The economic value of policies and programs to support children’s mental health

Summary

This paper explores evidence about the economic payoff of early childhood and school-based health care policies as they relate to the mental health of children. Following recent policy proposals and changes that have created new threats to health insurance coverage and school-based programs for children, we examine the value of public investment in children’s mental health. Using the Marginal Value of Public Funds (MVPF) approach, we treat public spending on coverage policies and school-based mental health–related programs as investments with potential downstream offsets. We contextualize this analysis with background demographics and the mental health landscape of children, and lessons we’ve learned from historical policy changes related to expanding health insurance coverage and programs intended to improve mental health. The discussion in this paper is focused on children in Pennsylvania, and thus, we apply empirical results from the literature to data that is specific to the state. We show that a variety of early childhood programs and school-based interventions pay for themselves over the long term. While our focus is on Pennsylvania, many of our observations and implications for future policy apply across the United States.

Introduction

Recent policy changes and policy proposals are creating new threats to health insurance coverage for low-income populations and to public subsidies for programs that support early childhood and school-based health care. The state of Pennsylvania has been consistently progressive in its coverage of low-income households by expanding Medicaid and establishing a state-run health insurance marketplace. The state has also actively promoted school-based health centers and supported early childhood programs. Each of these efforts has had significant impacts on the availability and design of mental health care for children. Yet, much of this has relied on federal funding. For this reason, states will need to examine public investment strategies as the rules governing federal funding change.

In this paper, we explore evidence about the economic payoff of early childhood and school-based health care policies and activities as they relate to the mental health of children. First, we concentrate on what has been learned from the set of Medicaid expansions that targeted children in the 1980s and 1990s. Then, we examine the benefits to children of the Medicaid expansions primarily aimed at adults, such as those created by the Affordable Care Act (ACA) over the last decade and a half. Our final set of lessons is based on what has been learned about programs that focus on mental health in schools. The lens we take is that of assessing the payoffs to public spending of coverage policies and school-based mental health–related programs. We pay particular attention to the long-run consequences of these programs, treating them in a manner akin to investments. To that end, we make use of the Marginal Value of Public Funds (MVPF) approach.1 The MVPF is a method of examining the full effects of public funding, including downstream offsets.

The paper proceeds in the following four sections. In the first section, we provide some background data on the demographics of children in Pennsylvania. The second section describes the mental health context of children living in Pennsylvania, including prevalence estimates and indicators of access to care. The third section describes what we know about the Marginal Value of Public Funds related to spending on children’s health and mental health care. We integrate broader studies into information relevant to the Pennsylvania context. The fourth and final section offers concluding observations on policy implications.

Basic demographics

In 2023, there were over 2.6 million children under the age of 18 living in Pennsylvania.2 Of those, 63% were white, 13% were Black, 15% were Hispanic, and 4% were Asian. In 2023, there were just under 1.7 million K-12 students in the state, of which about 551,000 were high school students.3 According to the Current Population Survey’s supplemental poverty measure, the child poverty rate was 9% in 2023.4 In 2023, 5% of children were without any form of health insurance (6% younger than age 6 and 5% between 6 and 18 years old, 5% for white children). Uninsured rates by poverty status were 7% for children under 100% of the federal poverty line (FPL), 7% for children between 100% and 300% of the FPL, and 3% for those above 300% of the FPL.5 Medicaid and other public insurance accounted for 39% of the health insurance coverage for children, while 49% had employer-sponsored health insurance, and 7% had other miscellaneous forms of coverage.

Prevalence

Using 2023 data from the National Survey of Children’s Health (NSCH), we examine the prevalence of various mental, emotional, and behavioral (MEB) disorders among children ages 3 to 17 in the United States broadly, and in Pennsylvania specifically. The NSCH is comprised of responses from parents of children ages 0 to 17 living in noninstitutional settings across the United States. Mental, emotional, and behavioral disorders include attention-deficit hyperactivity disorder (ADHD), behavioral or conduct problems, anxiety, and depression.6

Overall, we estimate that about 14.5% of children in Pennsylvania and 16.8% of children nationally had any MEB disorder in 2023. As seen in Figure 1, the prevalence of most MEB disorders is quite similar between Pennsylvania and the nation as a whole. The most significant difference is that the prevalence of ADHD is elevated nationally compared to Pennsylvania. Anxiety is the most prevalent disorder among children both in Pennsylvania and nationally, hovering around 11%.

Risk factors and markers of illness

There are a number of indicators of risk factors for mental illnesses and behaviors that may be the result of untreated mental illnesses in children. Suicidal behavior is the most prominent marker of behavioral health problems. The rate of suicide attempts among high school children in Pennsylvania was reported to be 12% compared to 9.5% nationally.11 Another metric indicating behavioral health problems is chronic school absenteeism. In Pennsylvania, the share of 4th graders who are chronically absent is 31%.12 The national data show that in 2021, the corresponding estimate was 28% and has declined to about 24%.13 Additionally, parental incarceration is a risk factor for children. In Pennsylvania, the share of children with an incarcerated parent is 6%.14 Food insecurity has also been linked to stress and behavior problems in children. Roughly 17% of children in Pennsylvania were food insecure in 2021.15

Policies and programs addressing children’s mental health

In this section, we assess a set of policies and programs that are aimed at promoting the health of children. We focus specifically on evidence of the economic consequences of investing in children’s mental health. We organize our review of evidence in three parts: 1) the expansions of Medicaid to cover children during the 1980s and 1990s; 2) the expansion of Medicaid under the Affordable Care Act (ACA) that affected children’s mental health largely through coverage of their parents, in addition to an overview of the economic consequences of Medicaid expansion overall; and 3) programs that address the mental health of children that are linked to pre-school and K-12 education. Our review of evidence related to Medicaid expansion focuses on mental health and programs aimed at prevention and treatment of mental illness in children.

It is important to note that there are a large number of interventions in school-based programs and community contexts for which the evidence suggests that the programs are unlikely to be effective, let alone cost-effective. That is, the chances of the net benefits of the program being positive are low to modest. So, the choice of programs or interventions given the context is extremely important to realizing potential benefits. In the review of evidence below, we have focused on programs that are most likely to be cost-effective.

The economic framework we use specifically assesses the value of government program spending, known as the Marginal Value of Public Funds (MVPF) approach.16 This approach considers the value of benefits to program recipients and associated beneficiaries alongside the net cost to taxpayers. That is, it measures the direct program costs along with offsets such as increased tax collections or savings in other government programs. In this way, the MVPF measures the social benefits stemming from each extra dollar of net public spending on a particular program. The MVPF can be expressed as a ratio:

The higher the MVPF, the larger the benefit it provides to recipients relative to the cost to the government in the long run. Alternatively, if the MVPF is less than one, it is more costly to the government to provide per dollar of benefit than it is valuable to beneficiaries. Therefore, the MVPF can be interpreted as “bang for the buck.”17 This approach explicitly recognizes government budget constraints and longer-term budgetary effects. In the case of the child expansions of the 1980s and 1990s, the MVPF has a relatively straightforward interpretation because the Medicaid policy was targeted at children, so the direct budgetary impacts are well defined. In the case of more general expansions like those under the ACA, the impact on children must be considered alongside other benefits. Therefore, we will identify specific child impacts and note findings related to adults.

The goal of this review is to provide information on the economic consequences of policies and programs that effectively promote the mental health of children. We focus on two sets of results: first, the potential offsets in public spending from sources like future tax revenues and averted costs to public programs; and second, the MVPF. An example of an MVPF calculation can be found in Appendix A.

Child expansion from the 1980s and 1990s

Congress expanded Medicaid eligibility rules for children and pregnant women in the 1980s and added the Children’s Health Insurance Program (CHIP) in 1997. These notably expanded coverage for low-income children, and through coverage improvements, increased access to mental health services. These expansions can be used to illustrate the economic consequences of a larger safety net for children, particularly as they relate to mental health–related outcomes. Broadly speaking, research examining the long-term effects of the Medicaid eligibility expansions for children and pregnant women during the 1980s and 1990s uses variation in eligibility rules across states and birth cohorts. Those studies consistently show that the expansions resulted in benefits to recipients in education, earnings, and health. They also show benefits to taxpayers. Table 2 summarizes findings from analyses of Medicaid expansion for children in the 1980s and 1990s. In the discussion below, we highlight results that apply most directly to behavioral health and those showing offsets to public funds.

Medicaid policies that increase the coverage eligibility of children have general effects on the future well-being of children through improved health, school completion, and ultimate economic success, as measured by earnings. Analysis by Brown, Kowalski, and Lurie (2020) estimates that the combination of lower mortality, greater educational attainment, and higher earnings allows taxpayers to recoup $0.58 per $1 spent on Medicaid expansion. We apply the estimates obtained from Brown and colleagues to 2024 Pennsylvania data to estimate the benefits to beneficiaries of the policy per net spending on Medicaid expansion, the MVPF. Our estimate is that beneficiaries obtained $1.97 per $1 of net public spending on expanded Medicaid eligibility.18

The Congressional Budget Office (CBO), in a similar vein, used the experiences of expanding eligibility for children in the Medicaid program to create a model of long-term payoffs to taxpayers and individuals of improved earnings caused by more inclusive Medicaid eligibility. They estimate that taxpayers would recoup between $0.31 and $0.40 per $1 spent on new Medicaid eligibles by age 65.19 The improved economic outcomes for the affected individuals, alongside the net cost to the government, result in a MVPF of 4.11, suggesting $4.11 of beneficiary value for each $1 of Medicaid spending on expanded eligibility. To assist with interpreting, we can compare the MVPF of 4.11 here to the MVPF of 1.97 found using Brown, Kowalski, and Lurie’s estimates. The higher MVPF associated with more inclusive Medicaid eligibility indicates a smaller cost of the policy to the government relative to the benefit provided to the recipients.

We now turn to studies related to child Medicaid expansions that are more directly linked to behavioral health issues. The analysis by Hendrix and Stock (2024) is focused on the impacts of the eligibility expansion on criminal activity. They show that an extra year of Medicaid eligibility leads to reduced criminal activity between the ages of 19 and 24. They estimate reduced rates of property crime by 9%, drug crimes by 7%, and driving under the influence (DUIs) by 4%, all of which were statistically significant at conventional levels. The reductions in crime result in taxpayers’ overall recovery of $0.19 per $1 of Medicaid spending. Combining that with the impacts reported by Brown, Kowalski, and Lurie (2020), taxpayer recoveries amount to $0.60 for females and $0.75 for males. Applying these findings to Pennsylvania data, the MVPF suggests about $2 to $3 of beneficiary benefit per $1 of additional Medicaid spending on children.

Jácome (2024) also examines the returns to expanded Medicaid coverage for young men. She notes the high prevalence of mental illnesses among males involved with the criminal justice system. Jácome used estimates obtained from studying criminal activity levels following the loss of Medicaid coverage in 19-year-old men in South Carolina to estimate the impact of restoring coverage for that population. She found that coverage generated taxpayer recoveries that accounted for $0.17 per $1 of spending on expanded coverage (in 2024 dollars) in addition to $1.48 per $1 of net government spending in benefits to victims and communities through averted crime. Nearly the entire impact was due to reduced crime associated with young men with histories of mental health conditions. It is important to note that such state-specific estimates are highly dependent on the criminal justice system and health care contexts in which they are made.

The study by Hamersma and Ye (2021) examined whether the State Children’s Health Insurance Program (SCHIP) expansion for children resulted in more mental health care. Using the National Survey of American Families, they found little evidence of increases in specialty mental health visits but did find an increase in overall health care utilization.

Impact of ACA expansions in Medicaid on children

In this section, we consider how increases in Medicaid coverage, broadly, primarily through the ACA, affected children’s mental health–related circumstances. There are several mechanisms that could result in impacts on children. First, there is the “welcome mat” effect, which refers to when a parent obtains coverage, and there are positive spillovers to the entire family. A second mechanism is that when parents have access to care and services, they will obtain treatment for their own mental health conditions that could, in turn, reduce risk factors for children, such as family violence and maltreatment. A third mechanism is that many states support home visiting programs for new parents that can reduce stress and improve parenting skills, which have been shown to promote mental health in children. Table 3 summarizes evidence on the impact of broad coverage expansion on children.

The estimates reported in Table 3 reflect a combination of impacts that stem from the ACA Medicaid expansions, which were primarily aimed at adults but affected children indirectly. The ACA expansions were aimed at low-income adults up to 138% of the FPL. An analysis by Hudson and Moriya (2017) estimated an 8% increase in Medicaid enrollment among parents with children who obtained coverage through the ACA Medicaid expansions. This so-called “welcome mat” effect resulted in an increase in Medicaid enrollment of children associated with the ACA Medicaid expansion that was 1.87 times as large as the increase in states that did not expand Medicaid. One might expect impacts on longer-term outcomes that are similar to what was found in the studies reported in Table 2 (CBO, 2023; Brown, Kowalski, and Lurie, 2020).

Two studies in Table 3 focus on the impact of the ACA coverage expansion on child maltreatment, a key risk factor for mental illness and longer-term disadvantages. The results from the two studies, applied to the Pennsylvania context, indicate that the estimated economic payoff of reduced maltreatment is very sensitive to the estimated reductions in maltreatment associated with Medicaid coverage of adults. Both studies show significant reductions in child neglect rates. However, the larger impacts reported by McGinty and colleagues suggest a substantially larger economic payoff to the adult expansion compared to the Brown et al. (2019) results.20 Applying the McGinty estimates to Pennsylvania, the social benefits stemming from reduced child maltreatment were $1.49 per $1 of government spending on Medicaid. In contrast, the corresponding figure from the Brown et al. analysis was $0.32 per $1 spent on the ACA Medicaid expansion.

Finally, we report an analysis of extending Medicaid payment to in-school behavioral health care targeted at children of parents with opioid use disorder (OUD). The estimates produced by Meinhofer and colleagues (2025) suggest small but significant gains in school-based access to care, which are also associated with reduced emergency department (ED) visits.

Publicly funded school-based prevention and treatment programs

In this section, we report on four examples of school-based programs aimed at behavioral health conditions. We also report on a somewhat older literature that focused on early childhood that has shown strong results.21 One initial observation is that making the economic case for such programs depends critically on the specific design, target population, and implementation. The programs discussed here represent the potential economic consequences of putting in place well-designed programs. Programs that are not well-designed and well-executed will not achieve these results. In our review, we found a limited number of programs that showed significant gains in mental health outcomes for children and reduced the subsequent social costs. Our analysis of the MVPF is focused on the programs that show such gains in outcomes, as these should be the targets of investments. Table 4 summarizes our findings. We note that there are other types of prevention programs and health models for child mental health, targeted at pregnant mothers, infants, and early childhood, that we have not included in our set of examples here.

Observations and policy considerations

The most important observation from the evidence offered in this paper is that there are significant payoffs to maintaining coverage for children. The evidence reported in Tables 2 and 3 makes that clear. Where the MVPF was calculated for coverage expansions focused on children, the values of benefits exceed the claims on public budgets. This, in part, results from long-run impacts on earnings and criminal justice involvement that offset some of the front-end costs of coverage in childhood. Therefore, efforts to maintain coverage of children pay off with respect to producing benefits of value at or above the program costs to the government.

The second major observation derived from the evidence review is that some well-designed programs that target early childhood and bring clinical expertise into schools to address problems like suicide-related behavior also provide good “bang for the public buck” (Table 4). Some targeted pre-K programs and high-risk parental training and support interventions modeled on well-established evidence have very large payoffs. The pre-K program, the Good Behavior Game, and the parental training program targeting risk of maltreatment all pay for themselves in the long term.

What is important to note is that program selection and design are determinative. We reviewed dozens of programs targeted at the same populations that did not provide strong evidence of similar payoffs. The school-based mental health treatment program in Minnesota was notable because it differentiated between expertise brought in from community resources and care delivered by school system personnel. The result was a division of labor between institutions that are focused on education and those with a mission concerned with the care of children with mental illnesses and related problems. The implication is that investing in carefully selected and well-implemented early childhood programs and school treatment can produce high levels of benefit to program participants, their families, communities, and taxpayers.

Appendix A: Example of developing the MVPF using Golberstein et al. (2023) study

The study by Golberstein et al. (2024) focused on a school-based mental health diagnosis and treatment program implemented in Minnesota’s Hennepin County. The program involved partnerships between 17 school districts and local mental health systems to place clinical staff in schools. The population involved children aged 5 to 18 years. The key outcome in our extension of the Golberstein et al. results was the impact of the program on suicide attempts. The estimated impact (Average Treatment Effect on the Treated, or ATT) was a reduction of 15% in suicide attempts that was statistically significant at conventional levels (0.05).

Baseline

Applying the estimates to the Pennsylvania context, we focus on high school-age children because the rates of suicide attempts are much higher than for younger age groups. The suicide attempt rate among Pennsylvania high school students is 8%.23 Data from the National Center for Education Statistics from 2022 to 2023 report that there were 558,849 high school students in Pennsylvania, implying 44,707 suicide attempts in 2022. Golberstein et al. report that the cost of the school-based mental health initiative was $117 per student in 2019, the last year of their data collection, which translates into roughly $144 in 2024, adjusting for inflation by the CPI-U. There are few American studies of the long-term impact on earnings of suicide attempts by children. One study conducted a longitudinal analysis of earnings from 2002 to 2017 in a sample from Quebec.24 They evaluated the impact of suicide attempts on earnings through age 37 and then calculated the Present Discounted Value (PDV) over a 40-year career in U.S. dollars. The result was an impact estimate of a loss of $98,000 in individual earnings. Adjusting for inflation by the CPI-U yielded an estimate of $145,000 in 2024.25

Program impacts

In applying the various parameters from Golberstein et al. and Orri et al., we assume the full cost of the program was paid for by public dollars (federal, state, and local). We also note that the impacts were for a single cohort of students exposed to the program. The Golberstein et al. estimates were a 15% reduction in suicide attempts, or 6,706 children, when applied to the Pennsylvania high school population. The PDV of earnings gains from the reduction of suicide attempts amounted to about $972.4 million. Applying the program cost to the high school population yielded a cost estimate of $80.47 million. Yet in order to obtain the net public funding impact, we need to adjust the cost estimate to reflect the tax collections realized from the increased earnings stemming from the program impacts. So, the elements of the MVPF are as follows:

2. PDV of beneficiary net earnings gains

= Total earnings – tax impacts of earnings gains

= $972.4 million – $327.4 million

= $645 million

3. PDV of net program costs

= Total program costs – tax impact of earnings gains

= $80.5 million – $327.4 million

= –$246.9 million

Interpretations and cautions

Assembling an MVPF from varied estimates and contexts should be kept in mind when interpreting results. In applying Minnesota findings to Pennsylvania, we assumed that the Hennepin County program cost estimates would be the same in Pennsylvania (inflation-adjusted to 2024). The overall cost of living in Minneapolis is about 5% lower than in Pennsylvania overall. The lower rates of suicide attempts in Hennepin County suggest that it may be easier and less costly at the margin to reduce suicide attempt rates in Pennsylvania. The suicide attempt rates in Quebec are higher than they are in Pennsylvania, which may have implications for earnings estimates, but that is unknown. The average earnings in Pennsylvania are higher (exchange rate adjusted) than in Quebec, which implies that our results may be underestimates of the earnings gains.