6

The Economic Costs of the Opportunity Gap¹

This chapter presents the committee’s estimation of the economic costs posed by the opportunity gap, using traditional approaches from cost-benefit analysis. These numbers, appropriately discounted, mark the upper bound of, or the maximum of, the benefit side of a calculation of the value of “investing in strategies, interventions, and policies to address opportunity gap concerns for children from birth to age 8,” which was part of the committee’s statement of task (Box 1-1 in Chapter 1). If interventions could entirely close this gap, that would represent the benefit side of the calculation, and the price of carrying out those interventions would represent the cost side (along with any unintended consequences).

After presenting these estimates, the chapter draws on existing high-quality causal research to draw inferences about a range of the share of this benefit that could be achieved by policy intervention, which likely is below 1 but surely is well above 0. The committee faced several challenges in finding this evidence. First, not all of this work is based on randomized controlled trials (RCTs), which are often viewed as the gold standard for research evidence. Thus, we had to rely on quasi-experimental studies and other approaches, and not all differences in outcomes can be examined in this fashion. Second, data are not always available with which to document differences in outcomes in order to consider the results of either RCT interventions or quasi-experimental designs, so the upper bound described above would be underestimated if the largest effects were for difficult-to-measure

¹ This chapter is adapted from a paper commissioned by the committee for this study (Bitler & Oh, 2022).

outcomes. Third, while the public costs of interventions are often easy to measure and are incurred right away, many of the benefits of these investments take many years to materialize. For example, if an intervention has intergenerational benefits, as has been documented for the Medicaid expansions of the late 1980s and early 1990s (e.g., East et al., 2023), it could take multiple decades for these benefits to materialize. Finally, some of the evidence stems from settings in the past (e.g., policy changes during the war on poverty leading to long-run improvements) or in different contexts (e.g., evidence about the effects of parental leave from Europe), and the literature on how well estimates from one setting apply to another is underdeveloped. Thus, there is a tension between high-quality evidence that provides comprehensive evaluations of total net benefits and interventions that fit today’s context and setting.

In the remaining sections of this chapter we briefly discuss approaches that have been used traditionally to evaluate these trade-offs (cost-benefit analysis, cost-effectiveness analysis), and then touch on the marginal value of public funds (e.g., Hendren & Sprung-Keyser, 2020), provide representative calculations from the literature for three interventions, and present the conclusions that can be drawn from these findings.

THE OPPORTUNITY GAP: COST ESTIMATES

This section provides rough estimates of the economic costs of the opportunity gap that are due to mortality and morbidity as a percentage of gross domestic product (GDP). We derive these estimates by determining a possible upper bound for the cost of disparities in education; health; and social-emotional development, well-being, and mental health.

Obviously, it is difficult to find an ethical way of determining the value to society of lives lost, but given this task, we require some way to value these differences quantitatively across groups. For purposes of regulatory impact evaluation, for example, the Department of Health and Human Services (HHS; 2016) summarizes the value of a statistical life (VSL) and value of a statistical life year appropriate for considering the benefits of reducing risks of mortality and morbidity through interventions.² The calculation is derived from the thought experiment of measuring a person’s willingness to pay to reduce the risk of dying from a particular cause or condition, and VSL is then presumed to be constant for small differences in risk. The midpoint in the review cited in these guidelines in 2014 dollars is $9.3 million,

² As stated in the HHS guidelines, “Values for mortality risk reduction reflect the rate of tradeoff between money and small changes in mortality risk, referred to as the marginal rate of substitution between wealth and risk” (Hammitt, 2000, p. 13). The guidelines describe many caveats and issues associated with using VSL (and quality-adjusted life year).

with a range of $4.4–14.2 million. It is important to note that this is not a value to always be applied to a certain mortality reduction. When the mortality reduction and/or any costs occur in a different time period than the present, the costs must be discounted to the present.

Morbidity reductions (improvements in health) are more difficult to value, but if alternatives are not available, the HHS guidance suggests using estimates of the monetary value of a quality-adjusted life year (QALY). QALY is a measure that incorporates “duration and severity of illness” to make it possible to “compare health status across individuals or population groups,” and represents the value of one more year in perfect health (Department of Health and Human Services, 2016, p. 18). The total QALYs of experiencing a particular condition for a specific length of time are calculated by determining how much time a person spends in that condition and multiplying that time by the health-related quality of life associated with the condition—a number ranging from 0 (death) to 1 (full health), measured from populations experiencing the condition. Thus the QALY gain of an intervention is the sum of the change in QALY across a statistical person’s lifespan after versus before the intervention of interest, and then for regulatory purposes, the future is discounted as in other cost-benefit or cost-effectiveness calculations. Table 3.2 of the guidelines (HHS, 2016) reports a monetized value per QALY in 2014 dollars, which ranges from $490,000 to $820,000 for average VSL.

Chapters 1 through 5 of this report summarize existing work on the existence and sources of differences across groups—racial/ethnic, immigrant/nonimmigrant, place of birth/residence, and family advantage (human capital, income, wealth). The simple statistics document significant differences in health outcomes by many of these characteristics. For example, a National Center for Health Statistics Data Brief from 2011 summarizes infant mortality rates by race and Hispanic origin of the mother, using 2007 data (MacDorman & Matthews, 2011). Babies born to non-Hispanic Black mothers had an infant mortality rate (mortality in the first year of life) of 13.31/1,000 live births. By comparison, the rate for babies born to American Indian or Alaska Native mothers was 9.22/1,000 live births; for those born to Puerto Rican mothers was 7.71/1,000 live births; for those born to non-Hispanic White mothers was 5.63/1,000 live births; and for those born to Mexican, Cuban, Asian or Pacific Islander, or Central and South American mothers was even lower (MacDorman & Matthews, 2011). Were we simply to use the middle level of VSL, moving the infant mortality rates for non-Hispanic Black and American Indian or Alaska Native mothers to the average would result in a savings of $67 billion in 2014 dollars. We note further that these different groups have different levels of low birthweight and preterm birth, which represent additional risks, so eliminating these

additional infant health differences would result in additional savings that would be much more difficult to value.

Next, we considered differences in mortality for children aged 1–8. Data on mortality rates in 2018 for children aged 1–8 from the Centers for Disease Control and Prevention’s WONDER database were used to assess the savings from moving mortality for groups with higher levels to the mean. With these 2018 data, the mortality rate for this age group was 17.7 per 100,000. Non-Hispanic Black children had a mortality rate nearly twice as high, 31.6 per 100,000. American Indian and Alaska Native children also had an elevated rate, at 18.3 per 100,000, and Native Hawaiian or other Pacific Islander children had higher mortality than the average, at 20.39 per 100,000. Non-Hispanic White children, those of more than one race, Asian children, and Hispanic White and Black children had lower rates than the average. Eliminating this disparity would lead to a savings of $5.8 billion in 2014 dollars. The mortality savings alone would yield $72.8 billion in savings, or 0.4% of GDP in 2014, based on these rough estimates.

Earlier chapters of this report also review at length other disparities among young children in learning, physical health, safety and security, social-emotional development, and other outcomes. Eliminating these disparities would undoubtedly add further savings. Monetizing these outcomes is even more complicated, however, and requires many more assumptions, especially as assessing morbidity requires using a specific discount factor, such as that suggested by the Office Management and Budget. There is also interest in measuring the effects of these gaps on children’s development and future earnings. Assessing the benefits of avoiding outcomes such as low birthweight due to follow-on effects on outcomes across the life course, such as earnings and children’s development, is complicated, and raises issues about double counting. If averting low birthweight, low gestational age, and other health differences at birth affects later adult earnings, one should count either the properly measured present discounted value of the lower birthweight or lower gestational age and other health differences or the positive carry-on effects of these improved outcomes, but not both.

SHARE OF OPPORTUNITY GAPS AND THEIR COSTS THAT CAN BE AFFECTED BY POLICY

The previous section presented some examples of calculations of the net savings to society from closing opportunity gaps, detailed some of the mortality differences across groups that would be eliminated, and described challenges with making even rough estimates for valuing benefits of these changes. This section addressed what is known from causal research about the share of opportunity gaps that policy could plausibly be expected to affect. We are certain the number is much higher than 0, but also suspect

it is not 1. Currie (2005) discusses the extent to which disparities in health contribute to racial gaps in school readiness, exploring a host of health conditions as well as maternal health and behaviors and suggests these factors may account for as much as a quarter of the racial gap in school readiness. If this is the case, policy interventions around birth and health after birth could at most address one-fourth of this gap. Currie points out that incomplete take-up—the fact that many individuals eligible for programs do not participate in them—is also an important issue. This idea is worth considering when suggesting interventions, in that universal rather than targeted programs may be more effective. However, well-targeted programs with low barriers to take-up appear promising.

It is of course challenging to understand the right share of the savings from eliminating disparities that can be attributed to particular interventions, and the share likely varies across programs, populations, and settings. But any program that will eliminate disparities must either leave advantaged children unaffected or have greater impacts on the disadvantaged groups.

APPROACHES TO EVALUATING THE NET BENEFITS OF INVESTMENT IN MITIGATING THE OPPORTUNITY GAP

Even though a specific policy intervention might prove to be highly effective in mitigating the opportunity gap, it may not be feasible if it incurs tremendous costs relative to benefits. On the other hand, if a long-term benefit generated by a policy is substantial and even greater than its costs, investing in that policy could have the dual advantage of closing the opportunity gap while not overtaxing a government budget. To evaluate the effectiveness of various public programs quantitatively, social scientists use welfare analysis. In this section, we summarize three methods commonly used to measure the welfare impact of a public policy: cost-benefit analysis, cost-effectiveness analysis, and the marginal value of public funds (MVPF) framework. Although results of these welfare analyses should be interpreted carefully and cautiously because they depend on strong assumptions, the analyses are valuable as they provide a guideline for policy makers.

Cost-benefit analysis is a traditional method used in public policy. It is also commonly adopted when governments attempt to assess whether proposed regulations or policies are “worth it.” From this perspective, interventions for which the present discounted value of the costs is outweighed by the present discounted value of the benefits are considered worthwhile. Suppose the total benefit of a policy is estimated to be greater than its cost, so that its cost-benefit ratio is larger than 1. In that case, the policy is deemed to generate a positive net social benefit. To encompass benefits and costs generated over the lifetime of a program, researchers usually calculate the present discounted value of the expected benefits and costs associated with

the program. Importantly, costs are not accounting costs, but incorporate the opportunity cost of the resources used (their next-best alternative use). Costs and benefits are discounted, as a dollar today is worth an amount 1 + r next year, where r is the rate of return if the dollar were not used on this program. Should the government need to borrow, the relevant interest rate is the net of inflation rate, and a large literature focuses on the right rate and how to calculate it in the short versus the long run. Additionally, it is important to note that this simple approach does not take account of the fact that different persons may experience the costs and the benefits differently, and many of the estimates, including those around the value of saving lives, face challenges with representing uncertainty.

Cost-benefit analysis is very intuitive and easy to understand. However, reporting of the costs and benefits is not comprehensive, making it difficult to compare the gain or loss in social welfare across different welfare policies because different strands of literature usually adopt different measures. For instance, studies on early childhood education programs such as Head Start report costs and benefits using the internal rate of return or the ratio of benefit to cost of a program (Heckman et al., 2010; Carneiro & Ginja, 2014). On the other hand, studies on taxes often focus on the marginal deadweight burden or marginal efficiency cost of funds (Saez, Slemrod, & Giertz, 2012).

Cost-effectiveness analysis (often referred to as cost-utility analysis) is another method commonly used for welfare analysis. When the benefits are considered difficult to monetize or, alternatively, it is agreed that the benefits are useful, cost-effectiveness analysis is a good alternative to cost-benefit analysis. With this method, the lowest-cost way of achieving a common goal is preferred. For instance, two options for reducing infant deaths by 50% can be considered: expanding Medicaid coverage to disadvantaged pregnant women or expanding Medicaid coverage to families with newborn babies. If the former could achieve the goal at less cost than the latter, the cost-effectiveness analysis would consider the first policy more effective. Since cost-effectiveness analysis is particularly useful when it is difficult to measure the benefits of a policy quantitatively, it is commonly adopted in health economics literature, where it is difficult or sometimes perceived as impossible to measure the value of the life or health of a human being quantitatively. For example, estimates of the effects of Medicaid expansion sometimes address cost-effectiveness by converting impacts into costs per life saved (Currie & Gruber, 1996) or a more simple estimate of the number of deaths that could have been avoided (Miller, Johnson, & Wherry, 2021).

Finally, the MVPF framework aims to improve the comparability of net benefits across programs. Hendren and Sprung-Keyser (2020) suggest use of this framework as a unified method for welfare analysis of different government policies (see Finkelstein & Hendren, 2020, for more

detail). A mathematical definition of MVPF is the ratio of the benefit of a specific policy to its net cost, accounting for any policy externalities (such as spillovers from a given intervention beyond the direct intended effects). The “net” cost in the denominator is calculated by adding up any negative externalities or subtracting positive externalities induced by the policy. For example, if increasing the benefit provided by a cash transfer program makes a recipient work less and thus decreases the labor supply, which in turn reduces tax revenue, then the net cost of the program to the government will be greater than the direct cost of program implementation. By contrast, if individuals work more and thereby increase the labor supply to be eligible for additional benefits from the earned income tax credit, that policy externality will decrease the net cost of that program.³

Intuitively, the larger the MVPF, the greater the net benefits generated by a policy. The maximum possible value of an MVPF is infinity, which implies that a long-term benefit from a policy is significant enough to cover the policy’s initial and long-term costs. In this case, the policy helps the government save money over time. If the MVPF of a welfare program is larger than 1, the program generates larger benefits than its cost; if the MVPF is smaller than 1—for instance, 0.5—the program generates $0.50 worth of benefits for every $1.00 spent. If a program has large or even infinite MVPF, there is no reason not to adopt it. However, a low MVPF does not necessarily mean that a program is worthless; the program’s worth depends on values the society pursues.

EVIDENCE FOR WHETHER IMPLEMENTATION MAKES SENSE: REPRESENTATIVE EXAMPLES OF EFFECTIVE POLICIES IN THREE ARENAS

Many high-quality causal studies have found that the safety net programs providing resources during early life have both short- and long-term positive effects on children’s later-life outcomes, including labor market outcomes, health, and educational achievement. This section summarizes findings about the effectiveness of such programs in closing the opportunity gap. In addition, we provide findings from previous research about whether a certain welfare program generates a larger benefit than its cost based on MVPF estimates. We focus on three major U.S. welfare programs: the Supplemental Nutrition Assistance Program (SNAP, formerly known

³ Although the mathematical formulas of the MVPF framework and the ratio of benefits to costs—an alternative way to carry out cost-benefit analysis—are quite similar, one big difference between the two lies in how positive externalities of a government policy are incorporated. The MVPF includes any positive externalities in its denominator together with any negative externalities. In contrast, the ratio of benefit to cost sums up all positive externalities in its numerator while placing negative externalities in the denominator.

as Food Stamps), Medicaid, and Head Start. These programs were chosen because there is a large body of work on them in the U.S. context, but we note that it is likely worth considering many other programs as well, such as family or parental leave, that have less complete U.S. evidence but have been shown to be effective.

Reducing the Opportunity Gap Due to Income Differences through Cash and Near-Cash Means-Tested Assistance

Unconditional cash assistance programs are traditional welfare programs in the United States. A typical example is Aid to Families with Dependent Children (AFDC), replaced by Temporary Assistance for Needy Families (TANF), a block grant program, in 1996. The AFDC program transferred a cash benefit to families with dependent children aged 0–18 as long as their income was below a certain threshold. Public spending on and participation rates for AFDC/TANF fell significantly after the 1996 welfare reform since the TANF program was no longer an entitlement, and many states chose to funnel spending on that program to other purposes. Nonetheless, some studies have found long-term positive effects of exposure to cash assistance programs during early childhood before this change occurred. For instance, Aizer and colleagues (2016) showed that people who received cash transfers during childhood from the Mother’s Pension program, a predecessor of AFDC operated between 1911 and 1931, experienced significant improvements in health, educational attainment, and income in adulthood. Moreover, the magnitude of those effects was greatest for children from the poorest families and those who were exposed at younger ages. These findings imply that this cash transfer program closed some portion of the opportunity gap in health, education, and adult income for children from the poorest families.

A large body of literature on the effects of the SNAP/Food Stamp program provides valuable information about its effects in closing many aspects of the opportunity gap. SNAP currently provides a voucher that can be used to purchase food by low-income families that have gross incomes below 130% of the federal poverty threshold or otherwise satisfy gross income limits and have net incomes below 100% of the federal poverty threshold. Starting in 1961 under the Kennedy administration as a core mechanism of the war on poverty, this program has supplemented the food budgets of many disadvantaged families. Leveraging the quasi-random variation of the program’s rollout, researchers have found that exposure to Food Stamps in utero and in early childhood was particularly effective in improving adult health, educational attainment, and labor market outcomes (Hoynes, Schanzenbach, & Almond, 2016; Bitler & Figinski, 2019; Bailey et al., 2020). This program has been highly effective in improving outcomes

in early and later life; however, it is worth noting that its effects have been found to be heterogeneous across different subgroups.

What are the cumulative benefits of exposure to SNAP starting in early childhood? Using their estimates of increased income and longevity and reduced incarceration rates and public assistance spending, Bailey and colleagues (2020) estimated that the MVPF of providing SNAP from in utero up to age 5 is 56, which implies that for each $1 spent on the program, children exposed to the program in early childhood garner benefits worth $56 in the long run.

Reducing the Opportunity Gap Due to Inequitable Access to Health Care and Health Insurance through Medicaid

Since being established in 1965 and gradually adopted by states, Medicaid has provided basic health insurance to low-income individuals. Medicaid coverage was closely tied to AFDC participation until the early 1980s, but eligibility dramatically expanded to a broader population (Hoynes & Schanzenbach, 2018). Today, Medicaid is the most extensive safety net program for low-income families in terms of the amount of federal spending: according to the Congressional Budget Office, Medicaid spending for children totals $89 billion annually (Hoynes & Schanzenbach, 2018) and covers almost half of all births in the United States (Markus et al., 2013).

Focusing on the introduction of Medicaid and its expansions during the 1980s and 1990s, a body of research has found that exposure to the program in utero and/or in early childhood had substantial positive effects on the educational attainment, adult health, and earnings in early adulthood of exposed children (Meyer & Wherry, 2012; Miller & Wherry, 2018; Brown, Kowalski, & Lurie, 2020; Goodman-Bacon, 2021). For instance, Goodman-Bacon (2021) provides specific estimates of the public savings due to the introduction and expansion of Medicaid, respectively, in the long run. He estimates that the introduction of the Medicaid program yielded a total of 10 million additional QALYs, meaning that the program resulted in 10 million more years of life in perfect health compared with a counter-factual world without Medicaid. East and colleagues (2023) show that the program’s effects are even multigenerational, as women who experience in utero exposure to Medicaid are more likely to have healthy babies with respect to birthweight and gestational age. This intergenerational transmission of positive health impacts implies even more considerable social benefits from the program. These findings strongly indicate that the Medicaid program may be particularly effective in reducing the opportunity gap by race and income in health, educational attainment, and earnings by improving outcomes in adulthood of exposed children, most of whom are likely to be from vulnerable families.

The enormous private and public benefits of Medicaid are captured by an infinite MVPF, meaning that its fiscal externalities are large enough to fully cover its initial cost (Hendren & Sprung-Keyser, 2020). In addition, according to the estimates of Goodman-Bacon (2021), Medicaid saved more than twice its initial cost for 1976 birth cohorts. Brown, Kowalski, and Lurie (2020) focus on the expansion of the Medicaid program during the 1980s and estimate that $1.00 in spending on Medicaid recoups 0.90 cents by age 28, meaning that 56% of the expenditure for the 1980s expansion would be paid after 60 years.

Reducing the Opportunity Gap Due to Unequal Access to Early Education through the Head Start Program

Launched in 1965 as part of the war on poverty, the Head Start program has offered comprehensive child development services for children from low-income families from birth to age 5 with the aim of promoting the school readiness of disadvantaged children. The program encompasses not only early learning but also health, nutrition, and family well-being. Approximately 900,000 children aged 3–5 participate in the program annually; the program’s budget in fiscal year 2022 was about $11 billion (Gibbs et al., 2016; Administration for Children and Families, 2022).

Many studies have found positive causal effects of exposure to Head Start on children’s long-run outcomes, although these findings should be interpreted cautiously as many derive from retrospective analyses. Findings include substantial impacts in terms of increases in high school graduation rates, college attendance, earnings, and household income (Currie & Thomas, 1995; Garces, Thomas, & Currie, 2002; Deming, 2009; Thompson, 2018; Johnson and Jackson, 2019; Bailey, Sun, & Timpe, 2021); improvements in health (Ludwig & Miller, 2007; Carneiro & Ginja, 2014); and reduction in crime rates (Deming, 2009; Carneiro & Ginja, 2014; Johnson & Jackson, 2019). Effects are larger for more disadvantaged children and women (De Haan & Leuven, 2020) and for African American children (Bauer & Schanzenbach, 2016).

The above findings imply that the early-life intervention of the Head Start program has been effective in mitigating opportunity gaps in education, earnings, and health across income and racial groups. However, another strand of studies suggests that the effects of the Head Start program have not lasted long. For instance, results from the Head Start Impact Study, a randomized evaluation of Head Start conducted in 2002, indicate that the immediate gain in test scores that followed participation in the program faded out quickly in the following years (HHS and Administration for Children and Families, 2010; Puma et al., 2012). Although such findings do not negate the Head Start program’s contribution to closing the opportunity

gap in early educational attainment, they do raise concern about the program’s cost-effectiveness (Haskins, 2004; Barnett, 2011). It would be useful to follow up with the children included in the Head Start Impact Study to see whether positive effects reemerge in the longer run.

Although there is not yet clear consensus on the size of the net social benefit of the Head Start program, two recent studies suggest that the program pays for itself. Kline and Walters (2016) estimate that the MVPF measured from the Head Start Impact Study data is well above 1 (1.86 according to their preferred estimation). The MVPF of 1.86 implies that expansions of the program are likely to improve social welfare by generating $1.86 in benefits for every $1.00 spent. Bailey, Sun, and Timpe (2021) provide a traditional cost-benefit analysis using the staggered rollout of the Head Start program between 1965 and 1980. Focusing on gains in the accumulation of human capital and improvement in economic self-sufficiency, they estimate that the Head Start program generates positive fiscal externalities by reducing public assistance expenditures and increasing tax revenue (from approximately $576 to $2,331 annually per program participant) as a result of the additional wages earned by the exposed children. This estimated gain could be larger if the positive effects of the Head Start program are transmitted intergenerationally (Barr & Gibbs, 2022) or if the program is a good complement to other programs, such as Medicaid (Bailey, Sun, & Timpe, 2021).

CONCLUSIONS

This chapter has examined—at the broadest level—the added benefit accompanying a reduction in infant and child mortality, showing substantial benefit even when conservative estimates of VSL are employed. The overall annual benefit of a reduction in infant mortality could safely be estimated at approximately $67 billion, while a reduction in child mortality (ages 1–8) would result in a savings of approximately $73 billion per year.

Our review of the literature on the evaluation of social welfare programs focused on the three most heavily studied programs: SNAP, Medicaid, and Head Start. Our analysis found that the benefits of these programs exceed the costs, although to varying degrees. For SNAP/Food Stamps, the evidence shows (while results differ across different subgroups in the population) that every $1 spent on the program results in about $56 in benefits across the life course. Medicaid (and its expansions) has been shown to have very high rates of return—often exceeding the initial costs during the program’s early years—due to improved health and longevity. Finally, the potential benefits from funding the Head Start program also tend to exceed the initial costs; an initial investment of $1 yields a return of about $1.86, indicating that the program is more than cost-effective.

The analyses presented in this chapter provide a conservative estimate of the potential benefits of targeted interventions aimed at closing the opportunity gap for children aged 0–8. Overall, research evidence provides some quantification of the monetary and fiscal returns to reducing the opportunity gap for these young children. These benefits accrue at the level of the individual, the family, and society. Analysis of additional policies and programs, while sometimes challenging, is a key area for future research related to understanding the economic costs of the opportunity gap.

REFERENCES

Administration for Children and Families. (2022). FY 2022 Head Start funding increase ACF-PI-HS-22-02. Department of Health and Human Services. Available: https://eclkc.ohs.acf.hhs.gov/policy/pi/acf-pi-hs-22-02

Aizer, A., Eli, S., Ferrie, J., & Lleras-Muney, A. (2016). The long-run impact of cash transfers to poor families. American Economic Review, 106(4), 935–71. https://doi.org/10.1257/aer.20140529

Bailey, M., Hoynes, H., Rossin-Slater, M., & Walker, R. (2020). Is the social safety net a long-term investment? Large-scale evidence from the food stamps program. (Working Paper No. 26942). National Bureau of Economic Research. https://doi.org/10.3386/w26942

Bailey, M.J., Sun, S., & Timpe, B. (2021). Prep school for poor kids: The long-run impacts of head start on human capital and economic self-sufficiency. American Economic Review, 111(12), 3963–4001. https://doi.org/10.1257/aer.20181801

Barnett, W.S. (2011). Effectiveness of early educational intervention. Science, 333(6045), 975–8. https://doi.org/10.1126/science.1204534

Barr, A., & Gibbs, C. (2022). Breaking the cycle? The intergenerational effects of an anti poverty program in early childhood. Journal of Political Economy, 130(12). Available: http://people.tamu.edu/~abarr/Head%20Start%20Intergen_1_31_2022.pdf

Bauer, L., & Schanzenbach, D.W. (2016). The long-term impact of the Head Start program: Economic analysis. The Brookings Institution, The Hamilton Project. https://www.hamiltonproject.org/papers/the_long_term_impacts_of_head_start?_ga=2.95710620.1446079228.1670529729-979225845.1663104239

Bitler, M., & Oh, S. (2022). [Economic costs of the Opportunity Gap]. Commissioned paper for the Committee on Exploring the Opportunity Gap for Young Children from Birth to Age 8.

Bitler, M.P., & Figinski, T.F. (2019). Long-run effects of food assistance: Evidence from the Food Stamp program economic self-sufficiency policy research institute. University of California at Davis. Available: https://arefiles.ucdavis.edu/uploads/pub/2021/02/09/bitler-figinski-fsp-2019-8-29-full-paper.pdf

Brown, D.W., Kowalski, A.E., & Lurie, I.Z. (2020). Long-term impacts of childhood Medicaid expansions on outcomes in adulthood. The Review of Economic Studies, 87(2), 792–821. https://doi.org/10.1093/restud/rdz039

Carneiro, P., & Ginja, R. (2014). Long-term impacts of compensatory preschool on health and behavior: Evidence from Head Start. American Economic Journal: Economic Policy, 6(4), 135–73. https://doi.org/10.1257/pol.6.4.135

Currie, J. (2005). Health disparities and gaps in school readiness. Future Child, 15(1), 117–38. https://doi.org/10.1353/foc.2005.0002

Currie, J., & Gruber, J. (1996). Saving babies: The efficacy and cost of recent changes in the Medicaid eligibility of pregnant women. Journal of Political Economy, 104(6), 1263–96. http://www.jstor.org/stable/2138939

Currie, J., & Thomas, D. (1995). Does Head Start make a difference? American Economic Review, 85(3), 341–64. http://www.jstor.org/stable/2118178

De Haan, M., & Leuven, E. (2020). Head Start and the distribution of long-term education and labor market outcomes. Journal of Labor Economics, 38(3), 727–65. http://dx.doi.org/10.1086/706090

Deming, D. (2009). Early childhood intervention and life-cycle skill development: Evidence from Head Start. American Economic Journal: Applied Economics, 1(3), 111–34. https://doi.org/10.1257/app.1.3.111

Department of Health and Human Services (HHS). (2016). Guidelines for regulatory impact analysis. Office of the Assistant Secretary for Planning and Evaluation. Available: https://aspe.hhs.gov/reports/guidelines-regulatory-impact-analysis

Department of Health and Human Services (HHS), Administration for Children and Families. (2010). Head Start impact study: Final report. Available: https://www.acf.hhs.gov/opre/report/head-start-impact-study-final-report-executive-summary

East, C.N., Miller, S., Page, M., & Wherry, L.R. (2023). Multigenerational impacts of childhood access to the safety net: Early life exposure to Medicaid and the next generation’s health. American Economic Review, 113(1), 98–135. https://doi.org/10.1257/aer.20210937

Finkelstein, A., & Hendren, N. (2020). Welfare analysis meets causal inference. The Journal of Economic Perspectives, 34(4), 146–67. https://www.jstor.org/stable/26940894

Garces, E., Thomas, D., & Currie, J. (2002). Longer-term effects of Head Start. American Economic Review, 92(4), 999–1012. Available: http://www.jstor.org/stable/3083291

Gibbs, C., Ludwig, J., Miller, D.L., & Shenhav, N. (2016). Short-run fade-out in Head Start and implications for long-run effectiveness. [Policy Brief]. Center for Poverty, University of California at Davis. Available: https://poverty.ucdavis.edu/policy-brief/short-run-fade-out-head-start-and-implications-long-run-effectiveness

Goodman-Bacon, A. (2021). The long-run effects of childhood insurance coverage: Medicaid implementation, adult health, and labor market outcomes. American Economic Review, 111(8), 2550–93. https://doi.org/10.1257/aer.20171671

Hammitt, J.K. (2000). Valuing mortality risk: Theory and practice. Environmental Science & Technology, 34(8), 1396–400. https://doi.org/10.1021/es990733n

Haskins, R. (2004). Competing visions. The Brookings Institution. Available: https://www.brookings.edu/articles/competing-visions/

Heckman, J.J., Moon, S.H., Pinto, R., Savelyev, P.A., & Yavitz, A. (2010). The rate of return to the High/Scope Perry Preschool Program. Journal of Public Economics, 94(1–2), 114–28. https://doi.org/10.1016/j.jpubeco.2009.11.001

Hendren, N., & Sprung-Keyser, B. (2020). A unified welfare analysis of government policies. The Quarterly Journal of Economics, 135(3), 1209–318. https://doi.org/10.1093/qje/qjaa006

Hoynes, H., & Schanzenbach, D. (2018). Safety net investments in children. Brookings Papers on Economic Activity, 2018, 89–150. Available: https://www.brookings.edu/bpea-articles/safety-net-investments-in-children/

Hoynes, H., Schanzenbach, D.W., & Almond, D. (2016). Long-run impacts of childhood access to the safety net. American Economic Review, 106(4), 903–34.

Johnson, R.C., & Jackson, C.K. (2019). Reducing inequality through dynamic complementarity: Evidence from Head Start and public school spending. American Economic Journal: Economic Policy, 11(4), 310–49. https://doi.org/10.1257/pol.20180510

Kline, P., & Walters, C.R. (2016). Evaluating public programs with close substitutes: The case of Head Start. The Quarterly Journal of Economics, 131(4), 1795–848. https://doi.org/10.1093/qje/qjw027

Ludwig, J., & Miller, D.L. (2007). Does Head Start improve children’s life chances? Evidence from a regression discontinuity design. The Quarterly Journal of Economics, 122(1), 159–208. https://doi.org/10.1162/qjec.122.1.159

MacDorman, M.F., & Matthews, T.J. (2011). Understanding racial and ethnic disparities in U.S. infant mortality rates. (NCHS Data Brief No. 74). Hyattsville, MD: National Center for Health Statistics.

Markus, A.R., Andres, E., West, K.D., Garro, N., & Pellegrini, C. (2013). Medicaid covered births, 2008 through 2010, in the context of the implementation of health reform. Womens Health Issues, 23(5), e273–80. https://doi.org/10.1016/j.whi.2013.06.006

Meyer, B., & Wherry, L. (2012). Saving teens: Using a policy discontinuity to estimate the effects of Medicaid eligibility. (Working Paper No. 18309). National Bureau of Economic Research. Available: https://www.nber.org/papers/w18309

Miller, S., & Wherry, L.R. (2018). The long-term health effects of early life Medicaid coverage. Journal of Human Resources, 54(3), 785–824. https://doi.org/10.3368/jhr.54.3.0816.8173R1

Miller, S., Johnson, N., & Wherry, L.R. (2021). Medicaid and mortality: New evidence from linked survey and administrative data. The Quarterly Journal of Economics, 136(3), 1783–829. https://doi.org/10.1093/qje/qjab004

Puma, M., Bell, S., Cook, R., Heid, C., Broene, P., Jenkins, F., Mashburn, A.J., & Downer, J.T. (2012). Third grade follow-up to the Head Start impact study: Final report. (OPRE Report No. 2012-45). Washington, DC: Office of Planning, Research and Evaluation, Administration for Children and Families, Department of Health and Human Services.

Saez, E., Slemrod, J., & Giertz, S. (2012). The elasticity of taxable income with respect to marginal tax rates: A critical review. Journal of Economic Literature, 50, 3–50. https://doi.org/10.1257/jel.50.1.3

Thompson, O. (2018). Head Start’s long-run impact: Evidence from the program’s introduction. Journal of Human Resources, 53. https://doi.org/10.3368/jhr.53.4.0216-7735R1