptom. The findings were similar to those presented here. We also performed sensitivity analyses where we changed the observation window of our study. First, we excluded the first month of the CTC payment, given evidence that families were somewhat less likely to have received the credit in the first month after implementation of the reforms (Pilkauskas & Cooney, 2021; Michelmore & Pilkauskas, forthcoming), and results were robust to that exclusion. Second, we excluded the last month of the observation window in case families responded differently as monthly payments were ending. Again the results were robust to the alternative observation window. Third, because COVID-19 cases surged in December 2021, we also tested the robustness of the results to the inclusion of state-specific monthly COVID-19 rates, and results were quite similar. Fourth, we included a control for the number of days between CTC receipt and survey completion in case associations depended on how much time had lapsed between the two events, and results were again unchanged. Last, because the majority of our study respondents were female, we tested the robustness of our results to dropping the men from the study and found they were largely unchanged.

4.4.1. Comparison to families without children

While our main analyses focused on households with children under the age of 18, in a supplemental analysis (shown in Appendix Table 4), we also conducted a traditional difference-in-differences analysis, where we compared changes in mental health symptoms following implementation of the 2021 CTC between treated (households with children under 18) and untreated (households without children under 18) groups. As we discuss above, we do not believe that childless individuals serve as an adequate comparison group for those with children, but present results for the sake of completeness. We conducted these analyses measuring the CTC in two ways: 1) with a pre-post indicator interacted with an indicator for the presence of children in the household and 2) with predicted CTC amounts.3 In general, many of the findings discussed thus far are insignificant in these models, though we continue to find evidence of reductions in anxiety in models that use families without children as a comparison group and take into account the generosity of the monthly CTC benefits.

4.5. Discussion

Despite growing policy interest in unconditional income supports (e.g., Shaefer et al., 2018) and significant correlational evidence of the positive relationship between income and mental health (e.g., Sareen et al., 2011), experimental and quasi-experimental evidence on the effects of income on mental health is more mixed. Studies taking a more causal approach to examining the relationship between income and mental health have reported both positive (e.g., Batra et al., 2023; Evans & Garthwaite, 2014) and null effects of unconditional cash on mental health (e.g., Collyer et al., 2022; Gennetian et al., 2022; Glasner et al., 2022). This study adds to this literature by using a quasi-experimental research method to examine the link between 2021 monthly CTC payments and mental health symptoms among parents with low incomes. We focus on parents with low incomes as they were particularly affected by the 2021 CTC expansion, with the credit made newly available to families at the very bottom of the income distribution (Collyer et al., 2019; Goldin & Michelmore, 2022). Our large sample of such parents also allowed us to test for temporal variation in effects, to consider whether it took several months for monthly payments to affect mental health symptoms. By studying the effects of unconditional cash transfers on parental mental health symptoms in particular, we focus on an outcome with major implications for both parents and their children (e.g., Lovejoy et al., 2000; Turney, 2011).

Our models indicated that among those who reported receiving the credit, an additional $100 in monthly CTC benefits was associated with a 0.1–0.5 percentage point decrease in anxiety symptoms and a 0.1–0.4 percentage point decrease in depression symptoms. Scaling these estimates by those who reported receiving the monthly CTC implies a TOT effect ranging from a 0.6–2.1 percentage point decline in anxiety symptoms, and a 0.3–1.7 percentage point decline in depression symptoms associated with a $100 increase in monthly CTC benefits. Although estimates were not significant in the model controlling for number of children, as noted earlier, this model absorbed the majority of variation in our predicted CTC measure. That all estimates were negatively signed supports that CTC payments improved parental mental health symptoms.

Additionally, in the analyses that considered CTC dosage or time effects, we observed reductions in anxiety symptoms that were robust across model specifications. The point estimates increased somewhat in size with additional months of the payment, but it appears that by the third monthly payment, effects on anxiety symptoms seem to have stabilized. This suggests that income stabilization, health inputs or general stress levels, took a few months to really take effect, rather than a simple dose-response (where the effects grow over time). We found no evidence of dosage or time effects for depression symptoms (despite results consistent with an overall decline in depression). It is unclear why the beneficial effects of CTC payments increased with additional payments for anxiety symptoms, but not for depression. However, it might be that anxiety symptoms are more sensitive to income than depression. Had the payment lasted beyond six months, depression symptoms may have also improved. The larger effect on anxiety symptoms is consistent with other work on the CTC (Batra et al., 2023). In the future, studies of unconditional cash transfers should assess the role of time and frequency of payments in mental health effects.

Because the average CTC payment received by parents in our sample was $500, estimates suggested around a 6–20% reduction in anxiety symptoms and a 3–18% reduction in depression symptoms for parents living in poverty. Other research on social safety net policies estimated that an additional $1,000 reduced mental distress by about 8 percent (Schmidt et al., 2021). The effect sizes we find here are somewhat larger, but the structure of this policy (a monthly, unconditional credit) is also somewhat unique in terms of U.S. safety net assistance. Large estimated effects in this study could also be explained by the fact that the monthly cash transfers were quite large relative to household income of the families in our study. On average, monthly household earnings were around $830, and almost 40% of parents reported monthly household earnings below $500. Thus, the credit increased monthly income by nearly 60%, on average. Prior evidence on cash transfers has shown that effects are often larger for households that receive larger transfers relative to income (e.g., Evans et al., 2016; Pilkauskas et al., 2023).

In subgroup analyses, we found some evidence that CTC payments had a stronger association with the mental health symptoms of non-Hispanic Black parents; however, the point estimates suggested reductions in anxiety and depression symptoms across all race/ethnic groups. Research suggests that Black families were more likely to have been entirely excluded from standard pre-2021 CTC design, whereas Hispanic families were less likely to receive the full CTC compared to White families (Collyer et al., 2019, Goldin and Michelmore, 2022). That the association was largest for Black families may be driven in part by Black families disproportionately becoming newly eligible for the CTC. These findings are also consistent with other research that suggests larger impacts of the CTC on material well-being for Black families (Parolin et al., 2021; Pilkauskas et al., 2022).

This study has some limitations. First, our study has limited generalizability. Although in general our study population looks largely like SNAP recipients (see Appendix Table 1), most of the respondents in our study are mothers. Thus, our results largely reflect the effects of the 2021 CTC on mothers with very low incomes. Second, although the PHQ-4 is a widely used measure for assessing anxiety and depression symptoms, it relies on self-reported information that could be subject to social desirability bias. It is also important to note that the PHQ-4 is a screening tool and not designed for formal clinical diagnosis. Last, this study used repeated cross-sectional data, thus we could not observe changes in mental health symptoms at the individual level (which would require panel data).

Nonetheless, our results point to the beneficial effects of the 2021 monthly CTC payments on the mental health symptoms of parents, in particular mothers, with low incomes in the short-term. Because monthly CTC payments were only delivered for a period of 6 months, our estimates primarily reflect the effect of a short-lived policy on mental health symptoms. What is less well understood is the relationship between unconditional cash transfers and mental well-being in the long-run or what sorts of effects a near-universal longer-term policy could have on families. Since many risk factors for poor mental health, such as exposure to chronic stressors in one's environment, develop over long time scales, we would not necessarily expect them to be impacted by a short-lived policy expansion. Our results for time effects – that the effects on depression took a few months to set in – suggest that a more permanent expansion of the CTC may have the potential for even larger effects on mental health symptoms. Overall, this research suggests that the 2021 CTC slightly improved anxiety and depression symptoms among parents with low incomes, those who were hit especially hard during the COVID-19 pandemic.

Financial disclosure statement

The authors were funded by the Charles and Lynn Schusterman Family Philanthropies ($80,000) and the Susan B. Meister Child Health Evaluation and Research Center at the University of Michigan ($20,000).

Ethical statement

The paper uses deidentified data from Propel's Household Pulse Survey. This project was reviewed by the University of Michigan's Institutional Review Board and determined not regulated.

Declaration of competing interest

The authors declare no conflict of interest.

Appendix

Appendix Table 1.

Comparing descriptive characteristics of the Propel to the ACS, SNAP QC and CPS samples

	Propel	ACS 2019		SNAP Quality Control Data 2019	CPS ASEC 2019
		Households below 100% of poverty	SNAP recipients	SNAP recipients	SNAP recipients
Age (%)
18–24	7	7	6	9	8
25–34	42	37	38	41	38
35–44	36	37	38	32	35
45–54	11	15	15	12	15
55+	4	3	3	5	5
Household structure
Household size	4.29	4.04	4.19	3.56	4.11
Partner/spouse	30	46	47	–	50
Race/Ethnicity (%)
Black non-Hispanic	36	24	27	31	27
White non-Hispanic	35	37	37	41	37
Hispanic	20	31	29	22	27
Other non-Hispanic	8	8	7	7	9
Education (%)
Less than High School	21	21	18	25	23
High School	40	44	45	55	37
Some College	28	25	27	15	31
Associates Degree +	11	10	9	5	10
Female	95	74	72	89	73
N	15,593	34,648	38,322	15,735	2,816

Notes: All samples restricted to houseolds with at least one coresident child under the age of 18. ACS and CPS samples are restricted to the reference person. The SNAP QC sample is restricted to household heads deemed eligible for the SNAP case under review. ACS = American Community Survey 2019; CPS = Current Population Survey ASEC 2019; SNAP QC = SNAP Quality Control 2019.

Appendix Table 2.

Estimated effects of monthly 2021 CTC benefits on the likelihood of anxiety and depression symptoms; varying the outcome measure

	Intent-to-treat				Treatment-on-the-treated
	(1)		(2)		(1)		(2)
Outcome
Alternate cut-off to measure anxiety (GAD-2 ≥4)	−0.005	**	−0.001		−0.021	**	−0.006
	(0.001)		(0.002)		(0.006)		(0.010)
Alternate cut-off to measure depression (PHQ-2 ≥4)	−0.004	**	−0.001		−0.016	**	−0.006
	(0.001)		(0.002)		(0.006)		(0.009)
GAD-2 score	−0.021	**	0.000		−0.092	**	0.001
	(0.006)		(0.008)		(0.027)		(0.041)
PHQ-2 score	−0.017	**	0.003		−0.074	**	0.015
	(0.006)		(0.008)		(0.026)		(0.040)
Household size controls	x				x
Number of children controls			x				x
N	15,593

Notes: "Intent-to-treat" estimates show the effect of monthly 2021 CTC benefits regardless of CTC receipt. "Treatment-on-the-treated" estimates show the effect of actual receipt of monthly 2021 CTC benefits among those who received the CTC. Coefficients represent the effect of an additional $100 in monthly CTC benefits on the likelihood of mental health symptoms. Mental health symptoms were evaluated using the GAD-2 and PHQ-2 scales. Models include the full set of controls.

**p < .01.

Appendix Table 3.

Estimated effects of monthly 2021 CTC benefits on the likelihood of anxiety and depression symptoms; sensitivity analyses

	Anxiety								Depression
	Intent-to-treat				Treatment-on-the-treated				Intent-to-treat				Treatment-on-the-treated
	(1)		(2)		(1)		(2)		(1)		(2)		(1)		(2)
Excluding August	−0.005	**	−0.001		−0.023	**	−0.007		−0.005	**	−0.001		−0.023	**	−0.007
	(0.002)		(0.002)		(0.006)		(0.010)		(0.002)		(0.002)		(0.006)		(0.010)
N	14,182
Excluding January	−0.005	**	−0.001		−0.021	**	−0.0040		−0.004	**	−0.001		−0.021	**	−0.0040
	(0.002)		(0.002)		(0.007)		(0.010)		(0.002)		(0.002)		(0.007)		(0.010)
N	14,173
Controlling for Covid rates	−0.005	**	−0.001		−0.021	**	−0.006		−0.004	**	−0.001		−0.017	**	−0.003
	(0.001)		(0.002)		(0.006)		(0.010)		(0.001)		(0.002)		(0.006)		(0.010)
N	15,593
Controlling for number of days since CTC was disbursed	−0.005	**	−0.001		−0.021	**	−0.005		−0.004	**	−0.001		−0.016	**	−0.003
	(0.001)		(0.002)		(0.006)		(0.010)		(0.001)		(0.002)		(0.007)		(0.010)
N	15,593
Dropping male respondents	−0.005	**	−0.001		−0.021	**	−0.003		−0.003	*	−0.001		−0.013	*	0.000
	(0.001)		(0.002)		(0.006)		(0.010)		(0.001)		(0.002)		(0.006)		(0.010)
N	14,695
Household size controls	x				x				x				x
Number of children controls			x				x				x				x

Notes:"Intent-to-treat" estimates show the effect of monthly 2021 CTC benefits regardless of CTC receipt. "Treatment-on-the-treated" estimates show the effect of actual receipt of monthly CTC benefits among those who received the CTC. Coefficients represent the effect of an additional $100 in monthly CTC benefits on the likelihood of mental health symptoms. Mental health symptoms were evaluated using the Patient Health Questionnaire-4 (PHQ-4). Models include the full set of controls.

*p < .05; **p < .01.

Appendix Table 4.

Pre-post difference-in-differences and continuous/parameterized difference-in-differences - Households without children as a comparison group. Estimated effects on anxiety and depression symptoms.

	Pre-post (0/1 indicator)								Continuous/parameterized
	Intent-to-treat				Treatment-on-the-treated				Intent-to-treat				Treatment-on-the-treated
	(1)		(2)		(1)		(2)		(1)		(2)		(1)		(2)
Outcome
Anxiety	0.011		0.013		0.017		0.020		−0.002	*	0.000		−0.007	*	0.000
	(0.015)		(0.015)		(0.022)		(0.033)		(0.001)		(0.001)		(0.003)		(0.004)
Depression	0.036	*	0.038	*	0.053	*	0.055		0.000		0.002		0.000		0.006
	(0.015)		(0.015)		(0.022)		(0.023)		(0.001)		(0.001)		(0.003)		(0.004)
Household size controls	x				x				x				x
Number of children controls			x				x				x				x
N	24,009

Notes: Sample includes households with and without children. Coefficients for the pre-post difference-in-differences model represent the estimated effect of the introduction of the monthly Child Tax Credit payments. Coefficients for the continuous/parameterized differences-in-differences model represent the effect of an additional $100 in monthly CTC benefits. Models include the full set of controls.

*p < .05.

Data availability

The authors do not have permission to share data.