From their infancy to their adolescence, children continuously develop capacities that are crucial for their physical and emotional well-being and their cognitive abilities, which in turn help to promote their success in school, their responsible behavior as adults, their eventual economic self-sufficiency, and lifelong health. These capacities, therefore, are the foundation of a well-functioning and prosperous society. Numerous studies suggest that a lack of adequate resources in childhood compromises the development of these capacities. Accordingly, the widespread poverty among American children today is cause for serious concern.
Using the Supplemental Poverty Measure (SPM) threshold of about $25,000 for a family of four, in 2017 the U.S. Census Bureau counted more than 11 million U.S. children—nearly one-sixth of all our children—living in families with incomes that fell short of that poverty line (Fox, 2018).1 It also determined that 3.5 million of those children were living in “deep poverty,” defined as having family resources less than one-half the SPM poverty line (Fox, 2018). As detailed in Chapter 2 of this report, child poverty rates are much higher for Black, Hispanic, and American Indian children than for White or Asian children. They are also much higher for children in single-parent families than those two-parent families
1 See Tables A-2 and A-4 in The Supplemental Poverty Measure: 2017 at https://www.census.gov/content/dam/Census/library/publications/2018/demo/p60-265.pdf. The number of children in deep poverty was calculated by multiplying the percentage of people under age 18 with family incomes below 50 percent of the SPM poverty threshold by the number of children under age 18 in the United States (estimated to be 73.7 million in 2017 by the U.S. Census Bureau).
and for children in families with no workers than for those in families with part- or full-time workers. By most measures, poverty among U.S. children is higher than in peer English-speaking countries such as Canada and Australia, and it is much higher than in most other industrialized countries.
A robust research literature (reviewed in Chapter 3) shows that children growing up in poverty fare much worse than other children. Differences favoring children in more affluent families are already evident in toddlers’ and preschoolers’ language, memory, self-regulation, and socioemotional skills, with corresponding differences observed in neural structure and function in the brain regions that support these skills. Children living in deep poverty have the worst outcomes among all children on important health and development indicators, such as blood lead levels, obesity, and a composite indicator of flourishing that measures children’s mood, affection, and resilience (Ekono, Jiang, and Smith, 2016). By the time they reached their 30s, individuals whose families had incomes below the poverty line during early childhood completed two fewer years of schooling and were earning less than one-half as much income, on average, when compared with peers whose family incomes were at least twice the poverty line (Duncan, Ziol-Guest, and Kalil, 2010). Not all these differences can be attributed to poverty per se. Nevertheless, our review of the literature on the causal effects of childhood poverty (see Chapter 3) shows that the weight of the evidence indicates that income poverty itself causes negative child outcomes. This is especially the case when poverty begins in early childhood and/or persists throughout a large share of a child’s life.
Whether a family’s income is above or below a poverty threshold depends on parents’ decisions regarding their own schooling, work, and marriage, as well on a host of structural factors such as the availability of work, housing, and public transportation, the prevalence of neighborhood crime, and institutional racism, all of which are well beyond the control of families. However, government programs also matter a great deal. Child poverty rates in the United States would be much higher were it not for programs such as the Supplemental Nutrition Assistance Program (SNAP), which provides nutrition assistance benefits to low-income individuals, the Earned Income Tax Credit (EITC), and the Child Tax Credit (CTC) (see Chapter 4).
If all countries’ child poverty rates were measured solely by the earned income of parents, U.S. children would have poverty rates that fell in the middle of the rankings among peer English-speaking countries. Part of what drives our child poverty rates so much higher than those in peer Anglophone and other high-income nations is the much smaller fraction of U.S. Gross Domestic Product that is devoted to redistributive social programs (see Chapter 4). According to Kids’ Share 2018 (Isaacs et al.,
2018), spending on children younger than age 19 accounted for 9 percent of the U.S. federal budget in 2017. This figure, which does not include state spending on education, is projected to fall to 6.9 percent by 2028, while at the same time spending on adults under Social Security, Medicare, and Medicaid, which accounted for 45 percent of the budget in 2017, is projected to rise to 50 percent by 2028.
Given the problems generated by child poverty in the United States and the demonstrated effectiveness of many child poverty programs, the omnibus appropriations bill signed into law in December 2015 included a provision directing the National Academies of Sciences, Engineering, and Medicine to conduct a comprehensive study of child poverty in the United States. Specifically, the study was to provide an evidence-based, nonpartisan analysis of the macroeconomic, health, and crime/social costs of child poverty, to study current efforts aimed at reducing poverty, and to propose recommendations with the goal of reducing the number of children living in poverty in the United States by one-half in 10 years.2 This policy goal mirrors the aims of anti-poverty initiatives that have been undertaken in other English-speaking countries in the past two decades, most notably in the United Kingdom beginning in 1997 (Waldfogel, 2010; see also Chapter 4).
The heart of the charge issued by the U.S. Congress to the National Academies is the goal of reducing the number of children living in poverty in the United States by one-half within 10 years. Congress has requested objective analyses of the existing research on the poverty-reducing effects of major assistance programs directed at children and families and specific policy and program recommendations for accomplishing this goal.
Ad hoc committees appointed by the National Academies are guided by a statement of task that defines and constrains their work.3 Committee reports are expected to address all of the issues raised in the statements of task but not to go beyond them unless the committee judges it absolutely necessary for carrying out the full scope of the statement of task. The statement of task for the present study is shown in Box 1-1.
In developing its list of policy and program proposals for reducing child poverty by half in 10 years, the committee considered existing federal programs as well as innovative programs developed by states and localities
2 See Consolidated Appropriations Act, 2016, Pub.L. No. 114-113.
3 This study’s statement of task was developed jointly by staff members from Congress, the Office of the Assistant Secretary for Planning and Evaluation within the Department of Health and Human Services, and the National Academies, as well as researchers and policy makers with expertise in the reduction of child poverty.
and in other countries, such as the United Kingdom and Canada. The scope of the programs the committee considered was broad. In addition to traditional anti-poverty programs, such as cash transfers, food and nutrition programs, and housing programs, the committee considered work support, health insurance, foster youth, juvenile and adult justice, and education and training programs.
For each program and policy option it developed, the committee attempted to estimate what impact it could have on reducing child poverty as defined using the SPM; how its poverty-reducing impacts would be distributed across demographic groups and across groups at three different levels of poverty: those at the poverty level; those in deep poverty; and
those in near poverty; and what would be the annual cost of implementing the program or policy at scale. To the extent possible, the committee examined the sensitivity of the impacts of its policy or program proposals to economic conditions, and it also considered other possible benefits the proposals could provide for government and society, such as improvements in child health, educational achievement, and welfare.
Because virtually none of the program and policy options we developed, if considered individually, would meet the 50 percent poverty-reduction goal, we also considered packages that would combine a number of policy and program changes to meet that goal. These are presented in Chapter 6. The task of designing these packages led us to identify interactions among programs that could result in synergies or redundancies.
Timing is a key element of the committee’s statement of task. The policies and programs identified by the committee are intended to “help reduce child poverty and deep poverty . . . by 50 percent within 10 years of the implementation of the policy approach.” This relatively brief, decade-long interval focuses attention on actions that aim to quickly increase the resources available to the families of poor children—programs and policies such as tax credits or work requirements. Although programs such as those that support early childhood education may boost family income by enabling a mother to work, their main goals are to reduce poverty among future—rather than current—generations of children. Accordingly, they fall outside the committee’s statement of task, although they may be important to reducing poverty over the longer term.
Reducing Poverty or Building Children’s Capacities and Health?
The concern that growing up in poverty compromises children’s opportunities to develop to their full potential provides a powerful motivation for seeking to reduce or even eliminate child poverty. However, with children’s development in mind, the goal of child poverty reduction alone, whether in the short or long term, is limiting because it focuses all our attention on family resources and ignores other important factors in healthy development. An alternative goal to poverty reduction might be to promote children’s human capital, conceived broadly to include cognitive and noncognitive capacities as well as physical and mental health, both during childhood and into the adult years. Poverty reduction will help to build children’s human capital, but so too will attention to a much broader range of factors that promote children’s health and development, both
within the family and in the schools, neighborhoods, and other contexts of children’s lives.
For example, a broader goal of human capital development might lead us to favor policies and programs to promote more nurturing homes or more effective school environments over equally costly programs and policies that would benefit children only by improving their material circumstances. This report responds to the committee’s short-term poverty-focused congressional charge, but readers should bear in mind that adequate family material resources are but a single, albeit important, input for the healthy long-term development of children.
That said, programs targeting child poverty can build human capital in other ways. As an example, consider food assistance programs. Child poverty, as measured by the SPM, falls when benefits from a program like SNAP boost family resources. But, as explained in Chapter 3, the evidence also indicates that SNAP’s predecessor program, food stamps, reduced the incidence of low birth weight among children born into low-income families and, if benefits were received during early childhood, improved that child’s cardiovascular health in adulthood as well. When making decisions, policy makers might want to consider these kinds of human capital impacts along with the reductions in shorter-term child poverty that a specific program or policy might achieve. With that in mind, the committee’s review of the poverty literature in Chapter 3 includes evidence on programs that both reduce child poverty and promote children’s health development.
The heart of the committee’s charge is to “identify policies and programs with the potential to help reduce child poverty and deep poverty . . . by 50 percent within 10 years.” To identify these programs and policies, the committee sought suggestions from its members and invited outside testimony from experts in the field. These included experts from universities, from policy organizations, and from practitioner organizations and represented a diverse array of political perspectives. In addition to holding two public information-gathering sessions, the committee received 25 policy memos, 19 of them from the 40 individuals we invited to submit memos and 6 more that were unsolicited. The committee also drew on the expertise of its own members to develop a list of possible policies and programs that might meet the charge. In addition, the committee commissioned papers from experts in Medicaid and American Indian and Alaska Native (AIAN) children living in poverty.4
Criteria for Selecting Programs and Policies
With hundreds of local, state, federal, and international anti-poverty program and policy models to choose from, the committee developed a set of criteria to guide its selection and then considered the strengths and weaknesses of each policy or program. The criteria are as follows:
- Strength of the research and evaluation evidence
- Magnitude of the reduction in child poverty
- Child poverty reduction within high-risk subgroups
- Cost of the program or policy
- Impacts on the widely held values of work, marriage, opportunity, and social inclusion
The most important criterion was the strength of the research and evaluation evidence indicating that, if enacted, the policy would reduce child poverty in the short run. Here the committee gave preference to evidence from random-assignment program evaluations as well as methodologically strong “natural experiments,” that is, those that examined the impacts of unanticipated changes in the timing and structure of policies on children and their families. To generate estimates of poverty reduction from the committee’s program and policy ideas, it commissioned research from the Urban Institute’s Transfer Income Model, Version 3 (TRIM3) microsimulation model.5
Second, with a target of reducing child poverty by one-half within 10 years, an obvious guiding criterion was the magnitude of the reduction in overall child poverty. The committee’s statement of task speaks of reductions in both the number of poor children and the fraction of children whose family incomes are below the poverty line. Since these two indicators may differ slightly in the context of a growing population of children, the committee chose to focus on reductions in the rate of child poverty.
Discussions with study sponsors led the committee to consider the distribution of poverty-reducing impacts across high-risk groups of children, defined by such characteristics as race, location, immigration status, and age of parent, who have above-average levels of poverty. Accordingly, the committee assigned importance to anti-poverty programs with relatively larger impacts on the children in these groups.
The fourth criterion was the likely cost of the program or policy. We defined cost as the incremental budgetary expense after accounting for all of the secondary impacts of the program or policy change such as
participation in other programs and changes in taxes paid resulting from changes in employment (for example, payroll taxes).
Fifth, the committee considered whether the program or policy was likely to promote widely agreed-upon values. Although not an explicit element of the statement of task, societal values have always figured prominently in debates over the nature of anti-poverty programs in the United States (Lamont and Small, 2008). We focus on four such values: work, marriage, opportunity, and social inclusion. None is without complications or qualifications. In the case of work, for example, expectations that program participants seek paid employment may be suspended in the case of a parent with an infant or a severely disabled child. In the case of marriage, relationship quality is also a criterion, so an abusive or violent relationship, for example, would not be valued. Considerations of social inclusion figure prominently in debates over whether programs should be offered universally rather than targeted to the neediest individuals (Garfinkel, Smeeding, and Rainwater, 2010). Universal programs are obviously more costly, but targeted programs can generate unforeseen incentives for people to qualify for or remain in programs, and recipients of targeted programs can run the risk of being stigmatized and confined to separate programs for the poor. In some cases, targeted programs that reward work, like the EITC, appear to generate a strong sense of social inclusion among recipients (Halpern-Meekin et al., 2015).
In keeping with the spirit of its charge, the committee omitted political feasibility from its list of criteria, although we understood that some policies and programs might be more politically feasible than others. As the charge from Congress directs, the committee endeavored to “provide an evidence-based, nonpartisan analysis.”
The committee did not insist that all of the anti-poverty programs and policies it identified meet all of its five criteria. Strong research evidence was vital, but at the same time the committee recognized the inevitable tradeoffs in any policy or program proposal. Some of the approaches it chose were stronger on some criteria and weaker on others. The committee sought to balance the strengths and weaknesses of each proposal in light of the criteria taken as a whole.
At first glance, estimating poverty reductions for any given program may appear to be a straightforward calculation. If Program A provides, say, $5 billion in additional benefits to families with children, why not just conduct a simulation in which the incomes of recipient families are increased by the value of the added benefits and then determine how many families
are raised above the poverty or deep poverty thresholds by the incremental income? A first complication is that in the course of reducing poverty, anti-poverty policies and programs can produce behavioral responses on the part of parents. For example, programs like the EITC boost the (after-tax) hourly earnings of some low-wage workers, which can induce them to work and earn more, and this would then increase the poverty-reducing impact of the EITC well beyond what is accomplished by the tax credits alone (Hoynes and Patel, 2017). Other programs can discourage work by reducing program benefits when earnings increase, or may discourage marriage by imposing rules that provide fewer benefits to married parents than to single parents. These kinds of behavioral responses are difficult to gauge but, as explained in Chapter 5, the committee, supported by the research literature, attempted to incorporate such responses in its estimates of child poverty reductions.
A second complication in some programs is that not every potential recipient will in fact take up the benefit. Housing vouchers are an obvious example, because a substantial number of families offered vouchers today are not able to use them. As explained in Appendix F, the TRIM3 microsimulation model the committee used attempts to incorporate adjustments for behavioral responses and incomplete program take-up.
In some cases, the committee concluded that while a program met its criteria, it was not amenable to a quantitative policy simulation. One example is a program to promote the use of long-acting reversible contraception (LARC) devices, which have the potential to reduce poverty by delaying or reducing births into poor families; however, evidence on program take-up and impacts is fragmentary (See Chapter 7 for more information). Indeed, a number of promising small-scale demonstration programs have never been scaled up sufficiently to show whether key program features could be preserved if they were to be implemented across the nation or even across a single state. Expansions of the Medicaid medical insurance program are another example. The committee’s literature review in Chapter 3 suggests that health insurance programs can improve child health, but estimating short-run impacts of program expansion on poverty reduction is complicated by the various ways poverty measures handle health care benefits and expenditures.
Therefore, Chapter 5 includes programs and policies for which evidence on behavioral responses, take-up, and other complicating issues is definitive enough to support a reasonably precise set of estimates of child poverty reduction. In Chapter 6, the committee anticipated that programs and policies interact and so they estimated synergies and redundancies across programs and policies in its examination of packages. Chapter 7 discusses programs for which the evidence base was sufficient to suggest
considerable promise but not strong enough to support precise estimates of national impacts on child poverty.
Poverty reduction may benefit children in some families more than others. Parents coping with the stresses of unstable work schedules, personal or family illnesses and disabilities, uninvolved partners, neighborhood crime, low-quality schools, or discriminatory workplaces may find it difficult to engage in responsive parenting or longer-run planning on behalf of their children (McLoyd, 1998; Mullainathan and Shafir, 2013). These problems, in turn, may dilute some of the possible benefits of policy-induced increases in material resources. Because these contextual considerations are so important, and most are not part of the simulation model, the committee devotes an entire chapter (Chapter 8) to them and their implications for the committee’s conclusions.
Finally, the expertise of committee members spans a wide range of disciplines and includes policy work in state and federal governments as well as in the nonprofit sector. All members share a commitment to the standards of evidence embraced by the National Academies but at the same time brought diverse political orientations to issues surrounding anti-poverty policies. For the programs featured in Chapters 5 and 6, it is important to understand that committee consensus on their inclusion was based solely on the strength of the evidence base supporting them and not on individual committee members’ endorsements of the policies themselves.
The report proper begins in Chapter 2 with a demographic portrait of child poverty in the United States. In this chapter we explain how poverty is measured and why the relatively new SPM, which our statement of task directs us to use, provides a somewhat different view of child poverty than the much older official measure. Child poverty rates are lower with the SPM than with the cash-based Official Poverty Measure (OPM). Over the past half-century, SPM-based child poverty has declined more rapidly than OPM-based poverty. In Chapter 2, we also compare child poverty in the United States and in peer anglophone countries. By and large, the United States has considerably higher rates of child poverty than these other countries, although the concentration of poor children among single-parent and nonworking families is broadly similar.
In Chapter 3, we respond to the first element of the statement of task by reviewing the literature on the consequences of child poverty, including macroeconomic, microeconomic, health, and social costs. The chapter explains how the technical sophistication of these literatures has increased markedly over time, as studies of the consequences of child poverty have progressed from an emphasis on correlational methods to the use of natural
experiments that track how measures of child well-being change in response to large changes in policies such as the EITC and SNAP.
Chapter 4 responds to the second element of the statement of task by providing an assessment of current local, state, federal, and international efforts to reduce child poverty. As directed by the statement of task, the committee provides a separate look at poverty lines drawn to distinguish deep poverty (defined as below 50% of the SPM poverty line), conventional poverty (as defined by the SPM), and near poverty (the upper limit of which is defined as 150% of SPM poverty). At the federal level, a noteworthy distinction can be made between program impacts on the poverty of children whose families are near the poverty threshold and impacts on children in families well below the threshold. Tax-based programs such as the EITC move millions of children above the SPM-based poverty line but have much smaller impacts on the economic status of children in families with little taxable income. On the other hand, income-tested programs such as SNAP proved most effective at increasing the economic resources of the families of children in deep poverty.
Peer English-speaking countries provide some interesting examples of efforts to reduce child poverty, most notably the United Kingdom, where the government pledged in 1999 to halve child poverty within a decade and to eradicate it completely within two decades (Waldfogel, 2010). More recently, Canada enacted a very substantial child benefit for low-income families that is estimated to have reduced poverty among Canadian children by 5 to 6 percent within a year of its 2016 enactment (Sherman, 2018). These efforts are also reviewed in Chapter 4.
A crucial element in the committee’s charge is to compose a list of promising anti-poverty policies and programs. As discussed above, we did so by drawing on the evaluation research literature as well as on ideas from individuals and groups representing a broad range of political orientations and experiences working in local and county governments, at the local social services and school systems level, and in state and federal government. Chapter 5 details the policy and program proposals that were amenable to a quantitative policy simulation to estimate net impacts. The summary section of Chapter 5 covers several issues that cut across the set of program and policy proposals the committee developed. Several are based on how the various proposals rank based on the selection criteria, for example, ranking proposals based on cost, degree of poverty reduction both overall and in key demographic subgroups, and impacts on employment.
In Chapter 6, the committee presents program packages that are projected to meet the 50 percent poverty-reduction goal set by its authorizing legislation. Chapter 7 describes additional programs and policies that were judged to be promising but for one reason or another were not amenable to precise estimates of impact on child poverty.
The focus of Chapter 8 is on contextual factors that affect child poverty—from program administration to discriminatory behaviors and criminal justice policies and practices. These factors are not typically incorporated in the simulation models, but they can have a profound effect on the success of programs, providing useful infrastructure in some cases and interfering with policy, thereby creating “leaky buckets,” in others.
The final chapter (Chapter 9) summarizes the committee’s recommendations and outlines a research agenda. Chapter 9 also discusses the importance of implementing high-quality monitoring and evaluation to measure progress and identify further steps.
Appendix A includes biosketches of committee members and project staff and Appendix B provides the agenda for the two public information-gathering sessions. Appendix C lists the individuals and organizations that submitted memos to the committee. Appendix D comprises the appendixes for Chapters 2, 3, 4, and 5. Appendix E includes the TRIM3 summary tables, and Appendix F contains the Urban Institute TRIM3 technical specifications.
Finally, a note on the overall organization of this report: As with all consensus reports produced by the National Academies, we provide evidence supporting all of our conclusions and recommendations. But in contrast to many of those reports, here the bulk of this evidence is presented in online appendixes associated with most of the chapters. Separating the detailed evidence in this way enabled us to write a shorter and, we hope, more accessible presentation of our analyses and conclusions. The online appendixes (D through F) are available on the National Academies Press webpage at http://www.nap.edu/25246 under the Resources tab.
Duncan, G.J., Ziol-Guest, K.M., and Kalil, A. (2010). Early childhood poverty and adult attainment, behavior, and health. Child Development, 81(1), 306–325.
Ekono, M.M., Jiang, Y., and Smith, S. (2016). Young Children in Deep Poverty. Columbia University Academic Commons. Available: https://doi.org/10.7916/D86Q1X3.
Fox, L. (2018). The Supplemental Poverty Measure: 2017. Current Population Reports, P60265. Washington, DC: U.S. Census Bureau.
Garfinkel, I., Smeeding, T., and Rainwater, L. (2010). Wealth and Welfare States: Is America a Laggard or Leader? New York: Oxford University Press.
Halpern-Meekin, S., Edin, K., Tach, L., and Sykes, J. (2015). It’s Not Like I’m Poor: How Working Families Make Ends Meet in a Post-Welfare World. Berkeley: University of California Press.
Hoynes, H.W., and Patel, A.J. (2017). Effective policy for reducing poverty and inequality? The Earned Income Tax Credit and the distribution of income. Journal of Human Resources. doi: 10.3368/jhr.53.4.1115.7494R1.
Isaacs, J.B., Lou, C., Hahn, H., Hong, A., Quakenbush, C., and Steuerle, C.E. (2018). Kids’ Share 2018: Report on Federal Expenditures on Children Through 2017 and Future Projections. Washington, DC: Urban Institute.
Lamont, M., and Small, M.L. (2008). How culture matters: Enriching our understandings of poverty. In A. Chih Lin and D.R. Harris (Eds.), The Colors of Poverty: Why Racial and Ethnic Disparities Persist (pp. 76–102). New York: Russell Sage Foundation.
McLoyd, V.C. (1998). Socioeconomic disadvantage and child development. American Psychologist, 53(2), 185.
Mullainathan, S., and Shafir, E. (2013). Scarcity: Why Having Too Little Means So Much. New York: Macmillan.
Sherman, A. (2018). Canadian-style Child Benefit Would Cut U.S. Child Poverty by More Than Half. Washington, DC: Center for Budget and Policy Priorities. Available: https://www.cbpp.org/blog/canadian-style-child-benefit-would-cut-us-child-poverty-by-more-than-half.
Waldfogel, J. (2010). Britain’s War on Poverty. New York: Russell Sage Foundation.