One Percent for the Kids: New Policies, Brighter Futures for America's Children

Isabel Sawhill, Editor r Futures for America’s Chi e t h g ldre Bri , n s e i c i l o P w Ne

One Percent for the Kids

1269-00 FM 04/18/03 17:22 Page i

One Percent for the Kids

1269-00 FM 04/18/03 17:22 Page ii

1269-00 FM 04/18/03 17:22 Page iii

One Percent for the Kids New Policies, Brighter Futures for America’s Children

I S A B E L V. S AW H I L L Editor

BROOKINGS INSTITUTION PRESS Washington, D.C.

1269-00 FM 04/18/03 17:22 Page iv

Copyright © 2003 THE BROOKINGS INSTITUTION

1775 Massachusetts Avenue, N.W., Washington, D.C. 20036 www.brookings.edu All rights reserved

Library of Congress Cataloging-in-Publication data One percent for the kids : new policies, brighter futures for America’s children / Isabel V. Sawhill, editor. p. cm. Includes bibliographical references and index. ISBN 0-8157-7722-1 (cloth : alk. paper) ISBN 0-8157-7721-3 (pbk. : alk. paper) 1. Child welfare—United States. 2. Children—Services for—United States. 3. Children—United States—Social conditions. I. Sawhill, Isabel V. HV741.O635 2003 362.71'0973—dc21 2003006185

987654321 The paper used in this publication meets minimum requirements of the American National Standard for Information Sciences—Permanence of Paper for Printed Library Materials: ANSI Z39.48-1992. Typeset in Minion Composition by Circle Graphics Columbia, Maryland Printed by R. R. Donnelley Harrisonburg, Virginia

1269-00 FM 04/18/03 17:22 Page v

Contents

F O R EWO R D

vii ix

ACKNOWLEDGMENTS

1 2

Introduction Isabel V. Sawhill

1

Promoting the Healthy Development of Young Children Greg J. Duncan and Katherine Magnuson

3

When Work Alone Is Not Enough Robert Haveman 40

4

Preventing Early Childbearing Andrea Kane and Isabel V. Sawhill

16

56

5

Strengthening Fragile Families Irwin Garfinkel and Sara McLanahan 76

6

Family Policy and Academic Achievement David J. Armor 93 v

1269-00 FM 04/18/03 17:22 Page vi



vi

7

Providing Universal Preschool for Four-Year-Olds Barbara Wolfe and Scott Scrivner 113

8

Improving Neighborhoods for Poor Children Jens Ludwig 136

9

The Blair Government and Child Poverty: An Extra One Percent for Children in the United Kingdom John Hills 156

CONTRIBUTORS

179

B R O O K I N G S R O U N D TA B L E O N C H I L D R E N INDEX

185

182

1269-00 FM 04/18/03 17:22 Page vii

Foreword

F

ive years ago, Isabel Sawhill, a senior fellow in our Economic Studies department (known to all of us here as Belle), initiated a project called the Brookings Roundtable on Children. It is a national network of scholars that has met twice a year to discuss how best to improve children’s life prospects. This book is the culmination of that effort, and it is a testament to Belle’s intellectual leadership, organizational skill, and commitment to what is surely one of the key issues for any society: the development of policies that will improve the future of children. The Roundtable’s goals have been to understand the factors that affect the social mobility of children in the United States, to determine what already is known about intervening effectively in children’s lives, and to evaluate the choices available to policymakers. As part of its work, the Roundtable has published a number of research papers and Policy Briefs (a regular Brookings product) to educate public officials, scholars, and interested citizens on issues affecting the welfare of children. As one example, a discussion paper done with the support of the Roundtable led to the enactment in 2001 of a partially refundable children’s tax credit. In this volume, each author was asked to reflect on what the research showed about promising or proven strategies for improvvii

1269-00 FM 04/18/03 17:22 Page viii

viii



ing children’s futures. Most of the chapters were originally presented and reviewed at a Roundtable meeting held in May of 2002. Several additional chapters were commissioned over the summer, and the authors met again in October of 2002 to discuss both the new chapters and the overall themes of the book. Although no votes were taken, the policy recommendations contained in the book received widespread support from the group as a whole. The authors’ collaboration during both the writing of the volume and the five years of meetings that preceded it has produced a comprehensive, actionable plan for new social and economic policies designed to assist disadvantaged children and their families. The first chapter spells out the authors’ priorities and their budgetary implications. Taken together, the recommended policies constitute an ambitious agenda that also is very expensive—although the cost is less than 1 percent of GDP. The title of the book and the budgetary limit the authors set for themselves derive from the fact that the Blair government in Britain has devoted an extra 1 percent of GDP to its own goal of eliminating child poverty over a twenty-year period. The United States is unlikely to make a similar commitment any time soon, but part of the Brookings mission is to put ideas like these on the national agenda, since that is the best way to encourage a broader and more informed debate that eventually might move national policy in a new direction. I join Belle and the authors in hoping that this book will prove to be a useful tool for policymakers charged with evaluating potential strategies for assisting low-income children and families—and for anyone who is looking for new ways to improve children’s futures. S TROBE TALBOTT President, Brookings Institution Washington, D.C. March 2003

1269-00 FM 04/18/03 17:22 Page ix

Acknowledgments

B

rookings is grateful to the Atlantic Philanthropies, the Foundation for Child Development, the Alcoa Foundation, and Dorothy Perlow for their financial support of the Brookings Roundtable on Children. The editor wishes to thank both the authors and the other members of the Roundtable: Henry Aaron, George Akerlof, Douglas Besharov, Jeanne Brooks-Gunn, Gary Burtless, Janet Currie, William Dickens, David Ellwood, Ronald Haskins, Christopher Jencks, Joyce Ladner, Susan Mayer, Kristin Moore, Deborah Phillips, and Lisbeth Schorr. Although these members did not contribute a chapter (and do not necessarily endorse the views represented in the volume) they made invaluable contributions to the Roundtable in the form of discussion and debate over the past four years. In addition, several of them contributed discussion papers or briefs to the project (see page 182 for a list of Roundtable publications). Finally, special thanks are due to Molly Fifer, Carolyn Malcom, and Eileen Hughes for their assistance with the final editing and production of the volume.

ix

1269-00 FM 04/18/03 17:22 Page x

1269-00 FM 04/18/03 17:22 Page xi

One Percent for the Kids

1269-00 FM 04/18/03 17:22 Page xii

1269-01 CH01 04/18/03 17:23 Page 1

1 Introduction I S A B E L V. S AW H I L L

A

lmost everyone talks about the need to invest in the next generation. Yet in the late 1990s federal spending on children represented only 2 percent of GDP, in contrast to the 7 percent of GDP spent on Social Security and Medicare, the 3 percent spent on defense, and the 2.5 percent spent on interest to service the national debt.1 When state and local spending on education is included, the proportion spent on children increases by 4 percentage points, almost equaling the amount spent on Social Security and Medicare.2 However, virtually all observers agree that the growth of spending on these big entitlements for the elderly— together with rising expenditures for the war on terrorism and the increased interest payments associated with higher projected budget deficits—are likely to crowd out spending on children’s programs over the next several decades. The issue is not just how much the nation is investing in the next generation but also whether it is investing the funds well. This book examines promising initiatives—outside the K-12 public 1. Clark and others (2001); Congressional Budget Office (2000). 2. U.S. Census Bureau (2001).

1

1269-01 CH01 04/18/03 17:23 Page 2

2

 . 

school arena—that research suggests offer effective strategies for improving children’s futures. The authors have chosen to focus on nonschool initiatives not because they believe that education is unimportant, but because their expertise lies in other areas. The goal of all the initiatives discussed in this volume is to improve children’s life chances. Despite deeply held beliefs in the United States about the importance of opportunity, children born into different kinds of families begin life with very unequal prospects. By the time they are mature enough to make their own decisions, the die is largely cast. Moreover, their prospects are becoming increasingly unequal, both because parental income gaps have widened and because of a related bifurcation in family structure. The result is that a growing group of children is being raised in families in which the mother begins childbearing at an early age—typically outside of marriage— completes very little education, and ends up poor and often dependent on public assistance. Another growing group of children has parents who delay marriage and childbearing until they have completed a high level of education, established themselves in a career, and married. These parents have a limited number of children at an older age and invest both time and money in making sure that their children lead more advantageous lives.3 This bifurcation in children’s prospects implies increasing social divisions several decades hence that might be prevented if public and private investments were to improve the prospects of those children at the bottom of the socioeconomic ladder. This volume is the outgrowth of a four-year-long discussion among members of the Brookings Roundtable on Children, a group of scholars convened to address ways of improving children’s life prospects. The authors document what they believe the best research suggests about new policy directions in each of the following domains: income support for families, family formation and parenting, health care, early education and care, and neighborhood environments. Not all of the ideas discussed have been adequately tested using rigorously evaluated demonstrations, but all seem promising based on the available research. For this reason we commend them to policymakers, foundations, and local communities for further consideration. Where insufficient evidence on effectiveness exists, the authors suggest ways of plugging the gaps. No exercise of this sort would have much credibility if it failed to estimate costs and establish funding priorities. The authors have handled this issue by 3. Data on the size of these groups and on the growing bifurcation can be found in Sawhill and Chadwick (1999); also see Ellwood and Jencks (2002) and Sawhill (1999).

1269-01 CH01 04/18/03 17:23 Page 3



3

taking a cue from the Blair government in the United Kingdom. As detailed by John Hills in chapter 9, since coming to office in 1997 the Blair government has devoted almost an extra 1 percent of the country’s GDP to reducing child poverty. Translated into U.S. terms, this represents almost $100 billion a year in added spending. Our final budget of $60 to $76 billion a year is somewhat less than that and also less than the portion of the 2001 tax cut allocated to the wealthiest 5 percent of American families ($88 billion a year when fully phased in). Still, it is a startling sum—especially in the context of the currently cramped domestic agenda in the United States. The Children’s Roundtable authors have no illusion that the nation will anytime soon devote resources of this magnitude to improving children’s futures. They nonetheless feel that it is useful to demonstrate what could be accomplished if the wealthiest Americans were to forgo a tax cut and the nation were to invest a major portion of the forgone funds in the next generation. If action proceeds from ideas, putting these policy options on the agenda for further discussion and debate is the first step toward changing existing budgetary and political constraints. Table 1-1 shows a rough allocation of our budget among the different policy proposals that compose the core of this book. Needless to say, we did not immediately agree on the best way to improve the lifetime prospects of children or how to divide up these dollars. Still a consensus emerged, and in this first chapter I want to reflect on those areas of agreement after highlighting some of the interesting debates that surfaced in the process of arriving at an agreement. One of the debates was about whether to target younger or older children for investment. Almost all of us wanted to focus on younger children, especially those in the preschool years. Research has shown that, as a general proposition, investments in young children produce bigger payoffs than later interventions.4 One exception is found in chapter 4, which focuses on adolescents. It does so in part because of emerging evidence that some interventions with this age group can reduce crime, drug abuse, and dropping out of school. In addition, such interventions can reduce early childbearing outside of marriage and all of the problems that it creates for the children born to very young single parents. In short, starting early means starting before, not after, a child is born and targeting those who have not yet had children. Another debate concerned whether to provide universal services to all children (or their families) or to direct limited funds to a smaller and more disadvantaged group. Most members of the group wanted to limit additional assistance to the most needy, but many cautioned that it was important to 4. Currie (2000); Heckman (2000).

1269-01 CH01 04/18/03 17:23 Page 4

 . 

4

Table 1-1. Estimated Cost of Improving Children’s Futures Billions of dollars

Program description Child allowancea Elimination of child credit for under 6 Increased federal income tax revenues Decreased food stamp outlays due to increased income Reduction of child care tax credit Decreased TANF outlays due to increased income Total savings Increased earnings supplementsb Parental leavec After-school programsd Marriage promotion demonstratione National Whole Family Campaignf Improved health services Universal prenatal/perinatal screeningg Insurance for all uninsured under 18 Intensive interventions for severe behavioral and emotional problemsg Early childhood educationh Universal preschool for 4-year-oldsi Reallocation of current child care subsidies for 4-year-olds Improved neighborhoods for poor childrenl Total cost

Gross cost

Savings

39.4

Net cost 10.6

16.8 4.6 3.4 1.7 2.3 28.8 20 0.3 2.5 0.15–0.25

20 0.3 2.5 0.15–0.25

0.4

0.4

1.7

1.7

10

10

5

5

10 20.8

10 0–15.4 j 5.4k

0

0 60.65–76.15

a. For families with children under five and income under $60, 000 (Duncan and Magnuson). b. Increase the minimum wage; accelerate the CTC phase-in; supplement the EITC for fulltime workers and cover any additional child care costs (Haveman).

1269-01 CH01 04/18/03 17:23 Page 5



5

extend assistance to the bottom third, or even half, of all families in order to avoid stigmatizing recipients or creating undue disincentives to work and to marry. In chapter 7, for example, Barbara Wolfe and Scott Scrivner admit that disadvantaged children are more likely to benefit from preschool education than their more advantaged peers, but the authors still argue for a universal program on the grounds that it will lead to less income stratification and more public support. And in chapter 2, Greg Duncan and Katherine Magnuson propose a child allowance extending well into the middle class but with a significant claw-back through the income tax system. A third debate centered around the importance of family environments and the ability of public policies to affect when people have children, how many children they have, whether they marry, and the quality of parents’ relationships with each other and with their children. Chapters 4, 5, and 6 deal with this set of issues. Almost everyone agrees that children are better off if they have two relatively mature, married parents who have completed their own education and are prepared to take on the responsibility of parenting. But how such an objective might be achieved is far less clear. Chapter 4 argues for reducing teenage pregnancy through new investments in after-school programs; chapter 5 for an experiment to strengthen the relationship between unwed parents; and chapter 6 for a campaign to educate parents and potential parents on how their decisions regarding marriage and childbearing affect their children’s chances of success. Still another unresolved issue is the importance of money in a child’s life. A lack of material resources is easier to replace through government action than a weak family. For this and other reasons, many policy options

c. Exemptions from all welfare-related work requirements until child is six months old and from full-time work requirements when child is between six months and one year of age; six months of parental leave (Duncan and Magnuson). d. Funds target low-income neighborhoods or school districts (Kane and Sawhill). e. Promote healthy marriages, fathers’ involvement, and child well-being among new unmarried parents (Garfinkel and McLanahan). f. Improve children’s intelligence and academic achievement by reducing family risk factors (Armor). g. Duncan and Magnuson. h. Intensive center-based programs for high-risk children up to three years old (Duncan and Magnuson). i. Wolfe and Scrivner. j. The higher figure assumes public funding; the lower figure assumes funding by parents. k. That is, 85 percent of the $6.4 billion currently being spent on four-year-olds through Head Start, CCDF, TANF, and tax credits (assuming an 85 percent participation rate among this age group). These offset funds could be used to improve the quality of care, rather than to reduce net costs. l. Provide housing vouchers to public housing residents; rebuild public housing stock (Ludwig).

1269-01 CH01 04/18/03 17:23 Page 6

 . 

6

designed to improve low-income children’s futures involve increasing their family’s income or access to needed resources. But the extent to which income matters, especially above some relatively low threshold amount, is not clear. Some scholars contend that existing safety-net programs have eliminated the most serious forms of material deprivation and that the remaining effect of income on children’s development observed in much of the literature results from the fact that income is correlated with such hardto-measure parental characteristics as poor mental health or low IQ. These factors rather than a lack of income per se cause some parents simultaneously to be less successful in the labor market and to adversely affect their children’s well-being.5 If that is the case, then providing more income to these families through a child allowance, for example, is not likely to have much effect on their children’s life chances. Other scholars believe that income plays a more causal role. They argue that it can affect the amount of stress experienced by parents as well as their ability to expose their children to various opportunities, such as good neighborhoods and good schools. These issues are addressed in chapters 2, 3, and 9, wherein the authors cite a variety of solid research done in both the United States and the United Kingdom that tends to support the “income matters” side of this debate. Nonetheless, the case is far from airtight. A final concern centered around whether any single programmatic intervention can succeed given the multiple needs of many children, especially the most disadvantaged, and the importance of making sure that programs are implemented well and staffed by competent people with a commitment to their mission and a willingness to learn from experience. Kristin Moore has written on the inability of single-purpose programs to address the multiple disadvantages found in a small number of very disadvantaged families, and Lisbeth Schorr has argued for greater attention to implementation issues, to the context in which programs operate, and to how interventions may combine and interact, the whole producing greater effects than the sum of its parts. Although their thinking is not represented in this volume, given their active participation in our discussions, it is worth noting here.6 With this as background, we considered the need for more resources for those with limited incomes and for stronger families, better health care, and more early education and child care as well as safer and less toxic neighborhoods. Clear support existed for a set of policies that would provide lowincome families with the resources they need to raise children. Our preference was for linking such assistance to a willingness to work. The biggest debate here 5. Mayer (1997). 6. Moore, Vandivere, and Redd (2003); Schorr (2003).

1269-01 CH01 04/18/03 17:23 Page 7



7

was over how much single parents should be expected to work, with some believing that they should be provided an allowance that would enable them to stay home at least while their children were very young and others arguing for conditioning new assistance not just on work but on full-time work. In chapter 3, Robert Haveman reports research showing that school-aged children in low-income families whose earned income is supplemented with government assistance have fared better in terms of academic achievement, behavior, and health. Other research, cited in chapter 2, suggests that infants whose mothers work extensively during the first year of life do not do as well as those whose mothers work fewer hours or not at all. In the end, we argue for work requirements for those receiving welfare with an exemption for the mothers of infants, and we recommend six months of paid parental leave for everyone, contingent on parents having established a work history. Families with incomes of less than $60,000 a year would receive a modest child allowance that is not contingent on work, and low-income working families would receive additional supplements. Supplementing the income of those raising children may be desirable, but all of the resources in the world will not save a child whose parents do not provide the nurturing, guidance, and discipline that researchers find children need. In chapter 6, David Armor suggests that poor home environments affect not only a child’s subsequent behavior but also his or her IQ or cognitive skills. He favors greater efforts to educate parents and parents-to-be about the effects of their own behaviors on their children’s intellectual functioning. Too many children are being born to young unwed parents, and we favor preventing such early births whenever possible (see chapter 4) and experimenting with efforts to make the parents’ relationship more successful and durable when it is not possible (see chapter 5). We have learned quite a bit about how to prevent early, unwanted pregnancies in recent years but less about how to produce durable relationships between men and women who already have had a child outside of marriage—or about how to produce good parenting among those who have never experienced it themselves. Health care was another high priority. The roundtable authors applauded recent expansions of health insurance through Medicaid and the State Child Health Insurance Program and the nutrition assistance provided through the Women, Infants, and Children program. These improvements need to be supplemented by adequate screening and a basic package of preventive health care for all children and pregnant women, including greater emphasis on the need to identify and treat behavioral and emotional problems early in life. Extending health insurance to all uninsured children under the age of eighteen would complete this agenda.

1269-01 CH01 04/18/03 17:23 Page 8

8

 . 

The authors believe that good-quality child care and early education are essential to children’s healthy growth and development. We favored not only establishing a universal pre-K program but also expanding Head Start or a similar early education program to cover all poor children at younger ages. We discussed opening the program to those somewhat above the poverty line along with extending the hours (or providing wraparound child care) for mothers who work. For nonpoor mothers, we saw a need for adequate and stable child care provided in part by the government and in part by parents’ contributions linked to their ability to pay. Chapter 7 details the many sources of current funding for such care. The child allowance proposed in chapter 2 would give parents more choice; they could use the funds either to pay for care or to support a mother’s decision to stay at home. Last, with respect to neighborhoods, we were impressed with research showing that children whose families move to low-poverty neighborhoods fare better in terms of academic achievement, health, and involvement in crime. We suspect that the more positive peer and adult influences in higherincome neighborhoods as well as better schools, more community-based activities, and better policing or adult supervision have something to do with the results. This suggests the need to invest in these basic neighborhood institutions and to avoid concentrating low-income families in public housing projects where peer influences are more negative. Better to provide such families with portable housing subsidies (so-called section 8 vouchers), which have the additional advantage of costing less than public housing. This policy menu is very broad and ambitious. Each chapter in this volume provides more details on what is known about each topic, how various authors would address the specific policy challenge in their area of expertise, what it might cost, and how more could be learned about the issues involved. A brief summary of each of the chapters follows. In chapter 2, Greg Duncan and Katherine Magnuson argue that early childhood is a crucial period in a child’s development and that the benefits associated with additional federal spending on young children easily exceed the costs of well-designed and well-targeted investments. They propose a package of policies for children in low-income families that contains the following components: universal prenatal and perinatal screening services plus health insurance coverage; intensive center-based early-education programs for high-risk children, beginning at age three; exemption from all welfare-related work requirements for mothers of children under six months of age and exemption from full-time (more than thirty hours a week) work requirements for mothers of children between six months and one year. Mothers with children

1269-01 CH01 04/18/03 17:23 Page 9



9

older than twelve months would not be affected. In addition, one working parent in each family would be entitled to six months of parental leave. To these components they add a taxable child allowance for families with children under the age of five and a family income below $60,000. The allowance would be equal to $300 a month during the child’s first year, $200 a month between ages one and five, and zero after that. Families with more than one child under age five would receive a maximum of two child allowances. An extensive body of early childhood research supports the argument that each of the components of Duncan and Magnuson’s proposal helps to ensure healthy child development. The medical care component of their proposal is supported by research identifying the negative consequences of compromising the quality of prenatal and perinatal care. The early education component is based on results from several evaluations of high-quality, center-based programs that have demonstrated improvements in a variety of outcomes among economically disadvantaged children enrolled in such programs. The work exemption and the child allowance components are based on research indicating that early maternal employment and low family income may be detrimental to young children. The cost of this proposal is high, making any uncertainty about its effectiveness especially critical. But unlike some advocates of such measures, the authors have done an especially careful job of providing a detailed assessment of the research literature along with the estimated costs of the proposed initiatives. In chapter 3, Robert Haveman addresses the need to increase the earnings of families raising children, many of whom have left welfare but are earning very low wages. He proposes some combination of a modest increase in the minimum wage, accelerating the child tax credit phase-in from the year 2010 to 2006, simplifying and expanding various tax credits affecting families with children, and supplementing the earned income tax credit for low-wage workers. These proposals are based on findings from demonstration programs conducted in Canada, Minnesota, and Milwaukee that have required or encouraged full-time work but also have supplemented the earnings of the families involved. Rigorous research shows that the programs have had beneficial effects for the school-age children in these families, improving their academic achievements in particular. In chapter 4, Andrea Kane and I address the issue of early unintended childbearing and its adverse consequences for children. We argue for an expanded investment in after-school programs that are less politically contentious than some of the alternative approaches to preventing teen pregnancy (for example, sex education, abstinence education, family planning) but potentially are as effective, especially for higher-risk youth. A major issue,

1269-01 CH01 04/18/03 17:23 Page 10

10

 . 

however, is their cost. The research literature suggests that the most effective programs can cost as much as $4,000 a year per participating teen, although some successful but less intensive efforts, such as the Teen Outreach Program, have cost considerably less. We propose a $2.5 billion investment in programs that replicate effective models. These models typically include a mix of academics, community service, and group activities; they are relatively long-lasting (at least a year), and they engage teens for a substantial number of hours, employ high-quality staff, and successfully recruit and retain high-risk youth. Funds would target low-income neighborhoods or school districts where young people have fewer opportunities to engage in constructive activities and where the risks of teen pregnancy are greatest. The benefits of this investment are likely to go beyond reducing teen pregnancy to include less crime, more academic achievement, and less drug use. Even when they are assessed in terms of their ability to accomplish only the single objective of ensuring that more teens delay childbearing, they can be a cost-effective investment for taxpayers. In chapter 5, Irwin Garfinkel and Sara McLanahan propose a demonstration that would evaluate new policies aimed at strengthening the relationship between unwed parents. Although the authors recognize the importance and desirability of marriage for children’s well-being and have made marriage-friendly policies an important part of their proposal, they also recognize that a substantial proportion of unwed parents choose to live apart. Therefore their proposal aims to promote more father involvement (including more child support) and better outcomes for children regardless of whether parents decide to marry. The demonstration consists of three components: providing better assessment and referral services to ensure that new parents are aware of the government-sponsored services for which they are eligible; reducing marriage penalties in existing programs; and providing education and mentoring programs to strengthen the communication skills and relationship of unwed parents. The proposals offered in this chapter are based, in part, on findings from the Fragile Families and Child Well-Being study, which shows that most new, unmarried parents are romantically involved and expect to marry each other at the time of their first child’s birth. However, few do marry, and we know very little about what, if anything, can be done to improve their relationships and increase their chances of marrying. The demonstration is designed to provide important guidance to policymakers by sorting out the relative effectiveness of different intervention strategies. In chapter 6, David Armor also focuses on family structure and related aspects of children’s early environments. His research has shown that chil-

1269-01 CH01 04/18/03 17:23 Page 11



11

dren’s IQs are most malleable when children are very young and that such measurable factors as family size, marital status of parents, and parents’ age, education, and parenting skills affect a child’s IQ. Early IQ, in turn, affects later IQ as well as educational attainment and income. Intervening once children have begun school is not as effective, in his view, as intervening very early; even preschool may be too late if we want to affect IQ rather than social skills and behavior. For these reasons—and because the most important influence in a young child’s life is his or her parents—he proposes “a whole family policy,” an effort to better educate parents and potential parents, starting in junior high school, about the best family environments for children. He suggests that if young people know how much their own decisions about schooling, childbearing, marriage, and parenting affect their children, they might be more motivated to finish school, marry before having a baby, limit the size of their families, and learn how to provide adequate instruction and emotional support to their children. His hope is to use education programs and media campaigns to change attitudes about the consequences of such decisions for children, much as the Surgeon General’s report changed attitudes about the consequences of smoking. In chapter 7, Barbara Wolfe and Scott Scrivner argue for a major new investment in preschool. Although they believe that the gains from such an investment would be especially large for children from more disadvantaged families, they support a universal program on the grounds that it would produce less social stratification and a greater willingness on the part of the public to pay for it. The authors propose a full day of preschool, with a part-day, education-intensive component taught by early education specialists and a child care component provided by aides supervised by teachers for the remainder of the day. In addition to describing program design and implementation, Wolfe and Scrivner’s chapter outlines a plan for financing universal preschool through a combination of federal and state funding and parent contributions. To support their proposal, the authors cite a large body of research showing that high-quality, highly structured preschool programs can improve children’s school readiness and academic achievement, reduce grade retention and use of special education, and reduce rates of crime and teen pregnancy at later ages. They cite the Chicago Child-Parent Program (CPC), an intervention operating in Chicago public schools in low-income areas, as the largest and potentially most convincing study. A recent CPC program evaluation found that participation in preschool at ages three and four is associated with higher school-completion rates by age twenty, lower arrest rates by age eighteen, and lower rates of special education and grade retention.

1269-01 CH01 04/18/03 17:23 Page 12

12

 . 

This evaluation provides a cost-benefit analysis as well, projecting a return to society of $47,759 per participant by age twenty-one, a benefit that compares favorably to the much lower cost of the program, at $7,000 per participant. To estimate the additional resources that might be needed to fund universal preschool, the authors first calculate how much currently is spent on early childhood education in the United States, including both public and private sources. Their proposed financing mechanism uses existing public funding sources (including the Child Care Development Fund, Head Start, and state child care funds), but it also relies on parents to subsidize a large portion of the costs. Parents would contribute between $125 and $800 a year over ten years, based on their ability to pay, which would be determined by the income taxes they have paid; parents with lower income would pay less than higher-income parents. To bolster their case, the authors cite investments in similar and even more extensive programs for preschool children in other developed nations, such as Belgium, France, Italy, and the Scandinavian countries. Discussing the political feasibility of enacting policies to support universal preschool, the authors note that both presidential candidates in the 2000 election emphasized increasing the federal commitment to improving early childhood education (Gore with a $50 billion proposed commitment to universal preschool and Bush with a commitment to expand Head Start and improve its quality). Wolfe and Scrivner suggest that although public opinion and political momentum may not yet support complete public financing of preschool, enacting their proposal (which includes substantial parent contributions) could eventually build support for public financing of preschool as an essential component of elementary education. In chapter 8, Jens Ludwig focuses on upgrading the quality of neighborhoods for poor children. Neighborhoods may matter for children’s life chances because interactions with other residents can change children’s attitudes and behaviors and because the availability of strong local institutions such as schools or community-based organizations can make a difference in their lives. One way to improve children’s environments is to offer housing vouchers to current public housing residents so that they can move to better neighborhoods. Other options include building new public housing units in mixed-income communities or requiring developers to include affordable rental units in their new projects. These dispersal strategies should be distinguished from direct efforts to improve low-income neighborhoods. As Ludwig notes, efforts to improve depressed areas through empowerment zones or community development programs have not been particularly successful. They also are more expensive than mobility strategies that repro-

1269-01 CH01 04/18/03 17:23 Page 13



13

gram existing housing subsidies, rely heavily on private housing markets to accomplish their objectives, and tap into people’s natural desire to live in better neighborhoods. Ludwig argues that dispersal strategies merit serious consideration. He bases his assessment on recent findings from a housing voucher experiment called Moving to Opportunity that the U.S. Department of Housing and Urban Development has operated in five major U.S. cities since 1994. The program offers low-income families with children currently living in public or project housing the opportunity to relocate to a better neighborhood. The adults who moved reported improvements in mental health as well as in parenting behavior, although no consistent changes were noted in employment or earnings. Their children experienced improvements in mental and physical health and a decline in antisocial behavior, including violent crime. In addition, the children who moved did better in school as measured by test scores in reading and math. These encouraging findings are somewhat preliminary and do not address the possible negative effects that moving large numbers of disadvantaged families to new neighborhoods might have on their new communities, an issue that is likely to create political opposition to any large-scale effort of this sort. Nonetheless, if carefully designed to deal with such problems, housing voucher programs combined with some counseling appear to have the potential to substantially improve children’s prospects. The last chapter, by John Hills, stands apart from the other chapters in the sense that it does not directly prescribe a particular set of U.S. policy reforms. Instead, this chapter discusses the Blair government’s commitment to eliminating child poverty over twenty years in the United Kingdom, the means adopted to achieve that objective, the effects to date, and the possible implications of this British experiment for U.S. antipoverty policy. The Labour party’s agenda has included tax and benefit reforms, a work-based approach to alleviating unemployment, improved public services in low-income neighborhoods and for vulnerable groups, and child care and early education programs. Increased emphasis is being placed on making sure that “work pays” but not by denying increased benefits to those, such as single mothers, who remain out of work. More emphasis has been placed on the carrots of assistance (for example, work supports) than on sticks (sanctions or work requirements). Heavy reliance has been placed on using the tax system to raise incomes at the bottom. But in contrast to the earned income tax credit’s annual lump-sum payments, the United Kingdom’s child tax credit system, introduced in April 2003, will disburse payments regularly throughout the year, usually to the mother. Previous research suggests that this form of disbursement results in spending patterns that are more favorable to children.

1269-01 CH01 04/18/03 17:23 Page 14

14

 . 

Thus far the reforms appear to have boosted incomes in a way that has enabled the government to claim that its interim objective of a one-quarter reduction in child poverty by 2004 is within reach. Although facing some of the same political constraints as the United States, the United Kingdom has, in the end, relied heavily on directly boosting the incomes of poor families rather than on such indirect approaches as encouraging work and marriage. Whether such a direct approach would be politically feasible in the United States remains questionable, but the apparent success of the effort in both substantive and political terms should be of interest to those U.S. policymakers with similar ambitions. Given its more conservative public philosophy and lack of parliamentary government, the United States is less likely than Britain to make the kind of investments in children suggested in this book. And even if it were to do so, there is no guarantee that all of them would work to improve children’s futures. But there is no reason to believe that current policies are optimal. Indeed, there is a very good chance that the ideas put forward in this volume would simultaneously reduce child poverty and a variety of other costly social problems, from welfare dependency to crime; save money over the longer run; increase social mobility; and bring the United States closer to being the “land of opportunity” celebrated in our history and culture.

References Clark, Rebecca L., and others. 2001. “Federal Expenditures on Children: 1960–1997.” Occasional Paper 45. Washington: Urban Institute. Congressional Budget Office. 2000. The Budget and Economic Outlook: Fiscal Years 2001–2010. Currie, Janet. 2000. “Early Childhood Intervention Programs: What Do We Know?” Discussion Paper. Brookings Roundtable on Children. Ellwood, David, and Christopher Jencks. 2003. “The Spread of Single-Parent Families in the U.S. since 1960.” In Public Policy and the Future of the Family, edited by Daniel Patrick Moynihan, Lee Rainwater, and Timothy Smeeding. Russell Sage Foundation (forthcoming). Heckman, James J. 2000. “Policies to Foster Human Capital.” Research in Economics 54 (1): 3–56. London: Academic Press. Mayer, Susan E. 1997. What Money Can’t Buy: Family Income and Children’s Life Chances. Harvard University Press. Moore, Kristin, Sharon Vandivere, and Zakia Redd. 2003.“Cumulative Measures of Stress, Risk, Turbulence, and Community Connectedness: Item Choice, Reliability, and Validity.” Assessing the New Federalism Working Paper. Washington: Child Trends. Sawhill, Isabel. 1999. “Families at Risk.” In Setting National Priorities: The 2000 Election and Beyond, edited by Henry J. Aaron and Robert D. Reischauer, 97–135. Brookings.

1269-01 CH01 04/18/03 17:23 Page 15



15

Sawhill, Isabel, and Laura Chadwick. 1999. “Children in Cities, Uncertain Futures.” Survey Series 1. Center on Urban and Metropolitan Policy. Brookings. Schorr, Lisbeth B. 2003. “What Works to Raise Community Rates of School Readiness: Findings from the Pathways Mapping Initiative.” Discussion Paper. Brookings Roundtable on Children. U.S. Census Bureau. 2001. Statistical Abstract of the United States: 2001, 21st ed. Washington: Hoover’s Business Press.

1269-02 CH02 04/18/03 17:23 Page 16

2 Promoting the Healthy Development of Young Children GREG J. DUNCAN AND K AT H E R I N E M AG N U S O N

A

s documented in From Neurons to Neighborhoods: The Science of Early Development, a comprehensive summary of research on children’s development in the first five years of life, children’s preschool years are a time of great risk—and opportunity— for their long-term development.1 Crucial needs during this period include prenatal and perinatal health care; adequate economic resources; responsive caregivers; and, when called for, intensive early education and mental health interventions. The results of improving children’s environments appear to differ depending on whether the initial environment is deprived or normal. Although avoiding deprivation is crucial to an infant or preschooler’s healthy development, research on the payoffs from enhancing already adequate environments often fails to discover significant impacts. What are the most important policy issues for children’s healthy development? There is clear evidence of the dangers of prenatal exposure to poor nutrition, infections, and environmental neurotoxins (for example, alcohol and lead), most of which can be readily prevented with established interventions. Routine pediatric 1. Shonkoff and Phillips (2000).

16

1269-02 CH02 04/18/03 17:23 Page 17

  

17

care for infants can effectively screen for and treat many important causes of early developmental delays. Developmental research also has established the importance of the quality of caregiving during the first year of life. How early maternal employment, whether voluntary or mandated, affects a child’s quality of care and subsequent development is unclear, but recent studies suggest that it is linked to problems for at least some groups of infants in lowincome families. As a result, it appears sensible to endorse policies that provide parents with genuine choices regarding arrangements for the care of their infants. Recent experiments and many nonexperimental studies suggest that economic deprivation during the preschool years is harmful to children’s development. Accordingly, public policies should minimize the chance that children experience severe economic hardship early in life, although policymakers should be mindful of the employment disincentives that such policies may create. Intervention research indicates that early childhood may be a particularly fruitful time for additional investments in educational programs if family and neighborhood environments fail to ensure that children are ready for school. As a whole, the study of early childhood suggests that policies should treat the preschool years as a distinct period and recognize that low-income families often have difficulties meeting some early childhood needs. Recognizing both intervention opportunities and resource constraints, it appears that too little is being spent on children, particularly younger children. We propose a package of policies that —provide prenatal care, perinatal screening, and health insurance coverage for other medical needs —provide early childhood educational and mental health services for needy children —offer parents genuine choices regarding infant care arrangements —ensure a stable and minimally adequate standard of living for preschool children. Elements of our proposal include —universal prenatal and perinatal screening services plus health insurance coverage. —intensive center-based early education programs for high-risk children, beginning at age three. These programs would be in addition to the universal pre-K programs for four-year-olds presented in chapter 7 of this volume. We also recommend conducting experiments to develop interventions

1269-02 CH02 04/18/03 17:23 Page 18

18

 .    

to address severe behavioral and emotional problems in young children and to guide the considerable investment in those problems that is needed. —exemption from all welfare-related work requirements for mothers of children younger than six months and exemption from full-time (more than thirty hours a week) work requirements for mothers of children between six months and one year of age. Work requirements imposed on mothers with children older than twelve months would not be affected. At the same time, working parents (one per family) would be entitled to six months of parental leave. —a child allowance for children under the age of five living in families with incomes under $60,000. The (taxable) allowance would be $300 per month during the child’s first year and $200 per month between ages one and five, with no benefits for older children. Families with two children under age five would receive allowances for both children. No additional allowances would be extended to families with more than two age-qualifying children. We are mindful of both the total cost and the distributional consequences of our policy proposals. We estimate our proposal’s net cost to be $29.1 billion per year. We expect its beneficial impacts on child development to be several times that amount. Children in low-income families would benefit most from the policies we advocate. Our recommended expansion of prenatal and perinatal medical care and intensive early education and behavioral/ emotional interventions targets children in low-income families almost exclusively. Our package of economic supports would boost the family income of single mothers earning $10,000 annually by between $1,900 and $3,000. The income increase for a single mother with $20,000 earnings could be as much as $3,200, although the figure depends crucially on her state’s child care subsidy schedule. Low-income married couples also would see an income boost, but in this case the income change depends somewhat on how much each couple earns. Couples earning between $30,000 and $50,000 per year would enjoy income increases of at least $1,400. Since we phase out benefits at $60,000 and eliminate the current child tax credit and child care tax deduction, the net income of a couple earning $100,000 would fall by $2,000 to $3,000.

Children’s Developmental Needs Children’s developmental needs, rather than maternal employment or other policy goals, determine the design of our package of proposals. We base

1269-02 CH02 04/18/03 17:23 Page 19

  

19

our policy ideas on the voluminous body of research on the minimum set of conditions needed to ensure children’s healthy development.2 The case for our recommended health and early childhood education interventions is strong. The consequences of prenatal and perinatal threats to children’s healthy development are well understood, and the policy responses to them are straightforward. Results from several evaluations of high-quality, centerbased interventions have demonstrated remarkable improvements in the lives of economically disadvantaged children. Our proposed work exemptions for mothers caring for infants and allowances for families with preschool children are based on research indicating that early maternal employment and low household income may be detrimental to young children. Accordingly, our policies are designed to give parents of young children genuine choices of care for their children and a modest but steady, nonstigmatizing source of economic support.3 We are not advocating a wholesale return to the failed support policies that produced the welfare reform revolution of the 1990s. Instead, we begin with scientific evidence on children’s developmental needs and fashion a set of policies that meet those needs. Despite their professed goal of promoting children’s well-being, current policies favor maternal employment over child development. We reverse those priorities in order not to promote maternal employment at the expense of children’s well-being. In fact, relatively few such conflicts emanate from our proposals. This is due in part to the age-specific nature of children’s needs. Since research suggests that meeting children’s economic needs is more important in the preschool years than later, our child allowance applies only to children under the age of six and falls by one-third after the first year of life. Since research warns only of adverse consequences of maternal employment in an infant’s first months, our proposal for supporting parents’ choice in care for their children applies to infants but not to older children. Since economic research suggests a rather modest work disincentive from an increase in unearned income, on average our proposed child allowance would have relatively little impact on mothers’ work effort.

Prenatal Care, Perinatal Screening, and Health Insurance Because developing fetuses and infants are especially sensitive to harmful environmental influences and biological insults, early and adequate prena2. Shonkoff and Phillips (2000). Duncan served on the committee that wrote From Neurons to Neighborhoods and Magnuson assisted in drafting portions of some of its chapters. 3. In these respects our proposals resemble those of Blau (2001).

1269-02 CH02 04/18/03 17:23 Page 20

20

 .    

tal care and screening provides one of the most cost-effective ways to promote the development of young children. Medical professionals recommend that women have at least one prenatal visit during the first trimester of pregnancy because it reduces the risk of costly and deleterious birth complications. It is especially important that medical professionals address lowincome mothers’ mental health needs. In the United States, rates of prenatal care have increased over the past fifteen years. As of 1998, some 83 percent of mothers received prenatal care in the first trimester, although rates were as low as 63 percent for mothers who were at higher risk for birth complications, such as adolescent and lowincome mothers. Only 4 percent of mothers received no prenatal care or very late care (during the third trimester).4 A 1997 survey indicated that of women who had not received prenatal care, about one-third lacked insurance or the financial means to pay for their visits.5 Because we are seeking to cover the costs of one or perhaps two additional prenatal visits, as well as education and outreach for a relatively small group of mothers, we suspect that the necessary funding may be rather modest. Trends in unmet health needs of young children also are encouraging. According to data collected from January to June 2001, some 11 percent of children under eighteen years of age were without health insurance coverage, down from nearly 14 percent in 1997. Maternal reports in the 1997 National Health Interview Survey show that only 2 percent of children between birth and four years of age had unmet health needs, 3 percent had not seen a doctor in the past year, and medical care for 4 percent had been delayed because of concerns about the cost of care. At the same time, the percentage of children with unmet health needs was five times higher among poor children than nonpoor children.6 Given an estimated per-child, per-year health care cost of $1,200 and an estimated 1.1 million children between birth and five years of age with unmet health care needs, the cost of the additional funding reaches $1.3 billion. The total cost of providing insurance to all uninsured children under age six is roughly $3.2 billion.7 The cost of providing health insurance to all uninsured children under age 18 would total approximately $9.8 billion. 4. Martin and others (2002). 5. Centers for Disease Control and Prevention (2000). 6. Bloom and Tonthat (2002). 7. As of February 2002, states were allowed to use their State Child Health Insurance Program (SCHIP) funds to cover prenatal health care for mothers. In addition, unspent SCHIP funds amounting to $3.2 billion are available for state spending through 2006. In the short term, these funds may be able to cover expansion costs. However, in the long term additional expenditures are likely to be necessary.

1269-02 CH02 04/18/03 17:23 Page 21

  

21

Early Education Interventions Intervention studies suggest that the preschool period may be a particularly profitable time for investing in a child’s education. Several recent comprehensive reviews of experimental evaluations of early, child-focused, centerbased programs have concluded that intensive programs improve children’s short-term cognitive development and long-term academic achievement while reducing their placement in special education classes and increasing their promotion to the next grade.8 Some of these programs also improve children’s long-term social behavior, as indicated by fewer arrests and reports of delinquent behavior.9 In addition, recent nonexperimental studies of Head Start, a less intensive and less expensive early education intervention, have shown important long-term behavioral and academic benefits.10 The current national evaluation of Head Start will provide crucial evidence on the effectiveness of program components such as curriculum and of part-day compared with full-day services. The ongoing evaluation of Early Head Start is providing information about the importance of starting centerbased care early in childhood. We propose that this evidence be used to guide a $10 billion expansion in Head Start–type early-education programs targeting high-risk children; that amount would be added to the money spent on the universal pre-K programs for four-year-olds presented in chapter 7. We suspect that the two major uses of additional funds would be to expand coverage and extend existing programs to cover a full day for three-year-olds, although longer-run results from the Early Head Start experiment may establish the wisdom of targeting children even younger. It is tempting to think that less intensive and less costly attempts to enrich the child care environment of preschoolers—for example, by offering child care subsidies to low-income parents and introducing more rigorous regulations—also will be cost-effective. But evidence here is mixed.11 The benefit of increasing the quality of more typical types of child care does not appear to be very large.12 In the absence of substantial general subsidies, attempts to impose higher standards may increase the cost of care to the point that low- and middle-income mothers are forced out of the formal child care market.13 Altogether, the evidence leads us to propose channeling resources 8. Barnett (1995); Farran (2000); Heckman (1999); Karoly and others (1998). 9. Schweinhart, Barnes, and Weikart (1993); Reynolds and others (2001); Yoshikawa (1994). 10. Garces, Thomas, and Currie (2002). 11. Vandell and Wolfe (2002). 12. NICHD Early Child Care Study and Duncan (2002). 13. Currie and Hotz (2002).

1269-02 CH02 04/18/03 17:23 Page 22

22

 .    

into child care allowances that enable parents to purchase higher-quality care and into intensive early childhood education programs, rather than into the less intensive but more universal enhancement of the quality of child care. It also is tempting to argue that some of the money spent on center-based early education programs would be better spent on teaching adults how to be better parents.14 However, a broader assessment of parent-based interventions suggests that although many of them improve parenting, with few exceptions those improvements in parenting do not translate into improvements in low-income children’s academic outcomes.15

Early Intervention for Severe Behavioral and Emotional Problems Research has established that children with serious behavioral and emotional problems have a limited ability to achieve the social and cognitive competencies that serve as the foundation for success in school. Debilitating levels of anxiety and conduct disorders can be diagnosed in some young children and may have long-term effects on their psychological, behavioral, and social adjustment.16 Limited evidence suggests that some therapeutic interventions are effective, but the waiting lists for such services suggests a lack of professionals trained to provide them. Although the Individuals with Disabilities Education Act requires that family services be provided for children with social and emotional problems, in practice nearly all the services received focus on cognitive, language, and motor impairments. Consequently, the policy has not provided comparable screening and intervention services to children with severe behavioral and emotional problems. Furthermore, many of the professionals who come into contact with children at risk for such problems (for example, child protective service workers) are not well trained or supervised in the area of children’s mental health.17 This is just one symptom of a larger lack of coordination in providing services for children with serious behavioral and emotional problems.18 We envision eventual federal support for behavioral and emotional interventions that parallel those currently afforded by Head Start, costing per14. Olds and others (1999). 15. Gomby and others (1999); Magnuson and Duncan (2002). Some programs for parents of children with high levels of aggressive behavior problems have been successful in improving children’s behavior. 16. Knitzer (2000); Shonkoff and Phillips (2000). 17. Shonkoff and Phillips (2000). 18. Knitzer (2000).

1269-02 CH02 04/18/03 17:23 Page 23

  

23

haps $5 billion annually. But our knowledge is insufficient to justify such an expenditure at the current time. Given the importance of young children’s mental health to their later development, we propose that we first lay the foundation for policies to provide more effective early detection and treatment of mental health problems in infants and young children by funding basic and intervention research.

Welfare Exemptions and Parental Leave for Infant Care It is crucial that children have a responsive caregiver during their early months of life. While the vast majority of children whose mothers choose to work in their first year develop normally, emerging research suggests that maternal employment during the first year of a child’s life may have negative effects on the child’s cognitive development and academic achievement that persist into early childhood. Furthermore, those negative effects are particularly pronounced when mothers work more than twenty hours a week and among low-income children.19 Prudent policy should ensure that all parents are given a genuine choice about working when their children are very young. Welfare reform has removed some of those choices for low-income families. Before the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) was passed in 1996, states were required to exempt mothers with infants (children less than twelve months old) from work requirements. Temporary Assistance for Needy Families (TANF) legislation gave states the option of completely eliminating those exemptions. Roughly half of the states have shortened the exemption period, and a few have eliminated it altogether. This is particularly worrisome in light of the forty-hour combined work and activity participation requirements likely to take effect when PRWORA is reauthorized. We propose exempting mothers from all welfare-related work requirements when their children are younger than six months of age and exempting them from full-time work requirements (more than thirty hours per week) when their children are between six months and one year of age. Parental leave policies also have the potential to improve parents’ choices with respect to the care of their infants. The Family and Medical Leave Act enables parents who work for companies with fifty or more employees to take up to three months of leave with job protection following the birth of a child. Here we propose extending the period of parental leave to six months. In

19. Waldfogel, Han, and Brooks-Gunn (2002); Brooks-Gunn, Han, and Waldfogel (2002).

1269-02 CH02 04/18/03 17:23 Page 24

24

 .    

combination with a child allowance of $300 a month, it amounts to partially paid leave financed by the government.

Avoiding Economic Deprivation with a Child Allowance Economic resources matter because they enable parents to pay for food, housing, stimulating home learning environments, and other “inputs” needed for children’s successful development. Children’s basic needs for food and shelter are self-evident. Research on children’s home learning environments shows that stimulation, emotional support, structure, and safety are associated with the well-being of children from all economic backgrounds.20 Poverty and persistent poverty are strongly associated with suboptimal home environments.21 Cognitive stimulation within the home appears to be particularly important for children’s cognitive development.22 The more positive home learning environments of high-income children account for as much as half of the gap in test scores between high- and low-income preschool children and as much as one-third of the gap between the achievement scores of high- and low-income elementary school children.23 Economic resources also may shape children’s lives through their impact on parents’ mental health. Low-income parents are at greater risk of depression and other forms of psychological distress, such as low self-worth and negative beliefs about their ability to control their lives.24 Poor mental health contributes to lower-quality parenting. For example, depressed mothers’ responses to the needs of their children tend to be less consistent, frequent, and positive. Research in this field has emphasized the associations among economic decline, economic strain, parents’ psychological well-being, and children’s outcomes.25 Despite vast correlational evidence, how and even whether family economic resources affect child development remains a controversial issue. Experimental evidence establishing the importance of economic resources comes from a series of evaluations of welfare reform undertaken during the 1990s. Some of these programs augmented family economic resources while 20. Bradley and others (1994). 21. Garrett, Ng’andu, and Ferron (1994); Votruba-Drzal (2002). 22. See chapter 6 of this book. 23. Smith, Brooks-Gunn, and Klebanov (1997). 24. Gazmararian, James, and Lepowski (1995); Pearlin and Schooler (1978); Rosenberg and Pearlin (1978). 25. McLoyd (1990).

1269-02 CH02 04/18/03 17:23 Page 25

  

25

others did not. In all cases, participants were randomly assigned to a treatment group that received the welfare-reform package or to a control group that continued to live under the old welfare rules. Comparable analyses of these data by Morris and others revealed that welfare reforms that both increased work and provided financial supports for working families generally promoted children’s achievement and positive behavior.26 In contrast, welfare reforms that mandated work but did not support families financially had few impacts—positive or negative—on children. Therefore it appeared that merely increasing maternal employment had no impact on children’s achievement but that increasing both work and income did. In the experimental findings, the welfare reform impacts on children depended crucially on the age of the children. Preschool and elementary school children were helped by reforms that increased family resources. (Infants were not studied.) For adolescents, even generous reforms that promoted maternal employment may have increased school problems and risky behavior.27 Turning to the nonexperimental literature, Duncan and Brooks-Gunn coordinated the analyses of twelve groups of researchers working with ten different data sets that followed children and their families through time.28 They found that family economic conditions in early childhood appear to be more important in shaping ability and achievement than do economic conditions during adolescence and that associations between income increments and achievement were much larger for children in low-income families than for those in higher-income families. Evidence from the broader nonexperimental literature on causal links between economic conditions early in life and child achievement is mixed. The analysis by Duncan and others relating children’s completed schooling to average household income in early and middle childhood and adolescence shows that income in early childhood affects children’s academic achievement, particularly for children in low-income families.29 Impressive in the study is that its statistical controls for family income later in childhood address many of the usual concerns about bias from unmeasured variables. A study relating differences in achievement among siblings to differences in their family income at a specific stage of childhood also suggests the importance of economic conditions in early childhood.30 26. Morris and others (2001). 27. Gennetian and others (2002). 28. Duncan and Brooks-Gunn (1997). 29. Duncan and others (1998). 30. Levy and Duncan (2000).

1269-02 CH02 04/18/03 17:23 Page 26

 .    

26

Other studies, few of which focus directly on the role of early childhood family income, reach mixed conclusions regarding the importance of income. Blau finds small and insignificant effects of current income on child achievement and larger (although still modest) effects of long-run income, but he failed to distinguish early childhood as a distinct period.31 Mayer provided a set of tests for omitted-variable bias and found large reductions in the estimated impact of parental income, leading her to conclude that much of the estimated effect of income in the literature is spurious.32 However, her methods are not well suited for testing the income impacts for preschool children living in low-income families. Studies using the instrumental variables technique are unable to estimate income impacts very precisely; some suggest that income matters a great deal33 and others that it does not.34 All in all, there is more than enough evidence—both experimental and nonexperimental—on the importance of income for children’s development to support the wisdom of policies that prevent episodes of deep or persistent poverty during children’s preschool years. Although we favor work-based approaches (for example, the earned income tax credit) to other ways of boosting the incomes of low-income families, we view as crucial the fact that young children’s needs for a minimally adequate family income exist regardless of whether their mothers work. Thus, we propose a modest child allowance that is independent of parental employment. Specifically, we propose a child allowance for preschool children in families with an annual income below $60,000. Payments would be $300 a month during the child’s first twelve months of life, $200 a month between the ages of one and five, and zero afterward. Economies of scale argue for smaller payments to families with two preschoolers, but since no economies of scale exist with formal child care arrangements, we would not reduce the payments in the case of a second child. To avoid creating incentives for people to have large families and in light of some evidence suggesting that family cap provisions in welfare reform may have discouraged additional childbearing, we would not pay more than two allowances to any one family. Our monthly child allowance would help stabilize and enhance economic conditions in the crucial preschool period, particularly for families with few resources. As with the new child tax credit, providing a child allowance to higher-income families (in our case to families with incomes as high as $60,000) recognizes the value of parenting in nonpoor families, avoids the 31. Blau (1999). 32. Mayer (1997). 33. Maurin (2002). 34. Shea (2000).

1269-02 CH02 04/18/03 17:23 Page 27

  

27

stigma associated with means-tested programs, and helps generate political support. The higher ($300 compared with $200 per month) payment during a child’s first year would give mothers of infants more choices with respect to child care and reflect the fact that infant care typically is more costly than care for older children. Compared with annual “payments” from tax credits or deductions, a monthly payment is a more visible recognition of the importance of parenting and is less subject to the exploitation of low-income parents by tax preparation services that offer them high-interest loans against their refunds. With approximately 4.6 million children under the age of twelve months, 3 million of whom are eligible for the child allowance, and 18.3 million between the ages of one and five years, 11 million of whom are eligible, the cost of the child allowance would total $39.4 billion a year. However, as detailed below, the net cost of the child allowance is considerably less since we would count the child allowance as taxable income; consider part (in the case of TANF benefits) or all (in the case of food stamps) of the allowance as “countable” income in determining eligibility for benefits; eliminate the child tax credit; and eliminate the child care tax deduction. With its unconditional payments to families with preschool children, our proposed child allowance swims against the strong political tide of welfare reform. Concerns that the old Aid to Families with Dependent Children (AFDC)–based welfare system discouraged work and encouraged out-of-wedlock births fueled the TANF-based reforms of 1996. Is our proposal so fundamentally at odds with American values that it will forever be politically unthinkable? We believe that the answer is no. First, our proposal rests on scientific evidence regarding the conditions needed to ensure that all children enter school ready to learn. Second, it is consistent with recent political successes in supporting children through the new child tax credit. Third, its structure of time-limited supports is not inconsistent with welfare reform’s message that parents need to work to support their children. None of our proposals would alter work requirements or change any other provisions for mothers with children older than age one.

Funding and Integration with Current Programs All told, the gross cost of our proposals totals $57.9 billion (see table 2-1). Their net cost is much less—$29.1 billion—since our child allowance overlaps with existing government-financed sources of support. Long-run benefits are likely to total more than $29.1 billion.

1269-02 CH02 04/18/03 17:23 Page 28

28

 .    

Table 2-1. Cost and Funding Summary Billions of dollars

Program component Costs Child allowance for children under 5 (infants: $300/month; ages 1–5, $200/month) Exemption of mothers from all welfare-related work requirements until children are 6 months of age; from full-time work requirement when children are 6 months to 1 year. Six months of unpaid maternity leave. Intensive center-based early education programs for high-risk children Intensive interventions for children with severe behavioral and emotional problems (eventual cost) Universal prenatal and perinatal screening and medical care for uninsured children under 5 Total cost Savings Elimination of child credit for children under 6 Considering child allowance as taxable income for federal income tax purposes Considering child allowance as countable income in food stamp and other in-kind assistance programs Reduction of child care tax credit Reduction in TANF block grant (cash assistance) due to states being able to consider one-half of child allowance as countable income Total savings Net cost

Cost/saving 39.4 0.3

10.0 5.0 3.2 57.9 16.8 4.6 3.4 1.7 2.3

28.8 29.1

Source: Authors’ calculations. Total benefits (increased adult productivity, reduced crime, and so forth) are not quantified.

We propose that child allowance income be counted as taxable income; that the existing child tax credit and child care tax deduction be eliminated; and that TANF cash assistance be reduced. Treating the child allowance as taxable income would reduce its net cost by $4.6 billion. Since the existing child tax credit, which is partially refundable, serves the same function as our proposed child allowance, we propose that the tax credit be eliminated for children age five and under but maintained for older children. By 2010, when the child tax credit will amount to $1,000 per child, this change would reduce

1269-02 CH02 04/18/03 17:23 Page 29

  

29

the net cost of our proposal by $16.8 billion. The existing child care tax credit serves some of the same purposes as our child allowance. Thus, we propose eliminating the credit for children age five and under, a step that would save roughly $1.7 billion per year. It is important to consider how our proposed child allowance would be integrated into existing income support programs such as TANF and the food stamp program. It is reasonable to count the child allowance as an income source in these programs, which in the case of the food stamp program would reduce expenditures by $3.4 billion. Although the child allowance serves many of the same purposes as TANF cash assistance, there are many more strings attached to the receipt of TANF than there would be to receipt of the child allowance. We already have spelled out our proposed exemption period (six months for any work or participation requirement; another six months for a work or participation requirement of more than thirty hours a week). Beyond this and federal law, states would be free to stipulate conditions for receipt of TANF benefits. As to how states count child allowance income in computing TANF cash benefits, we propose that states count no more than 50 cents of every dollar of child allowance income. A 50 percent benefit reduction rate ensures that the monthly income of TANF recipients with young children increases by at least $100 per child. But the remaining $100 a month per child for children aged one to five and $150 a month for infants constitute a cost saving worth close to $2.3 billion per year. It is difficult to forecast the economic benefits of our proposal. On the basis of existing intervention research, we would not expect the added income from the child allowance to produce measurable benefits to children growing up in middle-class families. However, research does suggest a possible payoff both in early education interventions and in stabilizing economic resources for children in families in the bottom 20 to 25 percent of the income distribution. Because close to $20 billion of our proposal’s cost goes to low-income families, benefit-cost ratios of 1.5 or higher for these funds would reduce the longrun overall net cost of our proposal to zero.35 Benefit-cost ratios of some intensive early intervention programs are much higher than 1.5.36 35. If low-income is defined as income that is 1.5 times the poverty line, then census data indicate that 5.7 million children live in low-income families. How many of the $27.6 billion in net benefits go to them? We calculate that $19.9 billion does. This is the sum of $14.9 billion in child allowance payments and $12 billion in Head Start and mental health intervention expenditures and health service, less $5.7 billion in food stamp and TANF expenditure reductions and their $1.3 billion share of increased taxes and reduced tax credits and deductions. 36. Karoly and others (1998).

1269-02 CH02 04/18/03 17:23 Page 30

30

 .    

Distribution of Net Benefits To investigate the distributional consequences of our proposals, we simulate their effects on single- and two-parent families with different earnings. Table 2-2 shows our simulations for single-parent families with annual earnings of $0, $10,000, and $20,000. In the case of parents with $20,000 of earnings, the impacts depend greatly on the generosity of child care subsidies in the state of residence. The details behind our assumptions are provided in the appendix to this chapter. For example, take the case of a single mother with two young children who is earning $10,000. Currently, she would be entitled to no TANF cash benefits, but she would qualify for an additional $2,340 in food stamps and a $4,008 earned income tax credit. She would pay $600 in payroll taxes and, net of child care subsidies, would pay $600 in out-of-pocket child care costs. Her net income would be $15,148, more than $5,000 more than her earnings. This example illustrates the work incentives currently in place for single parents working nearly full time at the minimum wage. How would our policies affect this mother’s income? First, the higher income occasioned by her $6,000 child allowance payment would reduce her work effort modestly, so that her earnings would fall from $10,000 to $9,500. Consequently, her earned income tax credit (EITC) payment would be reduced slightly, to $3,810. Because the child allowance would be treated as countable income for her food stamp allotment, federal income tax, and child care subsidy, her food stamp income would fall to $612, she would be taxed $32, and her net child care costs would nearly double, to $1,074. All in all, her net income would rise by $3,098 to $18,246. All told, our package of economic supports would boost the family income of a single mother earning $0 and $10,000 annually by about $2,000 and $3,300 respectively. The income increase for a single mother with $20,000 of earnings could be as large as $3,200, although this depends crucially on her state’s child care subsidy. Low-income married couples also would see an income boost, but in this case the change would depend somewhat on their respective earnings (tables 2-3 and 2-4). Working couples would lose their child care tax credit, while one-earner couples have no credit to lose in the first place. Two-earner couples making $30,000 and $50,000 a year would see income increases of $1,445 and $1,508 respectively. Corresponding income increases for singleearner couples would be $2,753 and $2,452. Since we phase out benefits at $60,000 and eliminate the current child tax credit and child care tax deduction, the net income of a couple earning $100,000 would fall by about

Source: Authors’ calculations.

Gross income Net income Net difference

Policy Child allowance TANF Food stamps EITC Child care tax credit Child tax (refundable) Income tax Payroll tax Child care expenditures Child care subsidies

Earnings

Income and expenses

Dollars

8,820 8,820

... 4,800 4,020 0 0 0 0 0 0 0

0

Current

10,752 10,752 1,932

6,000 2,400 2,352 0 ... ... 0 0 0 0

0

Proposed

16,348 15,148

... 0 2,340 4,008 0 0 0 −600 −600 8,760

10,000

Current

19,922 18,246 3,098

6,000 0 612 3,810 ... ... −32 −570 −1,074 8,286

9,500

Proposed

23,731 19,647

... 0 0 2,547 684 500 −604 −1,200 −2,280 7,080

20,000

Current

28,547 16,363 −3,284

6,000 0 0 2,547 ... ... −1,624 −1,200 −9,360 0

20,000

Proposed

Not generous subsidy

23,469 19,345

... 0 0 2,574 395 500 −724 −1,200 −2,200 7,160

20,000

Current

28,074 22,596 3,251

6,000 0 0 2,674 ... ... −1,534 −1,150 −2,794 6,566

19,400

Proposed

Generous subsidy

Table 2-2. Proposed Policies for Single-Parent Family with Two Children, by Earnings and State Child Care Subsidy

1269-02 CH02 04/18/03 17:23 Page 31

30,000 ... 0 0 441 960 2,000 −1,624 −1,800 −9,360 0 33,401 20,617

Earnings

Child allowance TANF Food stamps EITC Child care tax credit Child tax (refundable) Income tax Payroll tax Child care expenditures Child care subsidies

Gross income Net income Net difference

Source: Authors’ calculations.

Current

Income and expenses

Dollars

35,494 22,062 1,445

6,000 0 0 694 ... ... −2,344 −1,728 −9,360 0

28,800

Proposed

52,960 35,976

... 0 0 0 960 2,000 −4,624 −3,000 −9,360 0

50,000

Current

55,200 37,484 1,508

6,000 0 0 0 ... ... −5,404 −2,952 −9,360 0

49,200

Proposed

Table 2-3. Proposed Policies for Two-Child Family with Both Parents Working, by Earnings

102,960 69,904

... 0 0 0 960 2,000 −17,696 −6,000 −9,360 0

100,000

Current

100,000 66,944 −2,960

0 0 0 0 ... ... −17,696 −6,000 −9,360 0

100,000

Proposed

1269-02 CH02 04/18/03 17:23 Page 32

30,000 ... 0 0 441 0 2,000 −1,624 −1,800 0 0 32,441 29,017

Earnings

Child allowance TANF Food stamps EITC Child care tax credit Child tax (refundable) Income tax Payroll tax Child care expenditures Child care subsidies

Gross income Net income Net difference

Source: Authors’ calculations.

Current

Income and expenses

Dollars

35,968 31,770 2,753

6,000 0 0 568 ... ... −2,434 −1,764 0 0

29,400

Proposed

52,000 44,376

... 0 0 0 0 2,000 −4,624 −3,000 0 0

50,000

Current

54,800 46,828 2,452

6,000 0 0 0 ... ... −5,044 −2,928 0 0

48,800

Proposed

Table 2-4. Proposed Policies for Two-Child Family with One Parent Working, by Earnings

102,000 78,304

... 0 0 0 0 2,000 −17,696 −6,000 0 0

100,000

Current

100,000 76,304 −2,000

0 0 0 0 ... ... −17,696 −6,000 0 0

100,000

Proposed

1269-02 CH02 04/18/03 17:23 Page 33

1269-02 CH02 04/18/03 17:23 Page 34

34

 .    

$3,000 in the case of two-earner couples and by $2,000 for single-earner couples (see appendix for assumptions). Beyond these impacts on cash incomes are the benefits of the expanded medical care and early education and mental health interventions, most of which target children in lowincome families.

APPENDIX 2A Assumptions behind Tables Table 2-1: Costs of Program Components To obtain the $39.4 billion child allowance figure, we use published census data and assume that 65 percent of the 22.9 million children under six are living in families with incomes of less than $60,000; 3 million children qualify for the infant credit and 11.9 million qualify for the credit for children ages one through five: $3,600 × 3 million = $10.8 billion, and $2,400 × 11.9 million = $28.6 billion, for a total cost of $39.4 billion. To obtain the $0.3 billion welfare exemption figure, we rely on current caseload data and assume that 42 percent of children on TANF are under age six and that 5.4 percent are under one year of age. We assume that one-half of the TANF families live in states that exempt mothers from work requirements in the first year of their children’s lives. Of the remaining 2.1 children in states with work requirements, only 5.4 percent of children (0.11 million) receiving TANF are under one year of age. Although some states impose work requirements as early as six weeks, others impose requirements at six months. We assume that on average mothers would receive full TANF cash benefits for an additional three months. Given a median state benefit of $400, 0.11 million × 3 × $400 = $0.13 billion. The additional costs associated with requiring only part-time employment for mothers of children six to twelve months of age are determined in a similar fashion. Again, half of the TANF caseload lives in states that exempt mothers of infants from full-time work requirements. With only 5.4 percent of children under one year of age receiving TANF and an assumed marginal increase in TANF benefits of $300 for six months, the resulting cost is $0.2 billion (11 million × 6 × $300), and the resulting total cost is $0.3 billion. Our $10 billion for added early education intervention funding is an assumed value. Our $5 billion in eventual behavioral/emotional problem intervention funding is an assumed value.

1269-02 CH02 04/18/03 17:23 Page 35

  

35

To obtain the $1.3 billion for prenatal and perinatal screening, we begin with data indicating that 4.6 percent of all children (1.1 million) have unmet health care needs or delayed health care. We further assume that the average Medicaid expenditure per child per year is $1,200, yielding a cost of $1.3 billion (1.1 million × $1,200). If additional spending were to attempt to cover the health care costs of all uninsured children under age six (11.7 percent of children) the necessary funding would equal $3.2 billion.

Table 2-1: Cost Savings To obtain the $16.8 billion figure for cost savings due to the elimination of the child tax credit, we assume that of the families of the 22.9 million U.S. children under age six that are potentially eligible for tax credit, 15 percent are not receiving the child tax credit because their earned income is too low to qualify and 23 percent qualify only for one-half of the credit. The remaining 62 percent receive the full credit (0.23 × 22.9 × $500 = $2.6 billion; 0.62 × 22.9 × $1,000 = $14.2 billion). Total cost savings therefore is $16.8 billion. To obtain the $4.6 billion figure for tax savings due to considering the child allowance as taxable income, we again assume that 22 percent of the children’s families do not pay taxes. We assume a marginal tax rate of 0.15 (recall that the allowance is restricted to families with a gross income below $60,000) for the remaining 78 percent and their $30.5 billion in taxable child allowance income. The resulting estimate for savings is 0.78 × $39 billion × 0.15 = $4.6 billion. To obtain the $3.4 billion figure for reduced food stamp spending, we use recent caseload data indicating that 50 percent of the 22.9 million food stamp recipients are children. Assuming an even age distribution, that results in approximately 3.8 children under age six receiving food stamps. Consequently, about 26 percent of children eligible for the child allowance are receiving TANF/food stamps. Food stamp benefits are calculated by reducing the maximum monthly benefit for a family of three ($335) by onethird of the family income (excluding child care costs and some other items). Assuming that income is increased by $10.2 billion (0.26 × $39.4 billion), food stamp expenditures are reduced by one-third of this amount. Food stamp–related cost savings equal $3.4 billion (10.2 × 0.33). The $1.7 billion in reduced child care tax credits was estimated with data from the Congressional Budget Office. To obtain the $2.3 billion reduction in TANF funding, we assume that 1.7 million children under the age of six are receiving TANF, which is about 12 percent of all children who are receiving the child allowance. Since we

1269-02 CH02 04/18/03 17:23 Page 36

36

 .    

consider only 50 percent of the child allowance as countable income for the purposes of calculating TANF benefits, the result equals $2.3 billion (0.12 × $39 billion).

Tables 2-2, 2-3, and 2-4 These tables assume that families have two children, one twelve months of age or younger and the second between one and five years of age. In the case of married parents we assume that they file joint taxes. In the case of dual earner couples, we assume equal earnings for each spouse. The TANF and food stamp benefit calculations are based on information from the 2000 Green Book. TANF benefits are based on the median state maximum monthly benefit for eligible mothers with two children, $400. EITC benefits were calculated from the assumed levels of earned income and the 2001 benefit schedule. The child care tax credit is a refundable credit that amounts to between 20 and 30 percent of child care expenditures up to $4,800 for two children. It is available only to families in which both parents are employed or looking for work. The child tax credit is only partially refundable, and we calculate the amounts based on the rate of $1,000 per child. (It currently is $600 per child and is scheduled to increase to $1,000 in 2010.) Income tax is based on 2001 rates. Families are assumed to have claimed the standard deduction for a family with two children. Payroll tax is assumed to amount to 6 percent of earned income. We assume that unsubsidized child care costs amount to $390 a month per child, likely a conservative estimate for center-based child care but an overestimate for non-center-based child care. Wisconsin was chosen to illustrate the “generous child care subsidy” state. For working families with two children, subsidies are given to families with an income of up to approximately $27,000. The estimated copayment for child care is approximately 11 percent of the earned income and child allowance. Illinois illustrates the “not generous child care subsidy” state. In Illinois families with two children with a gross income exceeding $24,000 are no longer eligible for subsidies. We incorporate the inherent employment disincentive into our estimates by assuming that a 10 percent increase in unearned family income leads to a 2 percent reduction in labor supply, that is, an income elasticity of .20 (Blundell and MaCurdy 1999). In the case of two employed parents, the income elasticity is applied to each parent. In applying this estimate, we first

1269-02 CH02 04/18/03 17:23 Page 37

  

37

assume no labor supply response and calculate the net percentage change in family income from all of the other components of our policy package. We then reduce preproposal earnings by 20 percent of this percent change. Note: The authors are grateful to the Family and Child Well-being Research Network of the National Institute of Child Health and Human Development (2 U01 HD30947-07) for supporting this research.

References Barnett, W. Steven. 1995. “Long-Term Effects of Early Childhood Programs on Cognitive and School Outcomes.” Future of Children 5 (3): 25–50. Blau, David M. 1999. “The Effect of Income on Child Development.” Review of Economics and Statistics 34 (4): 261–76. ———. 2001. The Child Care Problem. Russell Sage. Bloom, Barbara, and Luong Tonthat. 2002.“Summary Statistics for U.S. Children: National Health Interview Survey, 1997.” Vital and Health Statistics 10 (203). Hyattsville, Md.: National Center for Health Statistics Blundell, Richard, and Thomas MaCurdy. 1999. “Labor Supply: A Review of Alternative Approaches.” In Handbook of Labor Economics, vol. 3, edited by Orly Ashenfleter and David Card, 1559–1695. Elsevier Science. Bradley, Robert H., and others. 1994. “Early Indications of Resilience and Their Relation to Experiences in the Home Environments of Low Birthweight, Premature Children Living in Poverty.” Child Development 65 (2): 346–60. Brooks-Gunn, Jeanne, Wen-jui Han, and Jane Waldfogel. 2002. “Maternal Employment and Child Cognitive Outcomes in the First Three Years of Life: The NICHD Study of Early Child Care.” Child Development 73 (4): 1052–72. Centers for Disease Control and Prevention. 2000. “Entry into Prenatal Care: United States, 1984–1997.” Morbidity and Mortality Weekly Report 49 (18): 393–98. Currie, Janet, and V. Joseph Hotz. 2002. “Accidents Will Happen? Unintentional Childhood Injuries and the Effects of Child Care Regulations.” JCPR Working Paper 268. Northwestern University. Duncan, Greg, and Jeanne Brooks-Gunn, eds. 1997. Consequences of Growing Up Poor. Russell Sage. Duncan, Greg, and others. 1998. “How Much Does Childhood Poverty Affect the Life Chances of Children?” American Sociological Review 63 (3): 406–23. Farran, Dale C. 2000. “Another Decade of Intervention for Children Who Are LowIncome or Disabled: What Do We Know Now?” In Handbook of Early Childhood Intervention, 2d ed., edited by Jack Shonkoff and Samuel Meisels, 510–48. Cambridge University Press. Garces, Eliana, Duncan Thomas, and Janet Currie. 2002. “Longer-Term Effects of Head Start.” American Economic Review 92 (4): 999–1012. Garrett, Patricia, Nicholas Ng’andu, and John Ferron. 1994. “Poverty Experiences of Young Children and the Quality of Their Home Environments.” Child Development 65 (2): 331–45.

1269-02 CH02 04/18/03 17:23 Page 38

38

 .    

Gazmararian, Julie A., Sherman A. James, and James M. Lepowski. 1995. “Depression and Black and White Women: The Role of Marriage and Socioeconomic Status.” Annals of Epidemiology 5: 455–63. Gennetian, Lisa, and others. 2002. How Welfare and Work Policies for Parents Affect Adolescents. New York: Manpower Demonstration Research Corporation. Gomby, Deanna S., Patti L. Culross, and Richard E. Behrman. 1999. “Home Visiting: Recent Program Evaluations Analysis and Recommendations.” Future of Children 9 (1): 4–26. Heckman, James. 1999. “Policies to Foster Human Capital.” Paper presented at the Aaron Wildavsky Forum, Richard and Rhoda Goldman School of Public Policy, University of California at Berkeley, April 8. Karoly, Lynn A., and others. 1998. Investing in Our Children: What We Know and Don’t Know about the Costs and Benefits of Early Childhood Interventions. Santa Monica, Calif.: RAND. Knitzer, Jane. 2000. “Early Childhood Mental Health Services: A Policy and Systems Development Perspective.” In Handbook of Early Childhood Intervention, 417–38. Levy, Dan, and Greg Duncan. 2000. “Using Sibling Samples to Assess the Effect of Childhood Family Income on Completed Schooling.” Institute for Policy Research, Northwestern University. Magnuson, Katherine, and Greg Duncan. 2003. “Parent- vs. Child-Based Intervention Strategies for Promoting Children’s Well-Being.” Monographs in Parenting (forthcoming). Martin, Joyce A., and others. 2002. “Births: Final Data for 2000.” National Vital Statistics Reports 50 (5). Hyattsville, Md.: National Center for Health Statistics. Maurin, Eric. 2002. “The Impact of Parental Income on Early Schooling Transitions.” Journal of Public Economics 85 (3): 301–32. Mayer, Susan. 1997. What Money Can’t Buy: The Effect of Parental Income on Children’s Outcomes. Harvard University Press. McLoyd, Vonnie. 1990. “The Impact of Economic Hardship on Black Children and Families: Psychological Distress, Parenting, and Socioemotional Development.” Child Development 61 (2): 311–46. Morris, Pamela A., and others. 2001. How Welfare and Work Policies Affect Children: A Synthesis of Research. New York: Manpower Demonstration Research Corporation. NICHD Early Child Care Research Network and Greg Duncan. 2002. “Modeling the Impacts of Child Care Quality on Children’s Preschool Cognitive Development.” Working paper. Northwestern University. Olds, David L., and others. 1999. “Prenatal and Infancy Home Visitation by Nurses: Recent Findings.” Future of Children 9 (1): 44–65. Pearlin, Leonard I., and Carli Schooler. 1978. “The Structure of Coping.” Journal of Health and Social Behavior 19 (1): 2–21. Reynolds, Arthur J., and others. 2001. “Long-Term Effects of an Early Childhood Intervention on Educational Achievement and Juvenile Arrest: A 15-Year Follow-up of Low-Income Children in Public Schools.” Journal of the American Medical Association 285 (18): 2339–46. Rosenberg, Michael, and Leonard I. Pearlin. 1978. “Social Class and Self-Esteem among Children and Adults.” American Journal of Sociology 84 (1): 53–77. Schweinhart, Lawrence J., Helen V. Barnes, and David P. Weikart. 1993. Significant Benefits: The High/Scope Perry Preschool Study through Age 27. Ypsilanti, Mich.: High/Scope Press.

1269-02 CH02 04/18/03 17:23 Page 39

  

39

Shea, John. 2000. “Does Parents’ Money Matter?” Journal of Public Economics 77 (2): 155– 84. Shonkoff, Jack P., and Deborah A. Phillips, eds. 2000. From Neurons to Neighborhoods: The Science of Early Childhood Development. Washington: National Academy Press. Smith, Judith, Jeanne Brooks-Gunn, and Pamela Klebanov. 1997. “The Consequences of Living in Poverty on Young Children’s Cognitive Development.” In Consequences of Growing Up Poor, edited by Greg J. Duncan and Jeanne Brooks-Gunn: 132–89. Russell Sage. U.S. House of Representatives, Committee on Ways and Means. 2000. 2000 Green Book: Background Material and Data on the Programs within the Jurisdiction of the Ways and Means Committee. 106th Cong., 2d sess. Vandell, Deborah, and Barbara Wolfe. 2002. Child Care Quality: Does It Matter and Does It Need to Be Improved? Washington: Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services. Votruba-Drzal, Elizabeth. 2002. “Income Changes and Cognitive Stimulation in Young Children’s Home Learning Environments.” Institute for Policy Research, Northwestern University. Waldfogel, Jane, Wen-jui Han, and Jeanne Brooks-Gunn. 2002. “The Effects of Early Maternal Employment on Child Cognitive Development.” Demography 39 (2): 369–92. Yoshikawa, Hirokazu. 1994. “Prevention as Cumulative Protection: Effects of Early Family Support and Education on Chronic Delinquency and Its Risk.” Psychological Bulletin 115 (1): 28–54.

1269-03 CH03 04/18/03 17:25 Page 40

3 When Work Alone Is Not Enough R O B E R T H AV E M A N

C

onsiderable evidence now indicates that the economic resources of children’s families have a crucial and substantial effect on children’s development and success.1 Evidence that the 1996 welfare reforms have contributed to the widely heralded increase in the employment and earnings of welfare mothers also is convincing.2 Other research, however, reports that the increased work and earnings of these mothers has not led to a marked increase in the overall income of their families.3 For some who have left welfare, the reduction in cash benefits has more than offset the increase in earnings, while for others the offset has led to only small gains in net income. Some welfare families have escaped poverty through work, but others have fared badly. Although 1. See Gershoff, Aber, and Raver (forthcoming), Duncan and others (1998), and Brooks-Gunn and Duncan (1997). Mayer (1997) raises doubts about this relationship, though her analysis has been criticized. 2. See Ziliak and others (1997). 3. See Polit and others (2001), Michalopoulos, Schwartz, and Adams-Ciardullo (2000), and Freedman (2000). In addition, “leavers” studies that compared the income level of families during periods after they left welfare with their income before they left welfare have found little overall increase in income. See, for example, Cancian and others (2002).

40

1269-03 CH03 04/18/03 17:25 Page 41

     

41

poverty rates for single mothers have fallen over the last decade, they remain shockingly high. At the beginning of the new millennium, more than 30 percent of children of single parents lived in poverty, even when federal earnings supplements and food stamps were taken into account. At first glance, that result would seem surprising. After all, over the last decade, the emphasis of federal government support for low-income families has shifted to those that are working and earning: at the beginning of the 1990s, low-income working families received about $13 billion (2000 dollars) in federal assistance; by the end of the decade they received more than $70 billion. Most of the increased spending was in the form of increased work supports, such as earnings supplements (for example, the earned income tax credit),4 child care subsidies, or health insurance coverage.5 What, then, accounts for the failure of earnings to override the loss of cash assistance benefits when single mothers leave welfare for the labor market? The obstacles to full-time, year-round work are severe for this population, suggesting limited opportunities for increasing earnings by increasing annual work hours. For these women, intermittent employment and parttime work are common, accounting for some of the earnings shortfall. There are various reasons why low-skill single mothers fail to hold a permanent full-time job, not the least of which is the difficulty of balancing the unexpected demands of children (for example, sickness and behavior problems) with the demands of work. Moreover, many of these mothers are encumbered by severe and multiple physical, mental, and emotional barriers to working.6 Without some yet-unforeseen intervention to overcome these conditions, the income gains that can be achieved by increasing working hours seem small. Another reason for the low earnings of these single mothers is the generally low level of human capital that they bring to the labor market; as a result, the wages they receive are well below the amount necessary to secure a minimally acceptable standard of living. Except in a few high-wage states, wage rates 4. The EITC adds 40 cents to every dollar of earnings up to about $10,000 for families with two or more children, for a maximum benefit of $4,000. This maximum benefit applies to all earnings from $10,000 to about $16,000; it then is phased out at a rate of about 20 percent. The credit is refundable, meaning that families whose tax liability is less than the credit will receive the credit (minus taxes owed) in the form of a payment from the government. 5. These expanded programs will be discussed in more detail. The estimates in the text update those found in Ellwood (2000b). An illustration of this shift in policy is found in Office of the Legislative Auditor, State of Minnesota (2002), which indicates that the four state and federal work-supportive programs that have grown most rapidly accounted for 27 percent of government assistance to low-income working families in 1999, up from 7 percent in 1989. 6. See Danziger and others (2000) and Olson and Pavetti (1996).

1269-03 CH03 04/18/03 17:25 Page 42

42

 

for these mothers average about $8.00 an hour. Working an average of thirty hours a week for a year yields annual earnings before taxes of $12,000. When payroll taxes and the federal earned income tax credit (EITC) are taken into account, the cash income of families earning this amount remains below the poverty line for a family of three ($15,200 in 2002).7 Although the federal government and some states have expanded tax-based work supplementation policies over the last decade, the net benefit of overall tax policy on family income is eroded by payroll and income taxes.8 If increasing the income of poor, working, single-mother families and thereby reducing their poverty and boosting their children’s development are important goals, finding a way to further supplement earnings should be high on welfare reformers’ to-do list. Fortunately, the cupboard of ideas for accomplishing the task is not bare. In this chapter I summarize those ideas, explore a catalogue of other policy approaches to supplementing earnings, and suggest a realistic and practical policy strategy that incorporates some of these ideas.

Earnings Supplementation in Welfare-to-Work Programs One of the most impressive bodies of new research has tracked a series of evaluations of state welfare waivers or programs explicitly designed by states to be evaluated.9 All of the evaluations in this research package are random-

7. The poverty line, payroll taxes, and federal EITC are calculated for a single mother with two children. 8. As the example indicates, federal (and some state) EITC programs do increase the take-home pay associated with work and earnings. In 2001, more than $30 billion of refundable credits were paid to low-income earners. However, because the EITC is at least partially offset by federal, state, and local taxes (especially the payroll tax), the income supplementation impact of these programs is muted. Office of the Legislative Auditor, State of Minnesota (2002) indicates that federal (income and payroll) and state taxes offset about $1,000 of the $4,900 federal and state working-family tax credits (20 percent) for a woman with two children who worked full-time at $6.00 an hour; about $1,400 of the $4,300 of the credits (32 percent) if she earned $8.00 an hour; $1,900 of the $3,100 of the credits (61 percent) if she earned $10.00 per hour; and $2,700 of the $1,900 of the credits (142 percent) if she earned $12.00 per hour. Danziger and others (2002) indicate that for wage-reliant mothers in their sample of low-income mothers from the Detroit area, federal payroll and income taxes reduced monthly income by $164 (table 1) and that the net effect of taxes paid less credits received (including the EITC) resulted in a decrease in family income (table 2). 9. The bulk of these studies have been undertaken by the Manpower Development Research Corporation (MDRC). See the MDRC website for several of these studies (www.mdrc.org), especially Morris and others (2001).

1269-03 CH03 04/18/03 17:25 Page 43

     

43

ized experiments, so their results can be reported with confidence. Some of the programs involved mandatory participation in employment services or work encouragement or both, together with earnings supplementation. Many of the programs had none of these components. My reading of their findings leads to three primary conclusions. First, the programs that require participation in employment services or that supplement earnings had the largest impact on the employment and annual earnings of recipients. Second, nearly all of the programs that have resulted in both increased employment/earnings and family income offered direct (and often substantial) earnings subsidies to low-skill workers. Few programs without supplements increased both work and family income.10 Finally, those programs that have resulted in both increased employment and total income have had substantial and positive effects on school-age children.11 Those programs that failed to result in increased income did not provide earnings supplementation. The primary earnings supplementation programs that provided this evidence are the Canadian Self-Sufficiency Program (SSP), the Minnesota Family Investment Program (MFIP), and Milwaukee’s New Hope Project. Only New Hope covered all families; SSP and MFIP were offered only to those who received or were applying for welfare benefits. While New Hope also offered child and health care subsidies and MFIP provided employment, training, and job search services, SSP did not. These experimental programs illustrate three approaches to increasing the payoff to work through earnings supplementation. They can be characterized as —an income guarantee (welfare) and a graduated earnings subsidy, contingent on full-time employment (SSP)12

10. See Sherman (2001), Berlin (2002), and Zaslow and others (2002). In a review of sixteen welfare-to-work programs by Sherman (2001), programs that emphasized earnings supplementation were nearly the only ones that yielded an increase of 5 percent or more in recipient income. For welfare recipients, the earnings supplements result in more income by working than by relying on cash and near-cash benefits. 11. The same programs also showed larger positive effects on indexes of children’s behavioral and emotional status, school performance, and health/safety. See Morris and others (2001). Sherman (2001) concludes: “Every program that substantially lifted income had mostly good effects on children” (p. 8). Gennetian and others (2002) report that these same programs, however, did not bring benefits to adolescents, who revealed somewhat worse school performance. 12. The earnings subsidy was equal to one-half the difference between actual earnings and an earnings benchmark, so that supplemented earnings were twice full-time minimum wage earnings. The annual income of those covered was $3,000 to $5,000 more than the income of those who worked and earned the same amount, but remained on welfare, without the supplement.

1269-03 CH03 04/18/03 17:25 Page 44

44

 

—an income guarantee (welfare), a work bonus, and a graduated earnings subsidy with no requirement for full-time work (MFIP)13 —a job guarantee plus a graduated earnings subsidy, contingent on fulltime work (New Hope).14 The crucial point is that each of these programs successfully increased both earnings and the total income received by these families.15

Other Reform Options for Increasing Income from Work In addition to the earnings supplementation approach suggested by these welfare-to-work programs, a number of other strategies exist for simultaneously promoting work and increasing family income. Most of them involve national legislation, and all have advantages and disadvantages. The following paragraphs briefly summarize selected strategies and identify their pros and cons.

Raising the Minimum Wage A straightforward approach to encouraging the lowest-paid workers to seek employment and to work is to pass legislation increasing the minimum wage. Opposition to this approach by economists has been muted by research indicating that its adverse employment effects are not as severe as was believed and feared.16 Nevertheless, lingering doubts about legislating wages persist. Moreover, research studies have shown that the benefits of a higher minimum wage do not flow mainly to poor families.17 These concerns notwithstand13. The basic welfare grant was increased by about 20 percent for those who took a job. Earnings are subsidized at a rate of .62, in the form of a .62 “earnings disregard” against the welfare guarantee. (In effect, actual earnings are offset by reduced welfare benefits at a .38 rate.) Child care subsidies and “cashed out” food stamps also were provided. 14. The guaranteed job is a time-limited, minimum-wage community job. The supplemented earnings exceed the poverty line for most low-skill workers. Child and health care subsidies also were provided. 15. On average, the SSP program raised annual income by $1,909 for adults with children six to twelve years of age, reflecting higher earnings from work ($1,209) and work-related supplementation ($700). The comparable figures for MFIP are an increase in income of $1,307, accounted for by higher earnings from work ($751) and work-related supplementation ($556). Calculations in Sherman (2001), based on Morris and Michalopoulos (2001). 16. See Card and Krueger (1995). 17. See Burkhauser, Couch, and Glenn (1996). However, see also Sawhill and Thomas (2001), who analyzed the effects of an increase in the minimum wage of $1 per hour. They conclude that about 60 percent of the increased earnings would go to families in the bottom one-third of the income distribution and that the poverty rate of families whose head earns the minimum wage would fall by nearly one-quarter.

1269-03 CH03 04/18/03 17:25 Page 45

     

45

ing, an increase in the minimum wage would increase the rewards of working for the nation’s lowest-skill workers.

Expanding the Earned Income Tax Credit The earned income tax credit provides both incentives to work through the lowest earnings levels and more than $30 billion in annual income supplements for several million low-income families. The work-related credits available through the program were substantially expanded during the 1990s. It is the nation’s primary weapon in the battle to make work pay, but for single mothers who work full-time at a wage rate of $8.00 or more, EITC benefits are largely or fully offset by federal and state taxes (primarily the payroll tax).18 The case for exempting the poorest of the nation’s families from taxation on their earnings is strong. When welfare checks are not taxed but paychecks for low-wage workers are, overall tax policy—including both work-related credits (for example, the EITC) and taxes on earnings—fails to make it more attractive for them to work than to receive benefits. Expanding the EITC program could provide additional earnings-related income to working poor families, and a number of proposals to do so have been made. One option is to add a second benefit tier to the EITC that would be available only to families with at least one full-time worker. At the average wage rate of low-skill single mothers, about $8.00 per hour, full-time work would be indicated by earnings greater than $10,000 to $12,000 a year. The additional credit—say, $2,000—would be available to all eligible families with earnings of up to $20,000; above that amount, it would be phased down much as the current credit is. Because only families with a full-time worker would be eligible, this extension of the credit, like the full-time work eligibility requirement in the SSP and New Hope programs, provides incentives for more intensive work activity.19 A recent study evaluates the cost and effectiveness of several options for reducing poverty, including those to reduce the marriage penalty implicit in the current EITC program, increase income supplements across the board, concentrate supplements on families with children, and increase the size of the credit for those who work substantial hours. All of these expansions would result in substantial income supplementation and could effectively narrow the poverty gap.20 18. See note 8, above. 19. Sawhill and Haskins (2002). 20. See Sawhill and Thomas (2001). They simulate the budgetary costs of the various options and the gains in family income generated. They conclude that the increased work-related income of the “make work pay” option they prefer would narrow the poverty gap (based on a poverty line equal to twice the official line) by $1.36 for every $1 in budgetary cost.

1269-03 CH03 04/18/03 17:25 Page 46

46

 

A difficulty, of course, is that several of the expansions supplement the income of families well up in the income distribution, where the number of families becomes large. The budgetary cost of such expansions would be substantial. To control costs, the rate at which supplementation is reduced and then eliminated as earnings increase could be increased. However, this option creates increased marginal work disincentives over a range of earnings that already is encumbered by high marginal tax rates.21

Integrating Existing Tax Benefits into a Reformed Tax Structure Until the mid-1990s, low-income families with children received worksupporting tax benefits from three primary sources—the dependent exemption, child care tax credits, and the EITC. In 1997, the child tax credit (CTC) was added to this arsenal; in 2001, it was made partially refundable and scheduled to increase to $1,000 per child by 2010. The annual budgetary cost for this constellation of programs is about $90 billion; its package of workand child-contingent supports provides substantial assistance to some families, and it does encourage low-wage workers to work full-time. However, the work-promoting effects of the programs are tarnished by two other program characteristics. First, the combination of benefits that they offer has a U-shaped distribution, favoring poor and upper-income families over near-poor and middle-income parents.22 Moreover, the combined structure of the programs creates large marginal work disincentives for middle-income parents. A variety of options for correcting these problems exist, although all of them involve additional budgetary costs.23 Sawicky, Cherry, and Denk propose an interesting integration and simplification plan involving a unified simplified family credit.24 In place of the existing set of programs, they suggest a single expanded refundable credit program for families with children. As earnings increase from zero, there would be a 50 percent supplementation rate, for a maximum credit of $3,500 on earnings of $7,000. Between earnings

21. See Wolfe (2002). Office of the Legislative Auditor, State of Minnesota (2002) presents a vivid calculation indicating that, under MFIP, total resources (after out-of-pocket child care expenses) for a single parent with two children who worked full time would not increase at all as the wage rate increased from the minimum wage to $20.00 an hour. The cumulative marginal benefit reduction rate of noncash assistance is largely responsible for this outcome. 22. This problem has been somewhat eased with the addition of the CTC, though until it is fully phased in the distributional problem will persist. 23. See Ellwood and Liebman (2001). 24. Sawicky, Cherry, and Denk (2002); see also Cherry and Sawicky (2000).

1269-03 CH03 04/18/03 17:25 Page 47

     

47

of $7,000 and $16,000, the credit would remain constant; it then would be phased out at a rate of 10 percent over the range of earnings from $16,000 to $36,000. Families with earnings greater than $36,000 would receive a credit of $1,500 (equal to approximately the value of the CTC and the dependent exemptions).25 This integration and simplification would increase the amount of earnings supplementation, yielding more work-related total income for the nation’s poorest working families. It also would substantially reduce the very high disincentives to increasing earnings over the earnings range where those disincentives are now the greatest. It would provide increased support to middleincome families, thus mitigating to some extent the existing U-shaped distribution of benefits. Moreover, the sizable marriage penalty that exists in the current arrangement would be reduced, especially for single parents. And it would simplify the federal tax code substantially, requiring a single table instead of tables for three separate tax subsidies. Increased income supplements, of course, cost money. The basic plan (without additional credits for more than one child) was estimated to cost about $7 billion in tax year 2001.

Removing Poor Working Families with Children from the Federal Tax Rolls Since the mid-1990s, the combination of welfare reform and a highemployment economy has led to a situation in which low-skill parents, especially welfare leavers, have had their untaxed welfare benefits exchanged for earnings subject to the payroll tax. As a result, the total federal tax burden of low-earnings families has grown. Another proposal, more modest than the simplified family credit, is to fully exempt poor families with children from federal tax liability. One way to do that is to provide a full refund of the employee portion of the payroll tax on the first $10,000 of earnings of one parent of a family with children earning less than $15,000 a year, reducing the amount of the refund as family earnings increase beyond $15,000. Such a plan would provide nearly $800 a year in income supplements to working poor families with children, and it would do so in a way that encouraged fulltime work for one parent in each family at a modest budgetary cost. 25. This example is for a family with one child. The proposed schedule for larger families would be scaled up, with a maximum credit for families with two children equal to about $5,200 and a maximum credit for a family with three or more children equal to $7,000. The marginal supplementation rate and the marginal phase-down rates would remain the same as for the onechild family example.

1269-03 CH03 04/18/03 17:25 Page 48

48

 

Adoption of a Marginal Employer-Based Employment Subsidy All of the policy measures discussed so far directly supplement the market earnings of low-wage workers. By making work pay, they seek to increase both the income of these workers and the supply of labor. If these policy measures are to be successful, the demand for the services of low-wage workers must be sufficient to absorb any increase in their supply. Complementary, demandside measures therefore can also play a role in increasing the rewards of working. One such measure involves providing financial incentives to employers to hire more low-skill workers than they would otherwise employ. One form of employment subsidy that targets enterprises might work as follows: The government would provide a tax credit (or other financial subsidy) to any enterprise equal to, for example, 50 percent of the first $10,000 of wages paid to as many as fifty workers hired by the firm over and above 102 percent of the workers hired by the firm in the previous year. This arrangement affects the marginal choices of firms, directly influencing their decisions regarding both the number of employees to hire and their composition. While this arrangement does not explicitly distinguish among workers by their unemployment or poverty status, the subsidy (and hence the incentive to hire workers) pays a higher percentage of the wages of low-skill than higher-skill workers. This form of employment subsidy would alter the terms on which such workers are employed, making the hiring of low-skill workers a more profitable proposition than it is now. It is designed to ensure that jobs are available for the increases in labor stimulated by the supplyside measures.26 A number of other proposals have been made to encourage the employment and increase the income of low-skill workers either by directly assisting them to find and retain jobs or by making the package of gains associated with increased work more attractive. There are a number of suggestions in the former category for providing work-related services in the form of child care and transportation assistance or programs to enhance job skills and retention.27 Expanding public service employment and reforming the 26. A possible structure for this program is that of the New Jobs Tax Credit that was in place in the United States during the late 1970s. Evaluations of the new jobs program concluded that it was a potent and cost-effective measure for increasing employment of low-skilled workers. For evidence on the effect of the New Jobs Tax Credit on employment levels, see Haveman and Bishop (1979) and Perloff and Wachter (1979). For discussion of the advantages and costs of a two-pronged, supplyand demand-side labor market strategy, see Haveman (1988). A similar two-pronged approach for increasing work and earnings of low-skill workers recently has been recommended for Germany by the prestigious German Institute for the Study of Labor (IZA) (2002). 27. For example, see Moffitt (2002) and Campbell, Maniha, and Rolston (2002).

1269-03 CH03 04/18/03 17:25 Page 49

     

49

federal unemployment insurance program, such as by reducing the time necessary to qualify for benefits or making those who lose part-time work eligible for benefits, fall into the second category.28 In recent reforms of welfare programs, a number of states have directly increased the number of welfare recipients returning to work by increasing “earnings disregards” in calculating cash assistance benefits. While that measure does increase the number of those receiving cash benefits who return to work, it has the negative side effect of encouraging some families to remain on the welfare rolls.

Toward Increased Earnings Supplements for Low-Skill Workers Here I suggest a modest, though coordinated, set of policy measures designed to enhance the take-home pay of low-skill workers and hence the income of their families. The components of this strategy draw on existing research studies and the lessons of experimental efforts to reform the nation’s incomesupport system. While they are separate initiatives, together they form an integrated, work-encouraging income supplementation plan.

A Modest Increase in the National Minimum Wage The case for a modest increase in the minimum wage has been strengthened by recent research indicating that its disemployment effects may have been overstated in past debates. These findings suggest that a modest increase will have little if any adverse effect on employers’ demand for the services of lowskill workers. The current national minimum wage is $5.15, and it has been stalled at this level since 1997. Since 1968, the real value of the minimum wage has fallen by 64 percent.29 Notwithstanding concerns about the “target efficiency” of this measure, a majority of the income gains from an increase of $1 an hour to $6.15 an hour would accrue to families in the bottom one-third of the income distribution and significantly decrease the poverty rate of families whose head earns the minimum wage.30 There are no budgetary costs of this measure, and the adverse impacts on businesses and employment seem minimal.

28. See Blank (2002) and Ellwood (2000a). 29. See www.dol.gov/esa/minwage/coverage.htm [February 12, 2003]. 30. See note 17 and Sawhill and Thomas (2001).

1269-03 CH03 04/18/03 17:25 Page 50

50

 

Accelerating the Phasing-in of the Child Tax Credit While the expanded CTC passed in 2001 ultimately provides $1,000 of childrelated support to families by 2010—and does so in a way that encourages full-time work and reduces the marriage penalty—those effects are muted for the next several years. Work-encouraging support in the near term could be offered to low-earnings families by a more rapid phasing-in of the plan, perhaps making it fully effective within five years of its passage, by 2006.

Supplementing the Earned Income Tax Credit In either its current or an expanded form, the EITC could be supplemented for workers who work a substantial number of hours by incorporating elements of the New Hope or the SSP programs. In both programs, the objective was to direct additional earnings-related income supplements to employees working at least thirty hours a week, which was considered fulltime work. For example, in the New Hope program, the earnings of such full-time workers were supplemented so that when earnings and the supplement were combined with the state and federal EITC, annual household income was near or above the poverty line. Attaining that objective meant that the after-tax income of families of full-time workers would be higher than it would be with only the EITC. The program was designed so that there was always a financial incentive to increase work hours and earn higher wages. The earnings supplement phased out at a rate slow enough that participants always saw at least a thirty-cent increase in total income for each $1 increase in earnings.31 Even though the SSP program was designed for welfare recipients, it also offers an intriguing model. As with New Hope, supplementation was offered only to full-time workers (those employed at least thirty hours a week). For these workers, a target earnings level was set. Any worker with earnings below the target amount received a supplement equal to a fraction (say, one-half) of the difference between the target earnings level and actual earnings. For example, if the target annual earnings level for such full-time work was set at $20,000 and the subsidy rate was 0.5, a worker earn31. The design of the earnings supplement was somewhat complex because it was intended to avoid rewarding unstable earnings patterns; because of interactions between the supplements and the federal and state EITC (which have eligibility rules that differ from those of New Hope) and between the earnings supplements and New Hope’s other financial benefits; and because of adjustments for family size and multiple earners.

1269-03 CH03 04/18/03 17:25 Page 51

     

51

ing $14,000 would receive a subsidy of $3,000 ([$20,000 − $14,000]/2).32 The person’s take-home earnings would then be $17,000. In the SSP program, the objective was to supplement earnings so that they were equal to twice the full-time minimum wage earnings, or about $15,500 in the current U.S. context. SSP offered a more generous earnings supplement to full-time workers than did the New Hope program.33 Either of these “target” supplementation strategies could be viewed as a distinct earnings supplementation program added to the existing or an expanded EITC.34 Alternatively, the structure of either could be integrated with that of the EITC to form a single earnings supplementation strategy. Clearly, analytic work would be necessary to ensure consistency among income and family definitions and to ensure that the combined or integrated structure resulted in a desirable pattern of incentives for work and family composition. Preliminary estimates suggest that an extension of the EITC designed to direct earnings-related income gains to those workers who work a substantial number of hours would simultaneously increase both employment and earnings and take-home income of families headed by low-skill workers.

The Question of Cost One final question remains: How much would this cost? Given the several work and earnings supports that already are in place, the nation could supplement earnings substantially at a relatively small budgetary cost. One way of proceeding would involve consolidating a number of existing family-based work support programs into a simplified and integrated measure with some supplementation. The simplified family credit proposal, discussed previously, is one possibility. In 2003 dollars, the proposal would cost about $10 billion

32. The target earnings level was set at C$37,000 in British Columbia, which is equivalent to about $25,000 in U.S. dollars. After tax obligations and credits were accounted for, most families had incomes of from $2,000 to $5,000 a year more with the supplement than they would have had if they had worked the same number of hours without the supplement. 33. Robins, Michalopoulos, and Pan (2001) simulate the effects of an SSP-type program in which earnings-related supplements are contingent on full-time work, using samples of welfare recipients who have been on the rolls for a year. They find that adding this earnings supplement to the package of TANF, EITC, and food stamp benefits would lead to significant increases in full-time work, recipient earnings, and income relative to the work and earnings incentives these workers faced due to earnings disregards. The costs of the EITC would fall, and the increased incomes of the recipients would substantially exceed the estimated net budgetary cost of the program. 34. See also the suggestion of Sawhill and Haskins (2002), described above.

1269-03 CH03 04/18/03 17:25 Page 52

52

 

per year and would simultaneously improve work-related incentives, provide additional work-related income supplements, and direct the supplements to near-poor and middle-income families.35 For $15 to $20 billion, the nation could supplement earnings even more and achieve a system that truly supports families headed by low-skill workers in their efforts to hold a full-time job. Another option would be to maintain the existing structure of work-based assistance to families but accelerate scheduled program expansions and add a special supplement to encourage full-time work. Implementation of the child care tax credit, which currently is being phased in, could be accelerated to be fully effective by, say, 2005 or 2006. That would provide earnings supplementation for low-income families with a full-time worker. Because the CTC already has been legislated, the cost of accelerating its implementation would be felt only in the first few years; by 2010 acceleration would have no effect on budget projections. Combining this with supplemented EITC benefits for full-time workers along the lines of a national New Hope program would yield a significant improvement in work support for families with lowskill workers able to commit themselves to working full-time. For about $15 billion a year, New Hope–type supplementation ensuring a poverty line income could be provided to all of the 3 million plus poor families with at least one full-time worker.36 Together, this pair of measures could be implemented for about $20 billion in additional costs. Both of these approaches would result in a sea change in work-related income support targeting poor families willing to make a full-time work commitment. Together with a modest increase in the minimum wage, previously discussed, such supplementation would bring about a sizable reduction in poverty for working poor families, together with noteworthy benefits for many of their children.

Conclusion That supplementing the earnings of working poor families will boost their well-being and achievement and promote the development of their schoolage children seems clear. The main question, assuming the political will to 35. For most near-poor families, this plan would provide an increase in earnings supplementation of around 25 percent. This cost “guesstimate” is consistent with an estimated annual cost of a 25 percent expansion of the EITC of about $6 billion (see Sawhill and Thomas, 2001). 36. The per-slot costs of the New Hope program are about $4,500 in 2003, but they include additional child and health care costs. The $15 billion estimate assumes that the earnings supplementation benefits could be provided to these full-time working families at a per-family cost of $3,500.

1269-03 CH03 04/18/03 17:25 Page 53

     

53

reduce poverty through earnings supplementation, is how to structure a set of policies to accomplish that objective effectively. I have summarized some of the evidence on the beneficial effects of earnings supplementation and identified the primary policy instruments for implementing it. The suggested package of policy changes would surely contribute to making work pay and hence decrease poverty in America.

References Berlin, Gordon L. 2002. What Works in Welfare Reform: Evidence and Lessons to Guide TANF Reauthorization. New York: Manpower Demonstration Research Corporation. Blank, Rebecca M. 2002. “Welfare and the Economy.” In Welfare Reform and Beyond: The Future of the Safety Net, edited by Ron Haskins and others. Brookings. Brooks-Gunn, Jeanne, and Greg J. Duncan. 1997. “The Effects of Poverty on Children.” Future of Children 7 (2): 55–71. Burkhauser, Richard V., Kenneth A. Couch, and Andrew Glenn. 1996. “Public Policies for the Working Poor: The Earned Income Tax Credit versus Minimum Wage Legislation.” In Research in Labor Economics, vol. 15, edited by Sol W. Polachek, 65–110. Greenwich, Conn.: JAI Press. Campbell, Nancye, John K. Maniha, and Howard Rolston. 2002. “Job Retention and Advancement in Welfare Reform,” In Welfare Reform and Beyond. Cancian, Maria, and others. 2002. “Before and after TANF: The Economic Well-Being of Women Leaving Welfare.” Social Services Review 76: 603–41. Card, David, and Alan B. Krueger. 1995. Myth and Measurement: The New Economics of the Minimum Wage. Princeton University Press. Cherry, Robert, and Max Sawicky. 2000. “Giving Tax Credit Where Credit Is Due.” Briefing paper. Economic Policy Institute (April) (www.epinet.org [December 31, 2002]). Danziger, Sandra, and others. 2000. “Barriers to the Employment of Welfare Recipients.” Manuscript (revised February 2000) (www.fordschool.umich.edu/poverty/ publications.htm [December 31, 2002]). Danziger, Sheldon, and others. “Does It Pay to Move from Welfare to Work?” Manuscript (revised April 2002) (www.fordschool.umich.edu/poverty/publications.htm [December 31, 2002]). Duncan, Greg J., and others. 1998. “How Much Does Childhood Poverty Affect the Life Chances of Children?” American Sociological Review 63 (3): 406–23. Ellwood, David T. 2000a. “Public Service Employment and Mandatory Work: A Policy Whose Time Has Come and Gone and Come Again?” In Finding Jobs: Work and Welfare Reform, edited by David Card and Rebecca M. Blank. Russell Sage. ———. 2000b. “The Impact of the Earned Income Tax Credit and Social Policy Reforms on Work, Marriage, and Living Arrangements.” National Tax Journal 53 (December): 1063–106. Ellwood, David T., and Jeffrey B. Liebman. 2001. “The Middle-Class Parent Penalty: Child Benefits in the U.S. Tax Code.” Tax Policy and the Economy, vol. 15. MIT Press (http://papers.nber.org/papers/W8031 [December 31, 2002]).

1269-03 CH03 04/18/03 17:25 Page 54

54

 

Freedman, Stephen. 2000. “National Evaluation of Welfare-to-Work Strategies: FourYear Impacts of Ten Programs on Employment Stability and Earnings Growth.” U.S. Department of Health and Human Services, Administration for Children and Families, and U.S. Department of Education, Office of Vocational and Adult Education. Gennetian, Lisa, and others. 2002. How Welfare and Work Policies for Parents Affect Adolescents: A Synthesis of Research. New York: Manpower Demonstration Research Corporation (www.mdrc.org [December 31, 2002]). Gershoff, Elizabeth T., J. L. Aber, and C. C. Raver. Forthcoming.“Child Poverty in the U.S.: An Evidence-Based Conceptual Framework for Programs and Policies.” In Promoting Positive Child, Adolescent, and Family Development: A Handbook of Program and Policy Innovations, edited by R. Lerner, F. Jacobs, and D. Wertlieb. Sage Publications. Haveman, Robert. 1988. Starting Even: An Equal Opportunity Program for the Nation’s New Poverty. Simon and Schuster. Haveman, Robert, and John Bishop. 1979. “Selective Employment Subsidies: Can Okun’s Law Be Repealed?” American Economic Review 69 (May): 124–30. Institute for the Study of Labor (IZA). 2002. “Employment Potential of a Dual Subsidizing Strategy in the Low-Wage Sector.” IZA Compact (May): 2– 4. Mayer, Susan E. 1997. What Money Can’t Buy: Family Income and Children’s Life Chances. Harvard University Press. Michalopoulos, Charles, Christine Schwartz, and Diana Adams-Ciardullo. 2000. “National Evaluation of Welfare-to-Work Strategies: What Works Best for Whom: Impacts of 20 Welfare-to-Work Programs by Subgroup.” U.S. Department of Health and Human Services, Administration for Children and Families, and U.S. Department of Education, Office of Vocational and Adult Education. Moffitt, Robert A. 2002. “From Welfare to Work: What the Evidence Shows.”In Welfare Reform and Beyond. Morris, Pamela, and Charles Michalopoulos. 2001. “The Self-Sufficiency Project at 36 Months: Effects on Children of a Program That Increased Parental Employment and Income.” New York: Manpower Demonstration Research Corporation (www.mdrc.org [December 31, 2002]). Morris, Pamela, and others. 2001. “How Welfare and Work Policies Affect Children: A Synthesis of Research.” New York: Manpower Demonstration Research Corporation. (www.mdrc.org [December 31, 2002]). Office of the Legislative Auditor, State of Minnesota. 2002. Economic Status of Welfare Recipients. Program Evaluation Report 02-05 (January). Olson, Krista, and L. Pavetti. 1996. “Personal and Family Challenges to the Successful Transition from Welfare to Work.” Washington, D.C.: Urban Institute, working paper. Perloff, Jeffrey M., and Michael Wachter. 1979.“The New Jobs Tax Credit: An Evaluation of the 1977–78 Wage Subsidy Program.” American Economic Review 69 (May): 173–179. Polit, Denise, and others. 2001. “Is Work Enough: The Experiences of Current and Former Welfare Mothers Who Work.” Manpower Development Research Corporation, November (available at www.mdrc.org.) Robins, Philip I., Charles Michalopoulos, and Elise Pan. 2001. “Financial Incentives and Welfare Reform in the United States.” Journal of Policy Analysis and Management 20(1): 129–149.

1269-03 CH03 04/18/03 17:25 Page 55

     

55

Sawhill, Isabel, and Adam Thomas. 2001. “A Hand Up for the Bottom Third: Toward a New Agenda for Low-Income Working Families.” Brookings (www.brook.edu/views/ papers/sawhill/20010522.htm [December 31, 2002]). Sawhill, Isabel, and Ron Haskins. 2002. “Welfare Reform and the Work Support System.” In Welfare Reform and Beyond. Sawicky, Max B., Robert Cherry, and Robert Denk. 2002. “The Next Tax Reform: Advancing Benefits for Children.” Economic Policy Institute Technical Paper (draft). Washington: Economic Policy Institute (March). Sherman, Arloc. 2001. “How Children Fare in Welfare Experiments: The Pivotal Role of Income.” Washington: Children’s Defense Fund. Wolfe, Barbara. 2002.“Incentives and Challenges of TANF Design: A Case Study.” Journal of Policy Analysis and Management 21 (4): 577–86. Zaslow, Martha J., and others. 2002. “Experimental Studies of Welfare Reform and Children.” Future of Children 12 (1): 79–95. Ziliak, James P., and others. 1997. “Accounting for the Decline in AFDC Caseloads: Welfare Reform or Economic Growth?” Institute for Research on Poverty Discussion Paper 1151–97. University of Wisconsin–Madison.

1269-04 CH04 04/18/03 17:24 Page 56

4 Preventing Early Childbearing ANDREA KANE AND I S A B E L V. S AW H I L L

O

ne of the most effective ways to improve the life prospects of children is to prevent early childbearing. The children of teen parents are at a disadvantage on numerous economic, social, and health measures. There has been encouraging progress in reducing teen pregnancy in the United States over the past decade, but still nearly four in ten young women become pregnant at least once before the age of twenty, and the United States has the dubious distinction of having the highest rates of teen pregnancy and childbearing among industrial nations. While there is general agreement on the goal of reducing teenage pregnancy, there is far less agreement on the means. Not all approaches to reducing teen pregnancy are effective, and some that are effective have proven to be controversial. However, research has established that there is an approach to preventing pregnancy among teens that is both effective and relatively noncontroversial: after-school programs, particularly those emphasizing community service and intensive engagement of teens in other productive activities during the nonschool hours. In this chapter we propose that this approach be expanded on the grounds that it is a costeffective way to improve children’s futures. 56

1269-04 CH04 04/18/03 17:24 Page 57

  

57

There is widespread support for after-school programs, and federal, state, local, and private investment in them has grown significantly over the past few years. Federal funding for the 21st Century Community Learning Center initiative grew from $40 million in 1998 to $1 billion in 2002, and it is authorized at $1.5 billion in the bipartisan No Child Left Behind Act of 2001. Large majorities of voters strongly support expanded access to after-school programs,1 and both presidential candidates called for investments in afterschool care during the 2000 campaign.2 This support reflects the convergence of several factors: the growing number of working parents; society’s interest in giving young people a safe place to go; and a desire to increase academic performance (although certainly not as a substitute for comprehensive education reform).3 As a result, after-school programs often have multiple goals: providing a safe, supervised environment for the children of working parents, promoting academic success, and encouraging meaningful participation in the community among the nation’s youth. They also can help more young people avoid early childbearing and other risky behaviors, and that in turn would increase their chances of finishing their education, getting a job, and marrying before starting a family. Teen parents have a strong likelihood of ending up poor, on welfare, and in unstable relationships.4 There is some evidence that delaying childbearing would significantly improve adolescents’ status relative to that of their peers who do not delay childbearing and even stronger evidence that it would improve the prospects for the next generation—that is, their children. This multigenerational effect makes the need to invest in teen pregnancy prevention particularly compelling. Unfortunately, only a small percent of adolescents currently participate in high-quality after-school or out-of-school programs that have been shown to have a positive impact on reducing sexual activity and pregnancy as well as other risky behaviors.

The Proposal in Brief To improve the life prospects of children, we propose that the nation make a serious investment in high-quality after-school programs. An investment 1. More than eight of ten voters agree that access to after-school programming in the community is important and that access must be available to all children (Afterschool Alliance 2001). 2. Sylvester and Reich (2000). 3. Hollister (2003). 4. National Campaign to Prevent Teen Pregnancy (2002a).

1269-04 CH04 04/18/03 17:24 Page 58

58

    . 

of $2.5 billion a year to replicate effective models would allow most highrisk teens to participate in after-school programs. Funds would target lowincome neighborhoods or school districts where young people have fewer opportunities for safe, structured activities and where the risks of teen pregnancy are greater.5 To ensure that the funds are well-spent and to deepen our understanding of which models are most effective and why, the investment should include a well-designed evaluation. This proposal could build on one of the existing federal programs, such as the 21st Century Community Learning Centers (21st CCLCs) initiative, which has expanded rapidly in the last few years and places a priority on serving those with the greatest need for expanded learning opportunities. In 2002, 66 percent of the centers were in schools in which at least half of the students were eligible for free or reduced-price lunches.6 However, since most after-school programs serve elementary and middle-school students, a substantial new investment would be required to serve at-risk teens without diverting resources from younger children.7 It should supplement, not displace, the important state, local, and private investments in after-school initiatives around the nation. A word on terminology: while we, and much of the literature, use the term “after-school,” many of the programs being described also operate during evenings, weekends, and the summer. Other common terms include “out-of-school time,” “extended-service schools,” and “extra learning opportunities.” The remainder of this chapter addresses the rationale for our proposal, including what we know about program effectiveness, and describes the proposal in more detail, discussing its estimated costs, benefits, and feasibility.

5. For example, Kirby, Coyle, and Gould (2001) compare teenage birthrates by zip code in California and find that the share of the population living below the poverty level is highly correlated with teenage birthrates. Another factor that could be used to target and recruit girls is whether their mother or an older sibling has experienced a teen pregnancy. 6. U.S. Department of Education (2003). As recently reauthorized, this program provides formula grants to states. States use these allocations to make competitive awards to eligible local education agencies or community-based organizations serving primarily concentrations of poor students. These entities may provide an array of academic and nonacademic activities during nonschool hours or periods when school is not in session. An evaluation conducted by Mathematica Policy Research and released in 2003, as this book was going to press, suggested the program had not been effective. The model outlined in this chapter differs substantially from the one evaluated. 7. In 2001, 80 percent of 21st CCLCs were in elementary and/or middle schools, while only 14 percent were in middle and/or high schools (U.S. Department of Education 2002).

1269-04 CH04 04/18/03 17:24 Page 59

  

59

Rationale Early childbearing imposes significant costs on society and especially on the children of teen mothers. If single parenting has adverse consequences for children—as almost everyone now believes, on the basis of numerous studies—and if a common pathway into single parenthood is an early birth outside of marriage, then a good way to improve children’s futures is to reduce early childbearing. Almost all of the increase in child poverty over the last two decades can be traced to the increase in single-parent families after 1970, a trend driven almost entirely by an increase in childbearing outside of marriage. The pattern of nonmarital childbearing tends to start in the teen years, with about half of all first nonmarital births occurring to teens.8 Teen mothers face a number of challenges. As Lichter and Graefe have demonstrated, young unwed mothers are less likely to marry than those who delay childbearing, and they are less likely to marry successfully.9 Since stable, married parents offer the best prospects for children, children born to young, unwed parents are at a distinct disadvantage. Even after controlling for other factors, teen mothers spend more time as single parents, are more likely to end up on welfare, and stay on welfare much longer than other mothers.10 In addition, their educational prospects are worse than those of their counterparts who wait even a few years to have a baby. For example, only 41 percent of teen mothers ever complete high school; in comparison, 61 percent of young women who delay childbearing until age twenty or twenty-one do so.11 There also is evidence suggesting that teen dads face adverse outcomes in terms of educational attainment, earnings, and incarceration. Teen mothers also have more children on average than women who delay childbearing, making it more difficult for them and their children to escape poverty.12 It is worth noting that all researchers do not agree on the consequences for the teen mothers themselves. Some studies have suggested that the disadvantaged backgrounds of teen mothers are what constrain their opportunities and that an early pregnancy and birth do not add significantly to the poorer outcomes typically experienced by poor mothers regardless of age at 8. Marshall and Sawhill (2003); Thomas and Sawhill (2002); U.S. House of Representatives, Committee on Ways and Means (2002); Ventura and Bachrach (2000). 9. Lichter and Graefe (2001). 10. National Campaign to Prevent Teen Pregnancy (2001). 11. Maynard (1997). 12. Kalmuss and Namerow (1994); Moore and others (1993).

1269-04 CH04 04/18/03 17:24 Page 60

    . 

60

childbearing.13 But these studies have not gone unchallenged, and most researchers continue to believe that although the backgrounds of teens who typically have a baby early in life are responsible for many of the burdens they later face, having a baby as a teen only makes matters worse.14 The adverse consequences for the children of teen mothers are even more striking. A much higher fraction of the children of teen mothers than of older mothers is born prematurely and at a low birth weight. Low birth weight (less than 5.5 pounds) raises the chances of infant death, blindness, deafness, chronic respiratory problems, mental retardation, mental illness, cerebral palsy, and other disabilities. The children of teen mothers also do worse in school. They are 50 percent more likely to repeat a grade, they perform much worse on standardized tests of performance, and they are much less likely to complete high school than the children of older mothers. Children of teen mothers also have poorer home environments and are more likely to experience abuse and neglect and end up in the child protective services or foster care system. Finally, once they become young adults, they are more likely to be poorly educated, to be economically inactive, and to have a baby themselves as a teen. Researchers have had some difficulty sorting out the extent to which these differences are the result of the timing of the birth per se rather than characteristics of the mothers that may have led them to have a baby early. Once again, efforts to control for these differences diminish but usually do not eliminate the adverse consequences for the children. To cite Rebecca Maynard, “what is unambiguously clear . . . is that young teen childbearing has significant adverse consequences for the children and that these consequences are costing taxpayers and society enough to merit close policy attention.”15 She and her coauthors estimated that the costs were at least $7 billion a year, or more than $2,800 per teen mother, in 1997.16 Adjusted for inflation, the cost is $3,300 per teen mother in 2003—which is also the amount of the benefit to taxpayers of preventing teen pregnancy. If after-school programs have other benefits, such as reducing crime or improving academic performance, or if the benefits of preventing teen pregnancies and births have been underestimated, then the savings to taxpayers will be larger. In short, the $3,300 figure is a very conservative estimate of the likely savings for each teen birth that is prevented.

13. Hotz, McElroy, and Sanders (1997). 14. Hoffman (1998); Maynard (1997). 15. Maynard (1997, pp. 19–20). 16. Maynard (1997, p. 18).

1269-04 CH04 04/18/03 17:24 Page 61

  

61

The fact that having a teen mother adversely affects a child’s chances for success would not make much difference if there were no way to prevent early births. But solid research now suggests that we know more about how to reduce teen pregnancies than we used to. Many people think that the most effective way to reduce teen pregnancies is to provide young people with sex education and family planning services. A number of sex education programs have been carefully evaluated, and several have proven effective in reducing sexual activity or increasing the use of contraception.17 Family planning services also can help reduce teen pregnancy by providing access to forms of contraception that are more effective than over-the-counter methods and by further educating teens about the risks of pregnancy and sexually transmitted diseases (STDs) and how to prevent them. There also is considerable political support for “abstinence-only” education and strong public support for teaching teens to abstain from sexual activity; however, in terms of rigorous research, the jury is still out on the effectiveness of specific abstinence-only education programs.18 In the meantime, as detailed below, recent research shows that one very effective way of preventing pregnancy is to involve adolescents in what might loosely be called after-school programs.

What Do We Know about What Works? Despite the strong public support for after-school programs, we know much less than we would like about their impact and about which models work best, for whom. As Fashola, Eccles and Templeton, and others point out, a number of factors contribute to the weak state of research on after-school programs, including their voluntary and dynamic nature, their rapid growth, and the difficulty in defining the treatment.19 Moreover, much of the research on after-school programs focuses on younger children. There are a half-dozen comprehensive reviews of the research on afterschool and youth development programs that have identified some models that have produced strong effects, some that have not, and many whose results are impossible to define due to weak evaluation design. Several youth development/vocational education programs that might have been expected to produce positive results in terms of teen pregnancy and childbearing 17. Kirby (2001). 18. Kirby (2001); National Campaign to Prevent Teen Pregnancy (2002b). 19. Fashola (1999); Eccles and Templeton (2001).

1269-04 CH04 04/18/03 17:24 Page 62

62

    . 

(Conservation and Youth Service Corps, Job Corps, or JOBSTART) did not.20 This literature on youth development programs has been summarized by Rob Hollister for the Brookings Roundtable on Children; in his review, Hollister focuses only on programs that have been rigorously evaluated.21 He finds that a handful of programs have reduced drug use, alcohol use, smoking, violence, school absences, and dropout rates.22 In a recent report for the National Academy of Sciences, Eccles and Gootman also summarize the research as well as the key features of successful programs, citing three in particular that have been well-evaluated and found to have positive effects on youth: the Teen Outreach Program; Big Brothers/Big Sisters; and Quantum Opportunities.23 Evaluations that focus on teen pregnancy as an outcome also have been reviewed, by Doug Kirby for the National Campaign to Prevent Teen Pregnancy and by Child Trends.24 The last two reviews both find that a few programs have been effective in reducing teen pregnancy. One such program is the Teen Outreach Program (TOP), a year-long community volunteer program for high school students that is coupled with weekly, teacher-facilitated group discussions about adolescent values, decisionmaking, communication, parenting, life options, and students’ experiences as volunteers. Experimental evaluation at multiple sites found that participants were much less likely to become pregnant during the school year than the control group.25 Another program, Reach for Health and Community Youth Service Learning, which has similar content but serves middleschool students, produced significant reductions in sexual activity, both in the short term and over a three-year follow-up period. Interestingly, these positive results did not hold up when the health education part of the curriculum was implemented without the service learning component.26 And because the effectiveness of TOP did not depend on faithful replication of the education curriculum, Kirby suggests that the service component may be the most important element of the program. There also is evidence that in addition to having positive impacts on teen sexual activity and pregnancy, service learning has positive academic effects. For example, TOP reduced course failure and school suspensions.

20. Kirby (2001). 21. Hollister (2003). 22. Hollister (2003). 23. Eccles and Gootman (2002). 24. Kirby (2001); Kirby (2002); Manlove and others (2002). 25. Eccles and Gootman (2002); Allen and others (1997). 26. Kirby (2001); Manlove and others (2002).

1269-04 CH04 04/18/03 17:24 Page 63

  

63

The fact that these two service learning models have had such positive effects in preventing teen pregnancy is particularly noteworthy given the growing interest in national and community service, including service learning for young people.27 The federal Learn and Serve America initiative currently provides $43 million to schools, colleges, and community organizations that engage approximately 1.5 million students in service to their communities, and there are a number of state and privately funded initiatives as well.28 It is not entirely clear why these service learning programs had positive effects. However, it is probably because the programs engaged teens in meaningful activities that made them feel needed and useful, provided positive ongoing relationships with program staff, and kept them productively occupied and better supervised than they otherwise would have been. Supervision may be the key. A recent report reveals that among teens who participated in school-based STD screening programs, those who were unsupervised for at least thirty hours a week were significantly more likely to be sexually active than those who were unsupervised for less than five hours a week.29 In addition, for the teenaged girls surveyed (though not teenaged boys), nonparticipation in after-school programs was associated with sexual activity: whereas 71 percent of nonparticipants were sexually active, only 59 percent of participants were. As a Public/Private Ventures (PPV) evaluation of a set of model programs states,“One key goal of after-school programs is to provide youth with productive ways to use their out-of-school time and, thus, reduce their opportunities for risk-taking behavior. Our findings are consistent with [the programs] having this effect.”30 Although most teenagers report having their first sexual experience after 6:00 p.m., a significant proportion have their first experience during the daytime hours, and that proportion may be higher for subsequent encounters.31 A third program, more intensive and of longer duration than the two just described, is the Children’s Aid Society-Carrera (CAS-Carrera) program, which engaged teens ages thirteen to fifteen in a comprehensive mix of activities, from tutoring to sports or work in the community.32 The activities were available five days a week throughout the school year and were supplemented

27. National Commission on Service Learning (2002). 28. Corporation for National and Community Service (2002). 29. Cohen and others (2002). 30. Grossman and others (2002). 31. Child Trends (2002). 32. Philliber and others (2002).

1269-04 CH04 04/18/03 17:24 Page 64

64

    . 

by evening and summer programs; on average, teens participated sixteen hours a month over three years. The program also provided comprehensive medical care, along with sexuality education. The program was rigorously evaluated over four years at multiple sites and was found to have significantly reduced sexual activity among girls (though not boys) and to have reduced the pregnancy rate by more than half: the pregnancy rate was 10 percent among participants but 22 percent among controls. In addition, teens who participated in the program were more likely to graduate from high school and to enroll in college than those in the control group. The program was operated through community organizations and placed a premium on creating close relationships between participants and staff. It served at-risk youth, many of whom were from poor, single-parent families. Still another highly intensive program of long duration, the Quantum Opportunities Program (QOP), also produced encouraging results, although the evidence in this case is less robust.33 This community-based program provided a mix of educational activities, life skills training, career and college planning, and community service and work experience, for which students received stipends. One year after the end of the program, 24 percent of this group of very disadvantaged participants had had a child compared with 38 percent among the control group.34 QOP also showed some positive academic outcomes, although there are some questions about the strength of the findings. Big Brothers/Big Sisters is a less intensive mentoring program that has proven through rigorous evaluation to have very strong positive impacts on a variety of risky behaviors, particularly use of drugs and alcohol, but as is too often the case, sexual activity and pregnancy rates were not tracked.35 All of these positive results have come from studies in which a random assignment evaluation was conducted, suggesting that we can have confidence in the results.36 Important lessons also can be gleaned from demonstrations such as the Extended-Service Schools (ESS) initiative. While the 33. According to Hollister (2003), the evaluation of QOP is misleading for several reasons. With respect to evaluating teen pregnancy prevention, he notes the following: first, a site that failed completely was excluded from the analysis, and second, the lower childbearing rates did not occur at the Philadelphia site; they came only from pooling across sites, several of which suffered from attrition bias. In addition, Eccles and Gootman (2002) discuss the limitations of QOP’s evaluation design, including the fact that most goals included in the program design were not evaluated and the program’s relatively small sample size. 34. Hollister (2003). 35. Hollister (2003); Fashola (1999); Teirney, Grossman, and Resch (1995). 36. However, Eccles and Gootman (2002) argue that, given the flaws in QOP’s evaluation design, we should view the program’s results with some caution.

1269-04 CH04 04/18/03 17:24 Page 65

  

65

multiyear evaluation of this initiative being conducted by the Manpower Demonstration Research Corporation and PPV does not include random assignment, it offers interesting insights about the effects of providing youth development programs during the after-school hours in twenty communities around the country. Each community is adapting one of four nationally recognized extended-service school models. (Six of the sixty ESS programs take place in high schools). The four models share common features and are designed to promote both academic and nonacademic development through a variety of tutoring, sports, recreation, community service, and other activities. The programs take place at a school, but they operate in partnership with a community-based organization or university that controls the funds and the programming. The ability to measure program impacts is limited without a randomly assigned control group, but the evaluation is producing some encouraging evidence that participation is associated with positive effects on attitudes and behavior.37 Finally, Girls Inc. engages girls from twelve to seventeen in a variety of activities, including group discussions, assertiveness training, education activities, and career planning. An evaluation found that when long-term program participants were compared with nonparticipants, younger teens were twice as likely to postpone sexual intercourse, while older teens were half as likely to become pregnant. However, these results are weaker than those for the other evaluations cited above because the Girls Inc. evaluation did not include a randomized control group.38

The Proposal in More Detail Looking across the handful of highly effective models as well as at the lessons and practices from a broader range of after-school programs, we believe that a strong case exists for making a major investment in after-school programs focused on teens from thirteen to seventeen years of age living in atrisk families. Funds should target poor neighborhoods or school districts. Although arguably all teens, and especially those with employed parents, need access to after-school programs, disadvantaged teens are most at risk of pregnancy and least likely to live in communities where structured activities are available.39 The mothers of many of these teenagers are now required 37. Grossman and others (2002). 38. Nicholson, Postrado, and Weiss (1991). 39. Kirby, Coyle, and Gould (2001).

1269-04 CH04 04/18/03 17:24 Page 66

66

    . 

to work as the result of the 1996 welfare law, and some research suggests that their employment has had some negative effects on their adolescent children, although the effects that have been identified to date have involved school performance rather than pregnancy or birth rates.40 Given the evaluation evidence, the variety of goals of after-school programs, and the diversity of communities across the country, we believe that it would be a mistake to fund a single model. Rather, funding should be available to expand any of the effective models cited above or to implement new programs that draw heavily on the lessons learned from these proven programs. Comprehensive research reviews reveal several features of effective programs (particularly those serving teens) that stand out and should guide future investments in after-school programs.41 They include a safe and structured environment, developmentally appropriate and engaging activities, supportive relationships with well-trained and stable staff (or volunteers), strong articulation of positive social norms for behavior, opportunities to master new challenges, a sense of meaning and belonging (through community service and other avenues), and positive relationships with peers. The literature suggests that any new programming should pay attention to the following elements: . The content should include a mix of academics, community service, and social/enrichment activities that are developmentally appropriate for the teens being served.42 Most important is a mix of interesting activities that challenge young people and give them a chance to develop new skills. Particularly for older teens, there is some suggestion that involving them in developing the activities also is important. . While this proposal is based on the benefits associated with reducing early childbearing, effective programs do not need to focus exclusively on that goal. Most of the effective programs have had multiple purposes, and some of them have achieved other desirable outcomes, such as reduced drinking or improved academic performance. According to several reviews of the literature as well as a report by the Department of Education and the Department of Justice, children who regularly attend high-quality programs have better peer relations, emotional adjustment, and better grades and conduct in school; they spend less time watching tele-

40. Gennetian and others (2002); Brooks, Hair, and Zaslow (2001). 41. Hollister (2003); Eccles and Templeton (2001); Fashola (1999); Kirby (2001); Eccles and Gootman (2002). 42. Urban Health Seminar Series on Children’s Health and Safety (2001).

1269-04 CH04 04/18/03 17:24 Page 67

  

67

vision; and they have a lower incidence of drug use, violence, and pregnancy.43 While the primary focus of most after-school programs, including the 21st Century Community Learning Centers, is on activities to boost academic achievement, some communities are using after-school funding to address teen pregnancy.   . While the intensity and duration of effective programs vary, it seems clear that the most effective interventions, especially for at-risk teens, are relatively long lasting (at least a year) and intensive (at least several hours per week). It also is important that new funding be flexible enough to serve teens during evening, weekend, and especially summer hours. This is consistent with the practice of most effective programs, as well as with the direction of the 21st CCLCs, 71 percent of which were open at least fifteen hours a week in the summer in 2002.44   . Effective programs generally offer a safe, supervised place where teens can connect with responsible adults. The quality of the relationships teens form with adults through after-school programs appears to be a key ingredient to success. However, staff turnover is cited as a major issue in several programs, which suggests that adequate resources are needed to hire and retain qualified and committed staff. Training for mentors and other volunteers also is critical. . Schools are one but not the only possible location of afterschool programs. They have clear advantages since they are well-known sites with relatively abundant and flexible space. School-based programs also offer a good way to target low-income children or neighborhoods and to relate the programs to what gets learned during the school day. However, schools have disadvantages as well. As Grossman and her colleagues found, schools may not be as underused as commonly believed and, especially for older teens and at-risk youth, schools may not be appealing sites.45 Some of the effective programs profiled above were operated in the facilities of community-based organizations. Other options include recreation centers and facilities of faithbased organizations. Another factor related to location is ensuring that program sites can be accessed by the teens they are trying to reach. Transportation is a challenge if students do not live near their schools and rely on school buses, which may not run after the school day ends. Effective programs need to find creative solutions to transportation problems: for example, by obtaining funding to employ school buses and drivers after hours, forming partnerships 43. U.S. Department of Education and U.S. Department of Justice (2000). 44. U.S. Department of Education (2003). 45. Grossman, Walker, and Raley (2001).

1269-04 CH04 04/18/03 17:24 Page 68

68

    . 

with community and faith-based organizations that have vans, and choosing locations that are close to concentrations of the students targeted. Another option, used by the New York City Beacon Program, is for school-based programs to serve teens who live nearby, whether or not they attend that particular school during the day.46 . If services are to reach those who might benefit the most, programs need to find creative ways to recruit and continually engage teens. In particular, older teens may be more interested in working (often out of necessity) or in community service. Olsen and others caution that the youth who are most at risk of negative behavior may be least likely to participate in after-school programs, reinforcing the importance of allowing enough flexibility in program design and delivery to allow programs to attract the teens they are trying to reach. QOP and CAS-Carrera, for example, offer a stipend for hours spent in community service or other unpaid employment.47

How Much Would It Cost? The costs of effective programs vary enormously, from about $600 a year per student for TOP to $4,000 a year for an intensive program such as CASCarrera.48 Currently, many after-school programs, including those funded by the 21st CCLC program, Boys and Girls Clubs, and others, spend less than that. (See table 4-1). However, the evidence suggests that the more intensive and expensive programs are the most effective, especially with high-risk teens, although it is not clear that one has to spend as much as the CAS-Carrera model on every teen to get good results. TOP, for example, seems to have been effective with a less intensive, lower-cost intervention, at least during the year in which students participated in the program. Based on all the evidence, spending around $1,000 a year per teen might be sufficient. We view this as an estimate of average costs. That is, some programs might spend more or less than that, depending on the community and the model being used. Multiplying the cost data by the estimated size of the population produces a total estimated cost of $2.5 billion ($1,000 × 2.5 million kids).49 The net

46. Fashola (1999). 47. Grossman and others (2002); Olsen (2000). 48. For more on costs, see Newman, Smith, and Murphy (2000). 49. There are about 20 million teens ages thirteen to seventeen in the United States. According to the Annie E. Casey Foundation (2002), 12 percent of all children live in high-risk families. Assuming that older children are as likely as younger ones to live in these families, about 2.5 million teens are at risk.

1269-04 CH04 04/18/03 17:24 Page 69

  

69

Table 4-1. Annual Program Costs Dollars

Program Teen Outreach Program (TOP) Carrera Big Brothers/Big Sisters After-School Corporation Boys and Girls Clubs

Cost per participant 100–600 572 4,000 1,000 1,000 139

Sources: TOP: first estimate from Cornerstone, “Teen Outreach Program: TOP FAQs” (www.cornerstone.to/top/faq3.html [October 11, 2002]); second estimate from Newman, Smith, and Murphy (2000). Carrera: Carrera, Kaye, and Philliber (2001). All others: Newman, Smith, and Murphy (2000).

cost would be somewhat smaller after funding for existing programs was subtracted from the total. Although data on the number of teens formally enrolled in existing programs are not readily available, by using data from the National Survey of America’s Families (NSAF), we estimate that in 1998 roughly 60 percent of teens between the ages of twelve and seventeen were participating in some kind of organized, structured after-school activities on a regular basis (broadly defined). That equates to approximately 14 million teens.50 In addition to whatever the net cost turns out to be, a small portion of the funding (perhaps $100 million) should be set aside for further evaluation of the programs. Note that one reason that our proposal targets low-income communities is that it makes it feasible to expand effective programs at a reasonable total cost. The downside, of course, is less political support than there would be for a more universally available set of services. However, targeting communities rather than individuals helps to offset any sense that some teens are being denied services and helps to ensure somewhat greater integration of teens from different types of families.

Costs and Benefits It is worth asking under what assumptions the costs of the program are more than offset by the benefits it will provide. Looking just at savings for taxpayers, 50. However, since the NSAF survey question employs broad definitions of qualifying programs and of participation in these programs, using these data may lead to an overestimation of the share of teens participating in after-school programs. These estimates are derived from the authors’ analysis of the Urban Institute’s 1999 National Survey of American Families database.

1269-04 CH04 04/18/03 17:24 Page 70

    . 

70

we believe that the benefits equal at least $3,000 for every teen birth avoided and probably much more than that, as noted above. Now suppose that an afterschool program targeting high-risk youth costs $1,000 per teen and leads to a 50 percent reduction in pregnancies and births. If 100 teens are enrolled and sixty-six would either have become pregnant or have gotten someone pregnant without the program and only thirty-three pregnancies occur with the program, the total costs are $1 million (100 teens × $1,000) and the total benefits are $1 million (thirty-three pregnancies avoided × $3,000 in savings per pregnancy avoided). Are these realistic assumptions? Recalling that nearly four in ten girls in the United States become pregnant at least once before the age of twenty and remembering that the pregnancy rate among girls from more disadvantaged families is far higher than that,51 the assumption that six of ten disadvantaged teens probably will become pregnant in the absence of the program does not seem unreasonable. Finally, recall that in some effective programs the pregnancy rate for those who participated was half the rate of those who did not. Whether a program costing only $1,000 per enrolled teen can be this effective is a little less clear; we wouldn’t want to bet the farm on it. We later recommend that any demonstration and evaluation of this initiative include some planned variations in its cost and intensity that will help to resolve the question. In the interim, it bears repeating that these programs have other benefits that are not included in the figures used here. All in all, we think these after-school programs would be a good investment and one that would largely if not completely pay for itself over the longer run.

More Evaluation Needed Given the uncertainties alluded to above, we think that more research is needed. The good news is that there is strong support among the public and policymakers of both parties for expanding after-school opportunities (and service learning). As a result, investments in such programs have grown rapidly. While this bodes well for additional investments, it also means that the field has grown more quickly than the research base, and we have no strong assurances that all the additional funds are being spent effectively. The quality of programs no doubt varies considerably. Moreover, given the large unmet need for after-school programming, there is a danger that funds may be spread too thinly, while the research generally indicates that the more intensive programs are the most effective. 51. Kirby, Coyle, and Gould (2001).

1269-04 CH04 04/18/03 17:24 Page 71

  

71

Given some uncertainty about costs and benefits and about why certain models do and others do not reduce early childbearing as well as other risky behaviors, some funding needs to be earmarked for careful evaluations that inform states and communities about what types of programs to fund in the future. There are a number of key questions that any evaluation should address. First, would financial incentives help programs attract and retain older youth? Second, how important is the service activity itself—is there something inherent in performing community service that makes a difference, or is it simply a way to engage young people in a constructive, supervised activity? There is some suggestion from TOP and Reach for Health that performing a service itself matters, but that warrants further study. Third, what is the optimal intensity of an after-school program? Given the obvious relationship between cost and intensity, it will be important to learn more about how many hours a week and how many weeks a year a teen needs to participate in order to benefit from the program. Is a program costing $1,000 per enrolled teen optimal or does more or less need to be spent? In short, what is the most cost-effective strategy?

Concluding Thoughts A common pattern in the history of policymaking is to take a few relatively effective models and dilute what is offered in the process of expanding them to serve a much larger group of participants. There is a danger that this mistake will be repeated in the case of after-school programs. This chapter argues for a more intensive and focused effort, but we recognize that political considerations may press in another direction. A second possible concern regarding this proposal is the political viability of linking a relatively noncontroversial activity (after-school programming) to an issue that can quickly provoke controversy (teen pregnancy prevention). While there is a strong consensus about the desirability of preventing teen pregnancy, there is less agreement about how to do so. Interestingly, the 21st Century Community Learning Center program’s purposes include drug and violence prevention and character education, but the enabling legislation fails to mention preventing sexual activity or pregnancy as a purpose, probably because of these controversies. On one hand, many communities are deeply committed to reducing teen pregnancy and could be encouraged to view this outcome as one of several that might be advanced by

1269-04 CH04 04/18/03 17:24 Page 72

72

    . 

effective after-school programs. On the other hand, it may be better to avoid the controversy that such a focus might create. A third likely concern is the extent to which a focus on teen pregnancy and early childbearing leads to an overly negative view of youth themselves. For many working in the field of youth policy and programs, there is a tension between focusing on youth problems or risks (such as teen pregnancy) and what often is described as a more positive, asset-based approach to youth development.52 Some may dismiss or criticize our proposal because it focuses on preventing a bad outcome. However, as Gary Walker notes, this debate is primarily the concern of a small group who spend their lives thinking about youth policy and programs. It is largely irrelevant to the public policymakers who are primarily concerned with solving such problems as crime, drugs, poor school performance, and teen pregnancy.53 A recent National Academy of Sciences report also addresses the tension between the “prevention/ problem-centered” and “youth development” perspectives and concludes that both approaches are valuable and necessary.54 Finally, a number of people, educators in particular, believe the primary focus of after-school programs should be academic achievement. We would agree that this is one of several appropriate goals and point out that the goals of preventing teen pregnancy and increasing academic success are complementary. After all, teens who participate in risky behaviors, including having sex and becoming pregnant, are less likely to finish high school, and, as noted, poorer educational outcomes for both the adolescent mother and her children are a major rationale for offering this proposal.55 Investments in high-quality after-school programs should help young people both to perform better in school and to avoid pregnancy. For all of the reasons noted, we suspect that this proposal will engender some criticism or opposition. Moreover, additional research is needed to answer a number of questions raised in this chapter. But it is not necessary to wait until we have answers to all the questions raised to make a substantial new investment in after-school programs that will help more teens avoid early childbearing and other consequential behaviors, including dropping out of school. Given the strong public support for more after-school programming, we should build on the substantial evidence we now have about effective models and commit to learning even more through well-designed evaluations.56 52. Hollister (2003); Walker (2001). 53. Walker (2001, p. 12). 54. Eccles and Gootman (2002). 55. Redd, Brooks, and McGarvey (2002). 56. Forum for Youth Investment (2002).

1269-04 CH04 04/18/03 17:24 Page 73

  

73

References Afterschool Alliance. 2001. Afterschool Alert Poll Report 4 (July/August). Washington: C.S. Mott Foundation and JCPenney Afterschool. Allen, Joseph P., and others. 1997. “Preventing Teen Pregnancy and Academic Failure: Experimental Evaluation of a Developmentally Based Approach.” Child Development 64 (4): 729– 42. Annie E. Casey Foundation. 2002. Children At Risk: State Trends 1990–2000. Kids Count Special Report. Baltimore, Md.: Population Reference Bureau. Brooks, Jennifer L., Elizabeth C. Hair, and Martha J. Zaslow. 2001. “Welfare Reform’s Impact on Adolescents: Early Warning Signs.” Child Trends Research Brief. Washington: Child Trends. Carrera, Kaye, and Susan Philliber. 2001. “Some Policy Implications from the Evaluation of a Successful Teen Pregnancy Program.” Paper prepared for the 2001 annual meeting of the Association for Public Policy Analysis and Management. Washington. Child Trends. 2002. Facts at a Glance. Washington. Cohen, Deborah A., and others. 2002. “When and Where Do Youths Have Sex? The Potential Role of Adult Supervision.” Pediatrics 110 (6). Corporation for National and Community Service. 2002. “National Service Agency Launches 3 New Initiatives to Enhance Citizenship.” Press release, September 7. Eccles, Jacquelynne S., and Jennifer Appleton Gootman, eds. 2002. Community Programs to Promote Youth Development. Washington: National Academy Press. Eccles, Jacquelynne S., and Janice Templeton. 2001. “Community-Based Programs for Youth: Lessons Learned from General Developmental Research and from Experimental and Quasi-Experimental Evaluations.” Paper prepared for the Urban Seminar Series, Urban Health Initiative, Harvard University. Fashola, Olatokunbo S. 1999. “Review of Extended-Day and After-School Programs and Their Effectiveness.” U.S. Department of Education: Office of Educational Research and Improvement. Forum for Youth Investment. 2002. Policy Commentary #1: After-School Research Meets After-School Policy. Washington: The Forum for Youth Investment. Gennetian, Lisa A., and others. 2002. How Welfare and Work Policies for Parents Affect Adolescents: A Synthesis of Research. New York: Manpower Demonstration Research Corporation. Grossman, Jean B., Karen Walker, and Rebecca Raley. 2001. “Challenges and Opportunities in After-School Programs: Lessons for Policymakers and Funders.” Philadelphia: Public/Private Ventures/Manpower Demonstration Research Corporation. Grossman, Jean B., and others. 2002. Multiple Choices after School: Findings from the Extended-Service Schools Initiative. Philadelphia: Public/Private Ventures. Hoffman, Saul D. 1998. “Teenage Childbearing Is Not So Bad After All . . . or Is It? A Review of the New Literature.” Family Planning Perspectives 30 (5): 236–39, 243. Hollister, Robinson. 2003. “The Growth in After-School Programs: Status, Issues, and Evaluation of Impacts.” Unpublished discussion paper. Brookings Roundtable on Children. Brookings. Hotz, Joseph, Susan McElroy, and Seth Sanders. 1997. “The Costs and Consequences of Teenage Childbearing for the Mothers and the Government.” In Kids Having Kids: The Economic Costs and Social Consequences of Teen Pregnancy, edited by Rebecca Maynard. Washington: Urban Institute Press.

1269-04 CH04 04/18/03 17:24 Page 74

74

    . 

Kalmuss, D. S., and P. B. Namerow. 1994. “Subsequent Childbearing among Teenage Mothers: The Determinants of Closely Spaced Second Birth.” Family Planning Perspectives 26 (4): 149–53, 159. Kirby, Douglas. 2001. Emerging Answers: Research Findings on Programs to Reduce Teen Pregnancy. Washington: National Campaign to Prevent Teen Pregnancy. ———. 2002. Preventing Teen Pregnancy: Youth Development and After-School Programs. ETR Publishing. Kirby, Douglas, Karin Coyle, and Jeffrey B. Gould. 2001. “Manifestations of Poverty and Birthrates among Young Teenagers in California Zip Code Areas.” Family Planning Perspectives 33 (2): 63–69. Lichter, Daniel T., and Deborah Roempke Graefe. 2001. “Finding a Mate? The Marital and Cohabitation Histories of Unwed Mothers.” In Out of Wedlock: Trends, Causes, and Consequences of Nonmarital Fertility, edited by Lawrence L. Wu and Barbara Wolfe. Russell Sage Foundation. Manlove, Jennifer, and others. 2002. “Preventing Teenage Pregnancy, Childbearing, and Sexually Transmitted Diseases: What the Research Shows.” Child Trends Research Brief: American Teens. Washington: Child Trends. Marshall, Will, and Isabel Sawhill. 2003. “Progressive Family Policy in the 21st Century.” In Public Policy and the Future of the Family, edited by Daniel Patrick Moynihan, Lee Rainwater, and Timothy Smeeding. Russell Sage Foundation (forthcoming). Maynard, Rebecca. 1997. Kids Having Kids: Economic Costs and Social Consequences of Teen Pregnancy. Washington: Urban Institute Press. Moore, K. A., and others. 1993. “Age at First Childbirth and Later Poverty.” Journal of Research on Adolescence 3 (4): 393–422. National Campaign to Prevent Teen Pregnancy. 2001. Halfway There: A Prescription for Continued Progress in Preventing Teen Pregnancy. Washington. ———. 2002a. Not Just Another Single Issue: Teen Pregnancy Prevention’s Link to Other Critical Social Issues. Washington. ———. 2002b. With One Voice 2002: America’s Adults and Teens Sound Off about Teen Pregnancy. Washington. National Commission on Service Learning. 2002. “Learning in Deed: The Power of Service-Learning for American Schools.” Newton, Mass. Newman, Robert P., Stephanie M. Smith, and Richard Murphy. 2000. “A Matter of Money: The Cost and Financing of Youth Development.” Youth Development: Issues, Challenges, and Directions. Fall 2000: 82–124. Philadelphia: Public/Private Ventures. Nicholson, Heather Johnson, Leticia T. Postrado, and Faedra Lazar Weiss. 1991. “Truth, Trust, and Technology: New Research on Preventing Adolescent Pregnancy.” Indianapolis: Girls Incorporated. Olsen, Davey. 2002. 12-Hour School Days? Why Government Should Leave Afterschool Arrangements to Parents. Washington: Cato Institute. Philliber, Susan, and others. 2002. “Preventing Pregnancy and Improving Health Care Access among Teenagers: An Evaluation of the Children’s Aid Society–Carrera Program.” Perspectives on Sexual and Reproductive Health 34 (5): 244–251. Redd, Zakia, Jennifer Brooks, and Ayelish McGarvey. 2002. “Educating America’s Youth: What Makes a Difference.” Child Trends Research Brief: American Teens 4 (August). Sylvester, Kathleen, and Kathy Reich. 2000. After-School Programs, Issues, and Ideas: A Guide for Journalists and Policymakers. Prepared for the David and Lucille Packard Foundation. Washington: Social Policy Action Network.

1269-04 CH04 04/18/03 17:24 Page 75

  

75

Teirney, Joseph P., Jean Baldwin Grossman, and Nancy L. Resch. 1995. Making a Difference: An Impact Study of Big Brothers/Big Sisters. Philadelphia: Public/Private Ventures. Thomas, Adam, and Isabel Sawhill. 2002. “For Richer or Poorer: Marriage as an Antipoverty Strategy.” Journal of Policy Analysis and Management 21 (3): 587–99. Urban Health Seminar Series on Children’s Health and Safety. 2001. “Urban Seminar on After-School Time.” Overview of the seminar held May 10–11, Harvard University. U.S. Department of Education. 2002. “21st Century Community Learning Centers: A Snapshot.” ———. 2003. “21st Century Community Learning Centers: Program Performance Report,” forthcoming. U.S. Department of Education and U.S. Department of Justice. 2000. “Working for Children and Families: Safe and Smart Afterschool Programs.” U.S. House of Representatives, Committee on Ways and Means. 2002. Strategies for Reducing Early Childbearing and the Growth of Single-Parent Families: Hearing before the Subcommittee on Human Resources. Testimony of Isabel Sawhill. 107th Cong., 11 April. Ventura, Stephanie J., and Christine A. Bachrach. 2000. “Nonmarital Childbearing in the United States, 1940–99.” National Vital Statistics Reports 48 (16). Washington: Centers for Disease Control and Prevention. Walker, Gary. 2001. “The Policy Climate for Early Adolescent Initiatives.” Public/Private Ventures Brief (January).

1269-05 CH05 04/18/03 17:24 Page 76

5 Strengthening Fragile Families IRWIN GARFINKEL SARA MCLANAHAN

AND

O

ut-of-wedlock childbearing increased dramatically in the United States during the last four decades of the twentieth century. In 1960 about 6 percent of all children were born outside of marriage; at the end of the century, the figure was 33 percent.1 The numbers were even higher among some racial and ethnic minorities. In 1999, 69 percent of African American children and 42 percent of Hispanic children were born to unmarried parents, compared with 22 percent of non-Hispanic white children. In response to these disturbing trends, Congress has made marriage a major issue in welfare reform. In 1996, the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) offered financial rewards to states that reduced nonmarital births. In 2002, as part of the reauthorization of PRWORA, the Bush administration proposed to spend up to $300 million dollars annually on demonstrations to promote “healthy marriages.” The administration also proposed to include “father involvement” and “child well-being” among the objectives of PRWORA.

1. Ventura and Bachrach (2000).

76

1269-05 CH05 04/18/03 17:24 Page 77

  

77

In this chapter we describe a demonstration designed to evaluate alternative strategies for promoting healthy marriages, father involvement, and child well-being among new unmarried parents. Whatever the success of pro-marriage policies, a substantial number of single parents will remain. Policies that address the needs of these single parents and their children may end up discouraging marriage. We discuss this dilemma and incorporate in our proposal a new strategy for resolving it. The demonstration would include new services as well as changes in existing policies. It would be evaluated by a random assignment experimental design. Although the proposed demonstration has many components and would be expensive if enacted as national policy, it is no panacea. It does nothing to prevent unintended pregnancies or early childbearing. (See chapter 4 for another proposal that does address this problem.) Still, we believe that the intervention described would lead to more stable unions, more father involvement, and better outcomes for children.

The Demonstration The demonstration would begin at birth and last for at least five years. It would consist of three components: providing a new assessment and referral service; reducing marriage penalties in existing programs; and providing new services aimed at strengthening parental relationships. Mothers and fathers would be treated alike in terms of services and policy changes. Before leaving the hospital, all new parents would be interviewed and assessed to determine their eligibility for the new program. If eligible, parents would be informed of the objectives of the experiment and asked to participate. Eligibility would be limited to unmarried parents who were having their first child; who had a joint income that was less than the median income of two-parent families with one child in their city; who still were in a romantic relationship with each other; and who both were willing to participate in the experiment. If the mother was interested in participating and the father could not be interviewed at the hospital, the couple would be eligible for participation as long as the father could be interviewed and agreed to participate within a week after the birth. Once the couple agreed to participate, they would be randomly assigned to one of five experimental groups or a control group.

Assessment and Referral Service Parents who participated in the assessment and referral service would receive a home visit within a week or two after the birth of their child. The assessment

1269-05 CH05 04/18/03 17:24 Page 78

78

    

would be conducted by a trained social worker and would include interviews with both mother and father. The assessment would cover parents’ human capital (education, employment, health) as well as their social capital (connections to family, friends, and community groups). Parents also would be assessed with respect to their eligibility for benefits and services such as the earned income tax credit (EITC), Temporary Assistance for Needy Families (TANF), food stamps, medical care, child care, substance abuse treatment, literacy and other education programs, employment services and job training, child support, and parenting classes. Parents who were eligible to participate in one or more of these programs but did not would receive information and instructions on how to apply for benefits or services. Except in cases where there was evidence of child abuse or neglect, the social workers would be instructed only to inform parents of the availability of relevant services and to make themselves available to help parents obtain services at any time over the next five years. Social workers would initiate no further direct contact with parents. They would explain to the parents, however, that the agency they worked for would send a birthday card each year, along with a reminder of the wide range of services for which they were potentially eligible.

Marriage-Friendly Policies Parents who participate in the marriage-friendly policy component would be eligible for an array of income support services that, compared with current policies, would be substantially more marriage friendly, though not quite neutral; provide more economic security; and improve human capital.      . Categorical limitations on two-parent families—whether created by law or practice—would be eliminated for all parents participating in this component of the demonstration. In addition to making TANF more marriage friendly, the TANF benefit for the second adult in a two-parent family would be equal to at least 50 percent of the benefit for a single adult in a one-parent family. Current benefits add less than 25 percent for a second adult. The benefit increase would allow parents who live together to capture more of the economies of scale that accrue from living together than under the current structure. The new benefit level would continue for parents who elected to marry after the birth of their child. Benefits would be contingent on parents’ working or engaging in some other socially productive activity (besides caring for their new infant). During the first year of the child’s life, the hours of work (or training or education)

1269-05 CH05 04/18/03 17:24 Page 79

  

79

required of two-parent families would be substantially less than twice the hours required of a single parent. For example, if thirty hours a week was required of a single mother, the requirement for two parents might be as little as forty and no more than fifty hours. Services would be provided to fathers as well as mothers. The services for fathers, like those for mothers in TANF, would be geared primarily toward obtaining employment. However, in cases where the mother or father demonstrates the potential to benefit from further education and training, TANF would provide such support. In addition to removing restrictions and making TANF more marriage neutral, the intervention would do the same for the EITC, housing subsidies, medical care, and daycare services. In addition, it would guarantee these services to all economically eligible participants. Currently only a minority of economically eligible families receive either housing or child care subsidies, and the waiting lists for both benefits are long.   . All unwed parents would be required to establish legal paternity for their child and to obtain a child support order. Child support obligations would be based on a percentage of the father’s actual income and would rise and fall automatically with changes in income. Child support payments would go directly to the mother regardless of whether she received other public benefits. Fathers who lived with their child would be exempt from making child support payments. Both parents would be fully informed of the nonresident parent’s potential child support obligation in the event of a separation. Finally, to increase the economic security of children whose parents elect to live apart and fathers’ incentives to pay child support and stay involved with their child, a government-financed child support benefit contingent on the establishment of a child support order and tied to fathers’ payments would be established.

Relationship Programs Parents participating in the relationship component of the demonstration would be offered a new service to help improve the quality of their relationship. Parents would receive services based on one of three models. The first model emphasizes communication skills. It is primarily didactic and provides new parents with an opportunity to practice communication skills through role-playing and discussion. Parents participating in this program would attend workshops with other new parents to learn about and practice communication skills and ways to resolve disagreements. The curriculum for these workshops is based on a set of generic tasks, and all participants practice the

1269-05 CH05 04/18/03 17:24 Page 80

80

    

same skills. An example of this approach is the PREP program developed by Howard Markman and his colleagues.2 A second model aims at increasing intimacy and helping couples resolve their personal issues. In this program, parents would meet together in small groups, under the guidance of a clinician, to discuss issues that have arisen around the birth of the first child. The content of these programs is partly predetermined, to address issues that are common to all new parents, and partly reactive, to address issues brought up by the parents in the group. An example of this approach is the Parents and Partners (PAP) program developed by Cowan and Cowan.3 A third model provides mentoring to new parents. The idea behind this approach is to match new parents with older, more experienced couples who have a strong, stable relationship and who are willing to serve as role models and confidants to new parents. In this model, new parents and mentors meet on a regular basis to discuss issues related to parenting and partnering. This model currently is being used in some churches.

Experimental Design and Costs The demonstrations would be conducted in four or five cities. In each city, parents participating in the program would be randomly assigned to one of five treatment groups (sample size 1,000 each) or to a control group (sample size 2,000) as described below. Both parents would be eligible for treatment for a period of up to five years after the birth of their child, and both would be interviewed at their child’s birth and again when the child was one, three, and five years old. Parents in the first three treatment groups would receive the relationship skills component only. Group 1 would receive PREP, group 2 would receive PAP, and group 3 would receive the mentor program. Relationship interventions recently have gained widespread attention, and it is essential to examine the effect of these interventions in the absence of other changes. Couples in experimental group 4 would receive both the in-home assessment and the policy changes, and couples in experimental group 5 would receive the assessment, the policy changes, and PREP. Group 4 would allow us to assess the effects of making existing services and benefits more marriage friendly, and group 5 would allow us to determine the effectiveness of one 2. Stanley, Blumberg, and Markman (1999). 3. Cowan, Powell, and Cowan (1998).

1269-05 CH05 04/18/03 17:24 Page 81

  

81

of the relationship skills interventions in combination with more marriagefriendly services and benefits. We selected PREP because it has received the most attention and support among policymakers. The total costs for the demonstration include both intervention and evaluation costs. Each is discussed below. We estimate that the total cost over five years would be between $150 and $250 million. These estimates are discussed in the appendix to this chapter.

The Rationale Our program is based on a model that assumes that child development depends on parenting quality and the structure and quality of the parents’ relationship; parents’ human capital; and macro factors such as policies, culture, and labor market conditions. Both macro factors and parents’ human capital affect relationship status and quality of parenting.

Improving Relationship Quality A basic assumption of our model is that children benefit from living in a healthy, two-parent family and that improving the quality of the parents’ relationship increases child well-being. Numerous studies have shown that children who grow up with both biological parents are more successful in adulthood than children who grow up with only one parent.4 While part of the advantage associated with having two parents comes from having greater economic security, another part comes from the quality of the parents’ relationship. Parents with high-quality relationships—low conflict and high mutual support—are more likely to marry and stay married than parents with low-quality relationships. Both relationship quality and union stability are associated with good parenting and better child outcomes.5 We suspect that a substantial proportion of new unmarried parents would benefit from and willingly participate in an intervention designed to strengthen their relationship and promote stable marriage, especially if the service is offered around the time of the birth of their child. Our thinking is based on evidence from a new survey of unmarried parents—the Fragile Families and Child Well-Being Study—in which mothers and fathers were interviewed shortly after the birth of their child and followed for four years.6 4. McLanahan and Sandefur (1994); Waite (1995). 5. Cummings and Davies (1994); Emery (1999). 6. Fragile Families (2001).

1269-05 CH05 04/18/03 17:24 Page 82

    

82

Figure 5-1. Parents’ Commitment at Time of Child’s Birth No relationship 9% Friends 8% Cohabiting 50%

Visiting 33%

Source: Fragile Families (2001).

One of the most striking findings from the study is the high level of motivation among new unmarried parents. At the time of their child’s birth, half of unmarried mothers are living with the father of the child and another third are romantically involved with the fathers but living apart (we call the latter “visiting relationships”). Eight percent of parents are “just friends” and only 9 percent have “little or no contact.” (See figure 5-1.) Table 5-1 shows that the majority of unwed parents are optimistic about their future together. Seventy-three percent of mothers believe that their chances of marrying the father are 50 percent or higher. Almost two-thirds of mothers “agree” or “strongly agree” with the statement “It is better for children if their parents are married.” Most fathers are highly involved during the pregnancy and around the time of birth. According to the mothers’ responses, four of five fathers provided some financial support during the pregnancy, 84 percent will have their name on the birth certificate, and 79 percent of the children will take the father’s surname. Most fathers say they want to help raise their child, and the overwhelming majority of mothers say they want the fathers to be involved.7 7. Fragile Families (2001).

1269-05 CH05 04/18/03 17:24 Page 83

  

83

Table 5-1. Unmarried Parents’ Views about Marriage and Commitment Percent

View Marriage Chances of marriage 50-50 or better Marriage is better than cohabiting Marriage is better for children Commitment Provided financial support during the pregnancy Plans to sign birth certificate Plans for child to take his surname Wants to help raise child Mother wants father involved

Mothers

Fathers

73 51 64

88 60 75 80 84 79 94 94

Source: Fragile Families (2001).

In sum, at the time their first child is born, the vast majority of unmarried parents are committed to each other and to their child. Most mothers and fathers have high hopes about their future together and view marriage as a positive institution that benefits children. Clearly, these parents are likely to respond positively to programs and policies that promote marriage, which is good news for policymakers who favor this strategy. Unfortunately, fewer than 10 percent of new unmarried parents have actually married by the time their child reaches his or her first birthday, and 41 percent are no longer romantically involved—twice the percentage at birth. The reasons for the lack of follow-through are discussed below.

Building Human Capital A second assumption in our model is that parents’ human capital (education, work experience, and physical and mental health) affects the quality of their relationship and their parenting. This assumption is based on a large body of empirical research that shows that education and employment are positively associated with marriage and negatively associated with divorce. A second set of studies shows that parents with high human capital engage in more effective parenting behavior than parents with low human capital. Parents themselves recognize the importance of a steady income to maintaining a stable union. When asked to name the most important components of a successful marriage, the vast majority of parents say that the “father having

1269-05 CH05 04/18/03 17:24 Page 84

84

    

a steady job” is essential.8 In-depth interviews with unmarried parents confirm this perception and also suggest that many parents view marriage as something that comes after they have achieved a middle-class life-style rather than as something they do together in order to move upward on the economic ladder.9 Unfortunately, most unmarried parents have very low human capital (table 5-2). Although nearly all fathers worked during the year, nearly three in ten were out of work in the week before their baby was born. In addition, the educational attainment of both parents was low: 37 percent of mothers and 34 percent of fathers lacked a high school degree, and less than one-third of parents had any education beyond high school. Although a majority of unmarried parents were in fairly good health, some engaged in unhealthy behaviors. Three percent of mothers and 5 percent of fathers reported that a drug or alcohol problem interfered with their work or personal relationships in the past year. Drug and alcohol problems are likely to be underreported, so we assume that the true prevalence of substance abuse is higher. Finally, 38 percent of unwed fathers had been incarcerated, which suggests that a substantial proportion had (or had had) life-styles that could be harmful to their children.10 Given this profile, we would expect the initial in-home assessment to identify a substantial number of parents who would benefit from TANF, EITC, Medicaid, housing subsidies, education, job training, substance abuse treatment, and other services designed to improve their human capital.

Making Welfare and Child Support Policies More Marriage Friendly Finally, our model indicates that public policies, along with culture and labor market conditions, affect relationship status. While culture and the economy are difficult to change, policies are more malleable. To the extent that welfare policies or practices favor one-parent families over two-parent families, they discourage marriage and cohabitation and push biological fathers out of the picture. Although many state TANF programs appear to have reduced or eliminated restrictions on two-parent families, others still retain them. Many states and localities also give preference to one-parent families in allocating scarce child care and housing subsidies. 8. Fragile Families (2001). 9. Gibson, Edin, and McLanahan (2002). 10. Fragile Families (2001).

1269-05 CH05 04/18/03 17:24 Page 85

  

85

Table 5-2. Parents’ Human Capital and Ability to Support a Child Percent

Characteristic Worked in past year Worked in past week Education Less than high school High school only Some college College or higher Age Under 20 20–24 25–29 30 and older Very good or excellent health Drug or alcohol problems Incarceration

Mothers

Fathers

84 ...

98 78

37 32 27 4

34 40 22 4

21 43 19 17 65 3 ...

10 34 26 30 69 5 38

Source: Fragile Families (2001).

Categorical restrictions are not the only way in which welfare policies favor single parents. Because welfare is income tested, it creates an incentive for fathers with earnings and mothers without earnings to live apart (or to feign living apart). Moreover, the current benefit schedule captures any economies of scale that accrue from marriage, which reduces the incentive to marry. Research indicates that higher welfare benefits lead to more nonmarital births and fewer marriages, although the effects are small.11 Research also indicates that extending benefits to two-parent families would reduce divorce and increase marriage and cohabitation.12 We also would expect benefits such as those for child care, housing, and health care to increase union stability and family formation if extended to two-parent families. Making public support programs in the United States more marriage friendly would be relatively costly because a relatively generous benefit package already is available to female-headed families. The total cost of the benefits is kept low by explicitly limiting them to single-parent families, by 11. Moffitt (1998). 12. Carlson and others (2002); Mincy and Dupree (2001); Knox, Miller, and Gennetian (2000); Harknett and Gennetian (2001).

1269-05 CH05 04/18/03 17:24 Page 86

86

    

giving single-parent families priority for scarce services, and by structuring the programs so that the government rather than the parents captures any economies of scale. The virtue of designing an experiment in which these support programs are made more marriage friendly is that this makes it possible to determine the effects, both positive and negative, without spending a lot of money. Furthermore, although previous research has shown that the effects on marriage of Aid to Families with Dependent Children (AFDC) and TANF are small, there is no research on the magnitude of the total effects of all the marriage penalties in our income support programs. Taken together, Medicaid, TANF, housing subsidies, child care subsidies, and the EITC loom very large in the lives of the poor. If each program reduced marriage by only 4 percent, the effect would be very small and very difficult to detect. But the sum of all the individual effects could be substantial. Therefore an experiment to test the effect of making public support programs more marriage friendly is potentially quite important. In addition to reducing the marriage penalties in various social welfare programs, changing child support policies may be necessary. Child support affects family formation by increasing the costs of children to their fathers. Stricter child support enforcement should encourage men to avoid fathering children if they do not want them and give men who father children a greater incentive to live with their children. Previous research supports these predictions. States with stricter child support enforcement have less divorce and lower rates of nonmarital childbearing than states with weaker enforcement.13 Marriage and cohabitation among fragile families can be encouraged by increasing the capabilities of parents and by reducing marriage and cohabitation penalties in our transfer and tax policies. No matter how successful such policies may be, however, a substantial proportion of unwed parents will live apart. Welfare and child support policies that address the needs of these parents have to be structured in ways that promote father involvement and the quality of father-child relationships. If child support obligations are grossly inconsistent with fathers’ ability to pay, they may drive fathers away and discourage their involvement. Under the current child support system, poor fathers are routinely required to pay a much higher proportion of their income than middle- and upper-income fathers, and many are required to pay unreasonable amounts of arrearages. These unrealistic arrearages arise because child support agencies or courts base orders not on fathers’ actual earnings, but on their presumptive earn13. Nixon (1997); Case (1998); Garfinkel and others (2002).

1269-05 CH05 04/18/03 17:24 Page 87

  

87

ings.14 Some fathers even are required to pay back the mother’s welfare or Medicaid costs. Many fathers build up huge arrearages when they become unemployed or incarcerated. Such onerous child support obligations are rarely paid in full, but they do prompt fathers to avoid legitimate work because their wages then can be easily intercepted, and they breed resentment on the part of fathers and mothers toward the system and perhaps each other. Imprisonment for nonpayment of support exacerbates this negative dynamic. Given the low earning capacity of most unwed fathers, such practices are not likely to be effective and may have unintended negative consequences. Finally, there are several reasons for enacting a public child support benefit that is tied to fathers’ payments and phased out at higher payment levels (much as the EITC is phased out with higher earnings). First, such a benefit would reduce the poverty and insecurity of single mothers and their children. Second, it would increase nonresident fathers’ support of and involvement with their children. The benefit would increase government expenditures, and more important, because it would be limited to children who live apart from a parent, it would increase their parents’ incentives for living apart. In combination with all of the other program changes that eliminate or at least substantially reduce marriage penalties, however, we believe that the benefits of the public child support benefit would outweigh the costs. In our view it makes sense to limit favoritism for single-parent families to child support enforcement since the latter is the one service that is inherently contingent on single parenthood.

Criticisms and Political Feasibility Several possible criticisms may be raised about our proposal. One is that some fathers may not be good candidates for healthy marriage. A large proportion of unwed parents have children by different partners. Indeed, more than half the births in the Fragile Families study were to couples with children from other relationships. In these families, a question arises: Who should marry whom?15 The multiple-partner complication is a major reason for limiting participation in the demonstration to first-time parents. If the intervention is successful, the prevalence of multiple-partner births will diminish in the future.

14. Garfinkel and others (1998); Waller and Plotnick (2001). 15. Mincy (2002).

1269-05 CH05 04/18/03 17:24 Page 88

88

    

Some fathers are violent and abuse their partners. About 8 percent of noncohabiting mothers in the Fragile Families survey report being seriously hurt by the fathers of their children in the first year following the birth, and 14 percent report being hit or slapped by the father. Arguably none of these fathers are good candidates for a healthy marriage. Yet nearly all unwed mothers say they want the father involved with raising the child, including mothers who report that their partners are violent. Thus, providing services to these couples is likely to have beneficial effects in terms of improving the parents’ relationships and ability to cooperate in raising their child. In instances in which the father is violent and the mother does not want him involved with the child, she may refuse to participate in the program. More generally, whatever makes a man an undesirable marriage partner is likely to work against his participation. Consequently, requiring both mother and father to participate is likely to exclude from the program most fathers who are incapable of building a healthy marriage. The topic of what services should be provided to these men and their families also should be addressed. Another objection to the demonstration is that the government should not be promoting marriage, even healthy marriages. Some believe that by promoting marriage, government implicitly stigmatizes single parenthood. We disagree with this argument. Promoting healthy marriages acknowledges that some marriages are unhealthy and teaches society (and parents) to distinguish between relationships that are good for people and those that are not. A related objection is that government should not favor marriage over cohabitation. With respect to government benefits and taxes, we share this view. But insofar as marriage connotes a long-term commitment on the part of both partners, we think it is superior to cohabitation in terms of promoting the welfare of adults and children. Long-term commitments have positive effects on relationships between parents and between parents and children. Whether such commitments benefit from being blessed by a religion or the state is not clear. We suspect that public acknowledgement of a commitment helps. Another variation of this objection is that government should not be in the business of promoting values. Whether people live together or get married is their business. We do not share this view. Societies communicate norms and values in many ways, including modern democratic societies in which governments express values through laws and social programs. Welfare programs reinforce the value that no one should starve or go homeless because of poverty. Child support enforcement emphasizes the value of parents supporting their children. We try to teach healthy behaviors in our schools, and we tax harmful behaviors like smoking. We even outlaw some unhealthy behaviors. When we disagree with a policy, our disagreement usually arises

1269-05 CH05 04/18/03 17:24 Page 89

  

89

not because values are involved but because we disagree with the particular value being expressed. A completely different kind of objection has to do with the political feasibility of our proposal. Some critics have urged us to abandon the component of our proposal that seeks to make income support programs marriage neutral because of the huge fiscal cost associated with achieving that aim. We agree that eliminating marriage penalties would be extremely costly and politically infeasible at this time. However, we view these costs as an argument for conducting an experiment to find out whether large benefits may be involved as well. Small benefits would reinforce the status quo—large benefits, in contrast, would increase political feasibility. None of the components in this experiment, if proposed as national policy, are currently politically feasible. The relationship programs are relatively cheap, but liberals will object to them as a substitute for material support for low-income families. Neither liberals nor conservatives are opposed, in principle, to making policies more marriage friendly. Nevertheless, they are sharply divided over how to accomplish this objective. Fiscal conservatives would level the playing field by lowering benefits to single mothers, whereas liberals would support expanding benefits to a large proportion of the population. Fearing the power of the fiscal conservatives and the potential loss of benefits to single mothers, liberals are likely to minimize the importance of making policies more marriage friendly. Given the debate and the uncertainties, we have chosen to propose an experiment and to carefully evaluate its effects. The experiment’s results, in turn, are likely to have some effect on political feasibility. If any of the components prove to be successful, their political feasibility will be enhanced. Conversely, if the components prove unsuccessful, their feasibility will be reduced.

APPENDIX 5A Costs of the Demonstration Program for Fragile Families Intervention Costs The costs of the components will depend on both take-up rates and the cost of each component. We develop a best and a maximum cost estimate. The maximum estimate assumes very high take-up rates for all elements of all components and makes no attempt to subtract existing expenditures to

1269-05 CH05 04/18/03 17:24 Page 90

90

    

obtain net rather than gross cost estimates. The best estimate is based on our best educated guesses about actual take-up rates and the levels of existing expenditures. Each intervention is discussed separately.   ⁽   ⁾ . Take-up rates are assumed to be 100 percent. We assume PREP costs per couple to be $500 per treatment and two treatments per couple to be given, for a total cost of $1 million. Although a better guess is that take-up rates will be 25 percent, we do not calculate a lower bound estimate for this component because the cost of the intervention already is so low and if the participation rate turns out to be only 25 percent, it would be appropriate to increase the sample size so that this component could be assessed more effectively. Thus the reduction in intervention costs would be offset by an increase in evaluation costs.   ⁽ ⁾. Take-up rates for PAP also are assumed to be 100 percent. We assume the cost per couple to be $500 per treatment and six treatments per couple to be given, for a total cost of $3,000. The maximum cost for this intervention therefore is $3 million. Although take-up rates of 25 percent probably are more realistic, as explained above, lower intervention costs could be offset by higher evaluation costs.   ⁽   ⁾ . Take-up rates are assumed to be 100 percent. The costs are assumed to be $3,000 per couple, for a total maximum cost of $3 million. We assume that the treatment lasts for six months. Again, although take-up rates of 25 percent probably are more realistic, as explained above, lower intervention costs could be offset by higher evaluation costs. -     ⁽     ⁾ . For our maximum estimate, we assume that 100 percent of mothers and fathers who elect to participate in this component will request an initial assessment and that 100 percent of all couples will make a request for additional help with referrals and applications for programs in each succeeding year. We also assume that each assessment of mothers and fathers will require eight hours to complete (including travel time, missed appointments, and so forth) and that subsequent requests for help will require sixteen hours per request. Providing these services to 1,000 couples will require a total of 8,000 professional social worker hours at baseline and 16,000 hours for each of the succeeding four years, which translates into four and eight full-time workers respectively. We also assume that before and during baseline assessments, learning about and negotiating access to services will require another 8,000 hours, so that eight full-time workers will be needed for all five years. Assuming that social workers are paid $60,000 per year plus 30 percent fringe benefits yields a professional cost of $624,000. Assuming overhead of 100 per-

1269-05 CH05 04/18/03 17:24 Page 91

  

91

cent to cover all other costs, such as rent, administrative assistance, supervision, travel, and so forth, the maximum total cost of this portion of the intervention equals nearly $1.25 million a year, or $6.25 million over five years. Take-up rates for all services except child care are assumed to be 100 percent each year. Child care take-up rates are assumed to be zero during year 1 and 100 percent in subsequent years. We assume the average cost per eligible family for the more marriage friendly EITC, housing subsidy, and TANF to be $5,000 per year. Child care is assumed to cost $8,000 per year and Medicaid $5,000 per year. The total maximum cost of this component of the demonstration therefore is $82 million. Take-up rates for other services will be 5 percent per adult per year for drug abuse and 5 percent for training, with the cost of both averaging $10,000, for a maximum five-year cost of $10 million. Adding all of the elements of assessment and policy change yields a maximum total cost estimate of nearly $100 million. Our best estimate is that after taking account of existing expenditures and much lower take-up rates the actual costs would be more like $50 million. - ,  ,   ⁽    ⁾. The cost of this intervention is equal to the sum of the costs of interventions 1 and 4, or a total maximum cost of $104 million and a bestguess estimate of $54 million.

Evaluation Costs Evaluation costs include those for both data collection and design and analysis. Data collection costs include baseline interviews of mothers and fathers at the hospital and telephone follow-up interviews with both parents when the child is one, three, and five years old. The third- and fifth-year follow-ups also would include an in-home interview for mother and child assessments. We assume the four father interviews to cost $500 each and the mother interviews to cost $300, $500, $800, and $1,000 respectively. Assuming 7,000 interviews each for mothers and fathers yields a total of $32.2 million. The costs of designing the questionnaire, overseeing the data collection subcontractor, and analyzing the data are estimated to be another $6 million. Total evaluation costs therefore are approximately $38 million.

References Carlson, Marcia, and others. 2002. “The Effects of Welfare and Child Support Policies on Union Formation.” Paper presented at the annual meeting of the Population Association of America, Atlanta, Georgia.

1269-05 CH05 04/18/03 17:24 Page 92

92

    

Case, Anne. 1998. “The Effects of Stronger Child Support Enforcement on Nonmarital Fertility.” In Fathers under Fire: The Revolution in Child Support Enforcement, edited by I. Garfinkel and others, 67–92. Russell Sage. Cowan, Philip A., D. Powell, and C. P. Cowan. 1998. “Parenting Interventions: A Family Systems Perspective.” In Handbook of Child Psychology, vol. 5, edited by W. Damon. John Wiley. Cummings, E. Mark, and P. Davies. 1994. Children and Marital Conflict: The Impact of Family Dispute and Resolution. New York: Guilford Press. Emery, Robert E. 1999. Marriage, Divorce, and Children’s Adjustment. Thousand Oaks, Calif.: Sage Publications. The Fragile Families and Child Well-Being Study Baseline Report: The National Report. 2001. Center for Research in Child Well-Being, Princeton University. Garfinkel, Irwin, and others, eds. 1998. Fathers under Fire: The Revolution in Child Support Enforcement. Russell Sage. Garfinkel, Irwin, and others. 2003. “The Roles of Child Support Enforcement and Welfare in Nonmarital Childbearing.” Journal of Population Economics (forthcoming). Gibson, Christina, K. Edin, and S. McLanahan. 2002.“High Hopes but Even Higher Expectations: The Retreat from Marriage among Low-Income Couples.” Paper presented at the annual meeting of the Population Association of America, Atlanta, Georgia. May. Harknett, Kristin, and L. A. Gennetian. 2001. “How an Earnings Supplement Can Affect the Marital Behaviour of Welfare Recipients: Evidence from the Self-Sufficiency Project.” Working Paper. Canada: Social Research and Demonstration Corporation. Knox, Virginia, C. Miller, and L. A. Gennetian. 2000. “Reforming Welfare and Rewarding Work: Final Report on the Minnesota Family Investment Program.” New York: Manpower Demonstration Research Corporation. McLanahan, Sara, and G. Sandefur. 1994. Growing Up with a Single Parent: What Hurts, What Helps. Harvard University Press. Mincy, Ronald B. 2002. “Who Should Marry Whom? Multiple Partner Fertility among New Parents.” CRCW Working Paper 2002–03. Princeton University. Mincy, Ronald B., and A. T. Dupree. 2001. “Welfare, Child Support, and Family Formation.” Children and Youth Services Review 23 (6/7): 577–601. Moffitt, Robert A. 1998. “The Effect of Welfare on Marriage and Fertility.” In Welfare, the Family, and Reproductive Behavior, edited by R. A. Moffitt, 50–97. Washington: National Academy Press. Nixon, Lucia. 1997. “The Effect of Child Support Enforcement on Marital Dissolution.” Journal of Human Resources 32 (1): 159–81. Stanley, Scott M., S. L. Blumberg, and H. J. Markman. 1999. “Helping Couples Fight for Their Marriages: The PREP Approach.” In Preventive Approaches in Couples’ Therapy, edited by R. Berger and M. T. Hannah, 279–393. Philadelphia: Brunner/Mazel. Waite, Linda J. 1995. “Does Marriage Matter?” Demography 32 (November): 483–507. Waller, Maureen R., and R. Plotnick. 2001. “Effective Child Support Policy for LowIncome Families: Evidence from Street-Level Research.” Journal of Public Policy Analysis and Management 20 (1): 89–110. Ventura, Stephanie J., and C. A. Bachrach. 2000. Nonmarital Childbearing in the United States, 1940–1999. National Vital Statistics Reports 48 (16). Hyattsville, Md.: National Center for Health Statistics.

1269-06 CH06 04/18/03 17:25 Page 93

6 Family Policy and Academic Achievement D AV I D J . A R M O R

F

ew social problems have received more attention than the low academic achievement of children living in poverty. Indeed, a major goal of social policy during the past thirty years or so has been to alleviate the problem, directly or indirectly. Compensatory education policies have tried to improve poor children’s academic skills directly through school and preschool programs, while social welfare policies have tried to improve their general development indirectly through economic support of families. Unfortunately, most compensatory education programs, such as Title 1 and Head Start, have done little to improve the academic skills of poor children, who are disproportionately African American and Hispanic, although some Head Start programs have improved motivation and reduced behavior problems.1 Traditional welfare policies have not fared much better; in fact, some social critics believe that instead of restoring family structure, welfare policies promoted single-parent families and teenage pregnancy.2 Most compensatory education programs failed because 1. Most notably the Perry Preschool Project; see Schweinhart (1993). 2. The best-known criticism is in Murray (1984).

93

1269-06 CH06 04/18/03 17:25 Page 94

 . 

94

they assume a model of intellectual development that today does not appear to be viable, as discussed below. Traditional welfare policies failed to improve children’s academic skills because they encouraged certain family characteristics that actually hamper intellectual development, such as being headed by a young, uneducated single mother with many children. This chapter proposes a national whole family campaign, a major goal of which is to promote the kind of family environments that are associated with higher academic achievement and intelligence levels. The campaign would be modeled on national campaigns aimed at reducing risk factors for various adverse outcomes, such as the surgeon general’s campaign against smoking or the National Campaign to Prevent Teen Pregnancy.3 The whole family campaign would aim to improve children’s intellectual functioning by affecting their family environments. A major rationale for the whole family policy is recent evidence that certain family characteristics and activities—such as marital status, teenage pregnancy, family size, income, nutrition, and parenting, including instruction and nurturing behaviors—have important effects on the intellectual development of infants and young children.4 Some psychologists have argued that intelligence is determined largely by heredity and that this explains the failure of compensatory education programs.5 This chapter puts greater emphasis on family risk factors during early childhood. Traditional compensatory education programs fail not because intellectual ability is inherited, but because intelligence is affected by family factors that exert their greatest influence during early childhood, and because it is very difficult to change a child’s intelligence once he or she reaches school or even preschool age.

A Whole Family Policy Proposal There are three major strategies for modifying children’s intelligence. The first strategy employs special programs for low-IQ children during the regular school and preschool years. The second strategy involves very early childhood interventions that offer intensive instruction and other services from infancy to preschool. The third strategy, called the whole family approach, aims to change all or most of the environmental risk factors simultaneously. 3. See National Campaign to Prevent Teen Pregnancy (2002). 4. Armor (2003). 5. See Herrnstein and Murray (1994).

1269-06 CH06 04/18/03 17:25 Page 95

    

95

The evidence on compensatory and regular preschool programs generally indicates weak and inconsistent effects on IQ and achievement, particularly when their effects are compared with the impact of family characteristics.6 Reducing class size and hiring teachers with greater mastery of the subjects they teach seem to be the most promising of the school approaches, but their impact on IQ and academic achievement is quite small and often inconsistent. These findings are consistent with the proposition that intelligence is most malleable at very young ages and that malleability diminishes substantially after the first year or two of elementary school, a topic discussed in more detail below. The promising results of very early intervention projects like the Abecedarian project and Early Head Start, on the other hand, are consistent with the idea that intelligence is most malleable at very young ages.7 These programs provide intensive educational and family services for infants and toddlers, and some Early Head Start programs begin during the mother’s pregnancy. Less intensive interventions (such as home visits) or interventions for older children (such as Head Start) have rarely had significant effects on IQ. The downside of these very early intervention projects is cost, approximately double the annual cost of preschool projects like Head Start, and potential practical problems for parents having their infants and toddlers away from home for long periods of time (some are full-day programs that run twelve months a year). Nonetheless, the Early Head Start program may be the best intervention strategy for high-risk children whose parents are unable to provide them with a family environment that is conducive to their intellectual development. Whole family approaches may have the greatest potential to improve children’s IQs. The purpose of a whole family approach is to change the family environment by promoting those characteristics that improve children’s intellectual functioning: for example, being raised in a two-parent family by parents with adequate income and good parenting skills. There is no comprehensive whole family policy in effect today, but some of the profamily initiatives adopted by states under welfare reform resemble this approach. The family policy approaches embodied in the 1996 welfare reform act are changing the traditional reward structure so that two-parent families that work get greater rewards than single mothers who do not. In addition to welfare payments, some states have started other profamily policies that encourage marriage, involve nonresident fathers in child rearing, eliminate benefits for children born after a mother starts welfare, discourage teenage pregnancy, and require teen mothers to complete school. 6. Finn and Achilles (1999); McKey and others (1985). 7. Center for the Future of Children (1995 and 1999); Mathematica (2002).

1269-06 CH06 04/18/03 17:25 Page 96

 . 

96

Research on specific effects of whole family policies is just beginning, and there is limited evidence on how effective the policies might be. It may not be coincidental, however, that the long-term decline of two-parent families among African Americans finally halted after 1996. For the first time since these statistics have been gathered, the rate of two-parent black families actually increased by 4 percentage points between 1996 and 2000 (see figure 6-1). An “ideal” whole family approach would try to improve all of the family risk factors that have an impact on children’s intelligence and academic achievement. It would begin with young people before they became parents, which would mean targeting teenagers or even the parents of teenagers. The program would first encourage young people to get as much education as possible, completing high school at a minimum. Other major goals for prospective parents would be to delay childbearing until all education is completed and to marry before having children. Higher marriage rates should result in greater income, although supplemental income from welfare still may be required for low-income families. Additional goals would be to ensure that the parents stay married, that both parents (even if not married) are involved in raising their children, and to discourage having large numbers of children if income is low. Finally, the program should offer training and education in parenting skills, including the topics of nutrition (for example, breast feeding), cognitive stimulation or instruction (for example, reading

Figure 6-1. Percent of Children under Eighteen in Two-Parent Families Percent White

90 80 70

Hispanic

60

Black

50 40 30 20 10 0

1960

1965

1970

1975

Source: U.S. Bureau of the Census (2000).

1980

1985

1990

1995

2000

1269-06 CH06 04/18/03 17:25 Page 97

    

97

and related educational activities), and emotional support or nurturing (for example, openly displaying love and affection, praising accomplishments, and avoiding excessive physical punishment). At this time there is no single program that corresponds to the ideal whole family approach just described. While the primary goal of welfare reform is to move families off welfare and into jobs that ultimately will allow for economic self-sufficiency, the new welfare legislation also provides incentives for various profamily initiatives. Under welfare reform, many states are developing aspects of the whole family approach. States such as Arkansas, Delaware, Florida, Minnesota, New Jersey, and Oklahoma have begun various profamily policies aimed at increasing the marriage rate, maintaining two-parent families, reducing family size, and reducing the number of children born out of wedlock.8 While these state initiatives are commendable, they do not constitute a national whole family policy. Few states have all of the components of a whole family approach, and many states have very few components. While welfare reform offers opportunities to reach disadvantaged children and hopefully to improve their environments, it has two limitations as the primary focus of a whole family policy. First, most new welfare initiatives do not challenge the culture that has led to a worsening of family environments. Welfare reform attempts to fix problems that the culture has created, rather than preventing problems by challenging the culture and values that create them. Second, welfare reform does little for families who are not poor but whose children still face many risks related to family structure, family size, and parenting skills. Few of these families receive welfare benefits, and therefore few would gain from whole family policies imbedded in welfare reform. I propose taking all the components that have worked reasonably well in one or more states and combining them in a model program that could be adopted as a comprehensive prototype for a whole family policy. I propose then mounting a national education and prevention program to promote the whole family concept similar to the campaigns waged against behaviors that create other adverse social and health conditions, such as the National Campaign to Prevent Teen Pregnancy and the surgeon general’s campaign against smoking. In the case of the surgeon general’s campaign, significant progress was made in reducing smoking by publicizing its serious consequences, such as lung cancer and other diseases. Similarly, the campaign to prevent teen pregnancy coupled with welfare reform appears to have contributed to reductions in teen pregnancy and birth rates.9 8. A recent discussion of state initiatives is in Fein, London, and Mauldon (2002). 9. National Campaign to Prevent Teen Pregnancy (2001).

1269-06 CH06 04/18/03 17:25 Page 98

98

 . 

A key feature of the whole family initiative would be its potential to improve children’s intelligence and academic achievement. Although family environment can affect many cognitive and behavioral outcomes, adverse impacts on intelligence—especially for young children—might generate greater motivation for change than other outcomes, given the links between intelligence, educational attainment, and long-term occupational success. In addition to a national campaign, perhaps the best way to reach highand moderate-risk families would be through education programs in junior high and high schools. Most secondary schools have “health” classes that often are used for disseminating information about reproduction, sexually transmitted diseases, and methods of preventing pregnancy and disease. These classes could easily incorporate the whole family concept, emphasizing how each of the family risk factors has adverse effects on children’s intelligence and other outcomes. By disseminating information to teenagers about the consequences of their own decisions for their children, such programs have the potential to reach most future parents, hopefully convincing many of them that by adopting certain behaviors they can improve their children’s life chances. For young parents, high schools and community colleges also could develop classes that offer education and training in parenting skills, again stressing the main message, that good parenting skills are essential for maximizing children’s intelligence as well as other outcomes. A national whole family campaign also should link to existing efforts to alter one or more of the family risk factors for intelligence. For example, the campaign could recognize current state and federal programs (including welfare reform) to reduce teenage pregnancy, encourage marriage before having children, increase fathers’ involvement in raising their children, limit family size (family cap policies), and increase income for low-income families. In addition, a national campaign should recognize other efforts to improve the cognitive skills of children who already are at risk because of adverse family conditions. The best example here is the Early Head Start program. The annual cost for a national whole family campaign, including an education component in secondary schools, is estimated at about $400 million a year, for at least five years. The cost would break down into three categories: —$100 million would be for the national campaign, including TV and print media advertising, development of a prototype policy and related educational materials, special conferences and training workshops, and so forth. —$200 million would be allocated to a secondary school education program, which would consist primarily of a $10,000 incentive grant for each of the approximately 20,000 public secondary schools nationwide to develop education classes and train teachers.

1269-06 CH06 04/18/03 17:25 Page 99

    

99

—$100 million would be set aside for comprehensive evaluation programs, including national studies to track attitude and behavior change with respect to the family risk factors as well as randomized trials for the secondary school education programs.

Rationale This initiative is based on research suggesting that early family environments have an impact on children’s intelligence and on the assumption that if parents or parents-to-be are more aware of these connections they will make better choices. Some definitions and assumptions are necessary. First, I am using the term intelligence to represent those types of mental ability that children need to succeed in school, including the ability to master reading and mathematics, and that adults need to master complex knowledge and skills. The term academic intelligence might be more precise; other equivalent terms are cognitive ability, academic ability, and intellectual ability. Second, I assume that scores from IQ tests, academic aptitude tests, and combined achievement tests in reading and math can serve as proxies for intelligence. While individuals differ in the degree to which they exhibit specific aptitudes or skills, especially verbal versus quantitative aptitudes, the major factors that influence the development of most academic aptitudes are similar. Third, I am using the term risk factor in the same way that it is used in medical research: that is, to identify conditions or behaviors that increase or decrease the risk (or probability) of getting various illnesses. In the case of intelligence, a risk factor is a condition or behavior that tends to increase or decrease a child’s intelligence. A risk factor can be either genetic (biological) or environmental in origin. The argument advanced here consists of four interrelated propositions about intelligence. First, intelligence is an important human attribute because of its critical role in academic success and many other outcomes. Second, intelligence is malleable, but its malleability decreases substantially with age. Third, the most important risk factors for intelligence are embedded in family characteristics and parenting behaviors. Fourth, because the family plays such a heavy role in determining the risk factors, the family is potentially the most important agent of change.

The Importance of Intelligence Despite frequent criticisms of the concept, intelligence is an important human attribute because it exerts a very strong influence on academic success and

1269-06 CH06 04/18/03 17:25 Page 100

 . 

100

Figure 6-2. Youth IQ in 1979 and Education Completed by 1996 Percent

90 80 70 60 50 40 30 20 10

1–19

20–39

40–59 IQ percentile

60–79

College graduate

High school graduate

Some college

High school dropout

80–99

Source: National Longitudinal Study of Youth.

educational attainment. Educational attainment in turn has a heavy impact on career choices because many jobs have specific education requirements— especially the requirement for a college degree. The chain of influence follows several stages. Preschool IQ is a good predictor of achievement scores in the early school grades, achievement scores in the early grades strongly predict achievement scores in the later grades, and achievement in the later grades predicts how far an individual goes in school. Perhaps the most decisive relationship is that found between high school IQ and years of schooling in the National Longitudinal Study of Youth (NLSY).10 The prediction here is very strong (see figure 6-2). Seventy-five percent of youths with IQs in the top fifth graduated from college, compared 10. Armor (2003, figure 2.8).

1269-06 CH06 04/18/03 17:25 Page 101

    

101

with only 3 percent of those in the bottom fifth. Nearly one-third of those in the bottom fifth dropped out of high school. Two-fifths of the second-highest quintile graduated from college, and only 1 percent dropped out of high school, while less than 10 percent of the second-lowest quintile graduated from college. Clearly, the probabilities of finishing high school, attending college, or completing college increase steeply as IQ rises. The relationship between youth IQ and later income is somewhat weaker. The overall correlation is .4, compared with more than .6 for years of education. The lower correlation for income makes sense. Some occupations, such as the teaching profession, require relatively high IQ levels but offer modest salaries compared with those in many business careers. Nevertheless, for every increase of 20 percentile points in youth IQ, annual income rises by $20,000.11 There are two caveats to this proposition. First, while these correlations are high and the predictions significant, they are not so high as to preclude other important human talents in social and economic success. Motivation, artistic creativity, and interpersonal skills certainly play significant roles in educational and career success, and such attributes explain why the correlation between IQ and educational attainment is less than perfect. But the fact that human abilities other than IQ help shape a person’s success is no reason to diminish the importance of cognitive ability, as has become fashionable in some quarters (seen, for example, in the University of California’s decision to drop the SAT as an admission requirement).12 Second, to say that IQ or achievement in the early grades affects IQ or achievement in the later grades does not mean that children are not learning throughout their school years. All children, with the exception of the severely mentally impaired, accumulate considerable knowledge and skills during their school years. Rather, since the norms for IQ and achievement tests are based on age, the relative ranking of a child’s intelligence compared with that of other children of the same age does not change very much during the school years. A child who starts school with an IQ that is 15 points higher than that of his or her peers is likely to finish school 15 points higher.

Age and the Malleability of Intelligence The malleability of intelligence is intertwined with the nature of intelligence and the extent to which intelligence depends on genetic factors. The 11. A person at the 40th percentile, for example, has a score above 40 percent and below 60 percent of the total population. 12. Atkinson (2001).

1269-06 CH06 04/18/03 17:25 Page 102

102

 . 

vast majority of psychologists and geneticists agree that genes play some role; they disagree over how much is due to heredity and how much depends on the environment.13 For the conceptual framework being proposed here, it is not necessary to know the exact portion of intelligence ascribable to genetic factors. It is more important to understand whether and when environmental factors are sufficiently involved to justify modifying those factors. For these reasons, the malleability of IQ is more important than its heritability. Several types of evidence establish the malleability of intelligence and the effect of age on malleability. First, studies show that IQ scores are increasing for societies as a whole, the so-called Flynn effect.14 These studies show rising absolute levels of cognitive skills averaged over a total population; hypotheses range from better nutrition, the growth of information technology, more widespread schooling, or other cultural factors. These studies do not show relative gains in IQ for some children compared with other children. Second, the National Assessment of Educational Progress (NAEP) shows substantial improvement in African American and Hispanic achievement between 1970 and 1990.15 The NAEP studies not only establish the malleability of intelligence, but the relative improvement of minority compared with white achievement demonstrates that some types of environmental change benefited black and Hispanic children more than white children. While there is no consensus about the environmental factors responsible for this reduction in the achievement gap, most experts agree that the improved socioeconomic conditions of minority families played an important role. Third, the stability coefficients for IQ (correlations between IQ scores at two points in time) decrease at younger ages. While even the earliest IQ-to-IQ correlations are quite high, indicating that early IQ is a good predictor of later IQ, the magnitude of the correlation diminishes at younger ages. This finding not only supports the proposition that intelligence is malleable, it also supports the corollary that malleability decreases as children get older.16 A fourth type of evidence supports both the malleability of intelligence and the effect of age. Educational interventions for infants and toddlers (Abecedarian project, Early Head Start) have raised IQ scores; major interventions for older children like regular Head Start and Title 1 have not.17 This supports the conclusion that intelligence is most malleable at younger ages. While regular Head Start programs have had other beneficial impacts 13. Neisser and others (1996). 14. Flynn (1998). 15. Grissmer, Flanagan, and Williamson (1998). 16. Armor (2003, chapter 2). 17. Center for the Future of Children (1995); Karoly and others (1998); Mathematica (2002).

1269-06 CH06 04/18/03 17:25 Page 103

    

103

on child development, such as improved behavioral outcomes (lower dropout rates, less crime, lower divorce rates), only the earliest interventions have been able to produce significant and lasting effects on IQ scores. Finally, the malleability of intelligence is supported by the relationship between various risk factors and early IQ scores. The fact that environmental risk factors have a strong correlation with IQ score by the ages of three to five, independent of parent IQ, is perhaps the most telling evidence about the malleability of intelligence during early childhood.

The Risk Factors for Intelligence There is considerable evidence in the research literature about the conditions or behaviors—what I am calling risk factors—that have the greatest potential influence on a child’s IQ. Risk factors can be either genetic (biological) or environmental in origin. Technically, a risk factor for intelligence is a condition or behavior that is significantly correlated with a child’s IQ by the age of four or five. According to existing empirical research, the ten most important risk factors for a child’s intelligence are as follows (ordered by timing):18 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Parents’ IQ Parents’ educational attainment Family income and poverty status Family structure (marital status and number of parents at home) Age of mother when child is born The number of a child’s siblings The child’s nutrition (including breast feeding) The child’s birth weight Parental instruction (cognitive stimulation) Parental nurturing (emotional support).

With the exception of parents’ IQ and low birth weight, these factors clearly are environmental conditions or behaviors, meaning that theoretically they can be modified by parents or supplemented by other caregivers. I classify parent IQ as a genetic risk factor, although it could include an environmental component (for example, more intelligent parents might have better parenting skills). For the purpose of this conceptual framework, it does not matter whether parent IQ is genetic or environmental, because prospective parents cannot change their IQ. Low birth weight might have biological and 18. Armor (2003).

1269-06 CH06 04/18/03 17:25 Page 104

104

 . 

environmental aspects. To the extent it is caused by medical or physiological conditions beyond a mother’s control, it is a biological risk factor. To the extent that it is influenced by other environmental risk factors, such as the age of the mother or poor nutrition, then low birth weight could be classified as partly environmental. The risk factors of family income, nutrition, instruction, and nurturing can be modified after a child is born, and they also can be supplemented to some extent by individuals besides the parents. Theoretically, the instruction and nurturing behavior of parents has the greatest potential influence here, since infants and toddlers spend most of their time with their parents or guardians. Clearly, instruction can be supplemented in child care settings by professional caregivers. It is less clear whether parent nurturing can be supplemented effectively by professional caregivers. Absence of affection and emotional support from parents may create emotional problems that cannot be overcome by outside intervention. Certain types of nutritional care can be offered outside the family, but breast feeding cannot. Finally, family income can be supplemented through welfare programs, but it also can be influenced directly by parents’ decisions and behaviors, such as marrying in order to have two incomes. The remaining risk factors (education, family structure, number of children, and age of mother) are largely determined by the decisions and behaviors of prospective parents before conception. Of course, because of the heavy influence of IQ, the amount of a parent’s education is not simply a decision, although it can be consciously changed within limits (for example, dropping out of high school usually can be avoided). Having children without being married, age at childbearing, and number of children clearly are the result of decisions made by prospective parents, especially the mother. While other adults can counsel young people and prospective parents about risk factors, the decisions themselves are solely in the hands of the prospective parents. The following is a brief summary of the evidence on the importance of each of the risk factors. All ten risk factors had significant correlations with children’s IQ scores at age five and math achievement at age nine in the Children of the National Longitudinal Study of Youth (CNLSY), as of 1996.19 In addition, each factor has a significant effect on a child’s IQ or math achievement in multiple regression analyses of all factors (as illustrated below, in table 6-2).20 Each of the risk factors also has been examined in numerous studies, many of which are based on national data. Table 6-1 19. See Armor (2003, chapter 3) 20. See Armor (2003, appendix A).

1269-06 CH06 04/18/03 17:25 Page 105

    

105

Table 6-1. Risk Factor Correlations with Child’s IQ and Math Achievement Risk factor Mother’s IQ Father’s education Mother’s education Family income Two-parent familya Mother’s age at first birth Number of children Breast feedingb Birth weight Instruction score (cognitive stimulation) Nurturing score (emotional support) Number of cases

IQ

Math

.54 .33 .33 .36 .30 .31 −.21 .25 .13 .42 .37 4,630

.48 .33 .33 .32 .24 .30 −.13 .20 .14 .35 .29 4,630

Source: Children of the National Longitudinal Study of Youth. a. Two parents = 2; married/father absent = 1; never married = 0. b. Yes = 1.

shows the simple correlations between each of the risk factors and IQ and math achievement from the CNLSY. ’ . For mother’s IQ, the CNLSY shows a correlation of .54 for child’s IQ and .48 for math achievement. This is consistent with other studies that show a range of correlations between parents and their biological children on the order of .4 to .5.21 It also is consistent, as of 1992, with another analysis of the CNLSY .22 High sibling and parent-child correlations could reflect the effects of similar home environments as well as genetic influences. High IQ parents may place greater emphasis on teaching various skills to their children, and they may have a richer intellectual interaction with their child involving more complex words, ideas, games, and so forth, which provides greater stimulation for the child’s intellectual development. ’ . The CNLSY shows correlations of .33 for mother’s years of education; they are the same for father’s education. In the 1996 NAEP, high school seniors whose parents were college graduates reached the “proficient” level 50 percent of the time while seniors whose parents were high school dropouts reached this level only 20 percent of the time.23 21. Rowe (1994). 22. Phillips (1998). 23. Donahue and others (1999, table 3.3).

1269-06 CH06 04/18/03 17:25 Page 106

106

 . 

 . Family income shows a somewhat higher correlation for child’s IQ and about the same correlation for math achievement in the CNLSY (.36 and .32, respectively). There are many national studies of the relationship between poverty and academic ability. In a national evaluation of Title 1, first-graders whose family was below the poverty level scored at only the 31st percentile on a standardized reading test (the national average is the 50th percentile).24 Similar results have been found in a more recent national study of 20,000 children just beginning kindergarten.25  . In my CNLSY analyses, family structure is scored 2 for two parents, 1 for a single mother who was married, and 0 for a nevermarried single mother. The correlations between this family structure variable and IQ and math achievement are .30 and .24, respectively. Some of the best evidence about family structure also comes from NAEP. In the 1992 national assessment of mathematics for eighth-graders, those with both parents at home scored 103, children with only the mother at home scored 97, and children with neither parent in the home scored 92. (For this analysis, NAEP scores were standardized like IQ scores; that is, the national mean is 100 and the standard deviation is 15.) ’   ’ . The CNLSY correlations are .31 for mother’s age and .30 for math achievement. A study using data from the CNLSY and the National Survey of Children concluded: “Children of the youngest teen mothers are less likely to have received well-baby care in the first year of life, have less cognitively stimulating and less nurturing home environments, and obtain lower cognitive achievement scores than peers whose mothers were 20 to 21 at their births.”26   . The CNLSY correlations are inverse, so that the larger the number of children in a family, the lower the IQ (−.21) and math achievement (−.13). The most common explanation for this phenomenon is the “dilution of resources” in larger families.27 Because more parenting resources and attention are required to develop higher cognitive ability, the larger the number of children, the lower the resources and time available to each child, and therefore the lower their academic ability.28  ⁽   ⁾ . The only nutrition variable in the CNLSY is breast feeding, which has correlations of .25 and .20 with IQ and math achievement. The specific nutritional deficiencies with the greatest 24. Puma and others (1997). 25. West, Denton, and Germino-Hausken (2000). 26. Moore, Morrison, and Greene (1997). 27. Blake (1989). 28. Downey (1995).

1269-06 CH06 04/18/03 17:25 Page 107

    

107

involvement in a child’s cognitive development concern iodine, iron, and fatty acids.29 Fatty acids are believed to be the reason for the correlations between IQ and breast feeding. While there is no consensus yet on the precise types of fatty acids involved, there is substantial evidence that breast-fed babies have better cognitive outcomes than babies who are not breast fed. A review of twenty separate studies on this issue has determined that after adjusting for the appropriate factors, breast feeding is associated with significantly higher cognitive scores than formula feeding.30  . Birth weight is generally believed to be a threshold issue (low birth weight is defined as less than 5.5 pounds), but the CNLSY shows somewhat higher IQ and math achievement along a continuum from three to ten pounds. The reasons for the adverse effects of lower birth weight involve various developmental factors such as inadequate nutrition, restricted blood supply (as in the case of teenage mothers), and stress.31 One of the largest studies followed 14,000 full-term infants born in Great Britain in 1970.32 Significant cognitive deficits for children with low birth weight were observed at both ages five and sixteen, and at age ten these children had significantly lower math scores.     . In the CNLSY the risk factor with the highest correlations after mother’s IQ is instruction (cognitive stimulation), which is .42 for IQ and .35 for math. The correlations for nurturing (emotional support) are also fairly high, at .37 for IQ and .29 for math. These are part of the HOME scale of parenting behaviors.33 Other studies have used the CNLSY and other national data to come to similar conclusions about the importance of HOME scores on a child’s IQ. Phillips and others found large effects of HOME scores on verbal IQ after controlling for many other family variables, including mother’s IQ, in both the CNLSY and the Infant Health and Development Study.34 Other studies using the HOME scale have come to similar conclusions.35 The simple correlations in table 6-1 do not convey the complexity of determining the “independent” effects of each factor, particularly the difficulty of separating the effects of the environmental factors from the potentially genetic effect of parent IQ. This task is complicated by the fact that all 29. Kretchmer, Beard, and Carlson (1996). 30. Anderson and others (1999). 31. Lagercrantz (1997). 32. Strauss (2000). 33. Bradley and Caldwell (1976) 34. Phillips and others (1998). 35. Molfese and others (1997).

1269-06 CH06 04/18/03 17:25 Page 108

108

 . 

of the risk factors are correlated with one another. For example, mother’s education has a strong correlation with child’s IQ, but when the effect of mother’s IQ is removed, the effect of mother’s education is diminished considerably. Mother’s IQ is the strongest single predictor of child’s IQ at age five, but after controlling for environmental factors the effect of mother’s IQ is reduced by half. After removing the effect of mother’s IQ, number of children has a stronger effect on child’s IQ than either family income or family structure. The direct effect of two-parent families is small when controlling for all other risk factors, but the indirect effect of two-parent families is substantial because of its very large effects on income and on instruction and nurturing. The estimated effects of the risk factors following appropriate multivariate analyses with the CNLSY data are shown in table 6-2.36 Since the CNLSY did not measure father’s IQ, in this analysis father’s education is used as a surrogate for father’s IQ and is grouped with mother’s IQ as a potential biological factor.37 The environmental risk factors with the strongest effects on a child’s IQ are two-parent family, number of children, breast feeding, and parents’ instruction and nurturing; birth weight and family income have weaker effects and mother’s age is not significant. For math achievement, parenting behaviors, mother’s age, and birth weight have the strongest effects, but neither family structure nor breast feeding is significant. However, it should be noted that as shown in the bottom panel of table 6-2, two-parent family has a strong indirect effect on math achievement operating through the parenting behaviors of instruction and nurturing. If these eight environmental factors could be optimized for a family, a child’s IQ might be raised by nearly 10 points, independent of the mother’s IQ, and math achievement could be raised by 5 points.38 A caveat for the risk factors is in order. A risk factor does not produce a fixed effect on IQ for a particular child; rather, a risk factor means children with favorable scores on that risk factor have a higher probability of higher IQ. Moreover, there is no guarantee that changing all eight environmental risk factors would raise the IQ of a given child. Rather, children who have optimal levels for all risk factors are most likely to have the highest IQs and the best academic records in school.

36. See Armor (2003), appendix A, for details of this analysis. 37. In the CNLSY analyses, mother’s IQ is strongly correlated with mother’s education, suggesting that father’s education can be used as an indicator of father’s IQ. 38. Both IQ and math achievement have standard deviations of 15 points, so these would be large effects.

1269-06 CH06 04/18/03 17:25 Page 109

    

109

Table 6-2. Estimated Effects of Risk Factors on Child’s IQ and Math Achievement Estimated effect Risk factor

Change

IQ

Math

Potential direct biological effects Mother’s IQ Father’s education (for IQ) Total

+ 10 points + 2 years

3.3 0.6 3.9

2.9 1.1 4.0

+ 10 points

1.8

1.3

+ 10 points

1.3

0.9

− 1 child yes v. no + 1 pound 2 v. mother never married +$10,000 + 5 years

1.8 1.7 0.4 1.8 0.3 * 9.0

0.5 * 0.7 * 0.2 1.2 4.8

+ 10 points + 2 years 2 v. mother never married − 1 child +$10,000

0.5 0.3 1.7 0.5 0.3

0.5 0.2 1.6 0.4 0.3

Direct environmental effects Instruction (cognitive stimulation) Nurturing (emotional support) Number of children Breast feeding Birth weight Family structurea Family income Mother’s age at first birth Total direct effects Selected indirect effectsb Mother’s IQ Father’s education (for IQ) Family structurea Number of children Family income

Source: Children of the National Longitudinal Study of Youth. *Not significant at p ≤ .05; all other estimates significant. a. Two parents = 2; married/father absent = 1; mother never married = 0. b. Through instruction and nurturing.

The Family as Agent of Change Since the environmental risk factors operate primarily through parental decisions and behaviors and since their influence is strongest during a child’s earliest years, parents have the greatest potential for influencing their children’s life chances. Whether parents actually are willing and able to change their behavior is another matter. For example, knowing that smoking and obesity are risk factors for heart attack does not tell us how to get people to

1269-06 CH06 04/18/03 17:25 Page 110

110

 . 

stop smoking and lose weight in order to lower their risk of heart disease. Likewise, knowing that parent behaviors are important does not tell us the most feasible ways of changing their behaviors. At least four of the risk factors—income, nutrition, cognitive stimulation, and emotional support— can be supplemented or provided by persons other than parents. In addition, many educators sincerely believe that regardless of the skill differences that children bring to school, well designed and adequately funded school programs should be able to overcome the cognitive skill deficiencies of disadvantaged children.

Concluding Comments While whole family policies are still in their infancy, the conceptual framework proposed here still points to the family as having the greatest potential for becoming an agent of change. Since parents are the only people who have control over all of the environmental risk factors, and since they have more contact with their children during early childhood than anyone else, they are in the best position to optimize their children’s early environment and thereby maximize their children’s intelligence. The policy challenge is to inform parents of their unique position and motivate them to improve these environments to the extent feasible. As with all education programs for preventing or changing certain behaviors, program effectiveness is determined by the perceived consequences of the behavior and the credibility of the source. If family policy groups agree on the importance of the family to a child’s intelligence, if the Secretary of Health and Human Services endorses the whole family policy to kick off a national campaign, and if education programs are started in secondary schools, it is possible that young people would start thinking about the consequences of their choices regarding education, sexual activity, marriage, and childbearing. Ultimately, the best way to maximize intelligence and academic achievement for the greatest number of children may be to convince parents-to-be that their children’s intelligence depends strongly on their own decisions and behaviors, and then give them information and support needed to make the best decisions and and engage in the best behaviors. Note: Major portions of this chapter are based on Armor (2003).

1269-06 CH06 04/18/03 17:25 Page 111

    

111

References Anderson, James W., and others. 1999. “Breast-Feeding and Cognitive Development: A Meta-Analysis.” American Journal of Clinical Nutrition 70: 525–35. Armor, David J. 2003. Maximizing Intelligence. New Brunswick, N.J.: Transaction Publishers. Atkinson, Richard C. 2001. “The 2001 Robert H. Atwell Distinguished Lecture,” delivered at the 83d Annual Meeting of the American Council on Education, Washington, D.C., February 18. Blake, Judith. 1989. Family Size and Achievement. University of California Press. Caldwell, B. M., and R. H. Bradley. 1984. Home Observation for Measurement of the Environment. Little Rock: Center for Child Development and Education, University of Arkansas. Center for the Future of Children. 1995. “Long-Term Outcomes of Early Childhood Programs.” Future of Children 5 (3). ———. 1999. “Home Visiting: Recent Program Evaluations.” Future of Children 9 (1). Donahue, Patricia L., and others. 1999. NAEP 1998 Reading: Report Card for the Nation and the States. Washington: National Center for Education Statistics. Downey, Douglas B. 1995. “Family Size, Parental Resources, and Children’s Education.” American Sociological Review 60: 746–61. Fein, David, Rebecca A. London, and Jane Mauldon. 2002. Welfare Reform and Family Formation. Research Brief 1. Abt Associates (www.abtassociates.com/wrffproject [January 3, 2003]). Finn, Jeremy D., and Charles M. Achilles. 1999. “Tennessee’s Class Size Study: Findings, Implications, Misconceptions.” Educational Evaluation and Policy Analysis 21: 97–109. Flynn, James. 1998. “IQ Gains over Time: Toward Finding the Causes.” In The Rising Curve, edited by Ulric Neisser. Washington: American Psychological Association. Grissmer, David, A. Flanagan, and S. Williamson. 1998. “Why Did the Black-White Score Gap Narrow in the 1970s and 1980s?” In The Black-White Test Score Gap, edited by Christopher Jencks and Meredith Phillips. Brookings. Herrnstein, R. J., and C. Murray. 1994. The Bell Curve. New York: Free Press. Karoly, Lynn A., and others. 1998. Investing in Our Children: What We Know and Don’t Know about the Costs and Benefits of Early Childhood Interventions. Report RB-5014. Santa Monica, Calif.: Rand Corporation. Kretchmer, Norman, John L. Beard, and Susan Carlson. 1996. “The Role of Nutrition in the Development of Normal Cognition.” American Journal of Clinical Nutrition 63 (6): 997S–1001S. Lagercrantz, Hugo. 1997. “Better Born Too Soon than Too Small.” Lancet 350 (October 11): n9084, p1044(2). Mathematica Policy Research. 2002. Making a Difference in the Lives of Infants and Toddlers and Their Families: The Impacts of Early Head Start. Head Start Bureau, U.S. Department of Health and Human Services. McKey, Ruth H., and others. 1985. The Impact of Head Start on Children, Families, and Communities. Washington: U.S. Government Printing Office. Molfese, Victoria J., and others. 1997. “Prediction of the Intelligence Test Scores of 3- to 8-Year-Old Children by Home Environment, Socioeconomic Status, and Biomedical Risks.” Merrill-Palmer Quarterly 43 (2): 219–34.

1269-06 CH06 04/18/03 17:25 Page 112

112

 . 

Moore, Kristin Anderson, Donna Ruane Morrison, and Angelo Dungee Greene. 1997. “Effects on the Children Born to Adolescent Mothers.” In Kids Having Kids, edited by Rebecca A. Maynard. Washington: Urban Institute Press. Murray, Charles. 1984. Losing Ground: American Social Policy, 1950–1980. New York: Basic Books. National Campaign to Prevent Teen Pregnancy. 2001. Halfway There (www.teenpregnancy.org/resources/data/pdf/halfway.pdf [February 19, 2003]). ——— . 2002. Not Just Another Single Issue (www.teenpregnancy.org/resources [February 19, 2003]) Neisser, Ulric, and others. 1996. “Intelligence: Knowns and Unknowns.” American Psychologist 51: 77–101. Phillips, Meredith, and others. 1998. “Family Background, Parenting Practices, and the Black-White Test Score Gap.” In The Black-White Test Score Gap. Puma, Michael, and others. 1997. Prospects: Final Report on Student Outcomes. Cambridge, Mass.: Abt Associates (April). Rowe, David C. 1994. The Limits of Family Influence. New York: Guilford Press. Schweinhart, Lawrence J., and others. 1993. Significant Benefits. Ypsilanti, Mich.: High/Scope Press. Strauss, Richard S. 2000. “Adult Functional Outcome of Those Born Small for Gestational Age: Twenty-Six-Year Follow-up of the 1970 British Birth Cohort.” Journal of the American Medical Association 283 (5): 625. U.S. Bureau of the Census. 2000. “Families and Living Arrangements.” Current Population Survey Reports. Historical time series. Table FM-2 (www.census.gov/population/ www/socdemo/hh-fam.html [March 26, 2003]). West, Jerry, Kristin Denton, and Elvira Germino-Hausken. 2000. America’s Kindergartners. Washington: National Center for Education Statistics.

1269-07 CH07 04/18/03 17:26 Page 113

7 Providing Universal Preschool for Four-Year-Olds BARBARA WOLFE SCOTT SCRIVNER

AND

T

here is general agreement among experts in child development and early education that four-year-olds gain from being in a stimulating environment with other children. This chapter builds on their findings to argue that all children, regardless of family income, family composition, and prior experience in child care, would gain from being in a well-implemented preschool program at the age of four.1 Such an argument calls for a major new investment in preschool, which could take one of two forms: a program targeting children from low-income families, such as those eligible for Head Start, or from families with incomes below, say, 185 to 200 percent of the poverty line; or a universal program for all four-year-olds. Although the greatest potential gain per child is likely to occur among four-year-olds in lowincome families, we believe that the country would be even better off with a program for all four-year-olds. All children, regardless of income, would then be able to start kindergarten better prepared for school and experience the improved developmental outcomes 1. See, for example, Blau (2001), Currie (2001), Karoly and others (1998), Heckman (2000), and Peisner-Feinberg and Burchinal (1997).

113

1269-07 CH07 04/18/03 17:26 Page 114

114

    

associated with an early start. Such a program, through its universality as well as through children’s improved developmental outcomes, might also have a beneficial secondary effect: increased appreciation of and support for early education programs, perhaps leading to improved program quality, greater participation by families, less stratification of young children by income, and greater public willingness to acknowledge the benefits of financing preschool for all four-year-olds. As we later discuss, greater public support for early childhood education (ECE) and recognition of its benefits are cited as important factors in the more comprehensive ECE systems used in other developed nations.

The Argument for a Universal Preschool The years from birth to the age of five increasingly are viewed as a critical period for establishing the foundations of thinking, behaving, and emotional well-being. A recent National Academy of Sciences study, From Neurons to Neighborhoods, cites a large research literature showing that during these years, children develop linguistic, cognitive, social, emotional, and regulatory skills that predict their later functioning in many domains.2 The largest and perhaps most convincing study of the potential gains from attending prekindergarten is based on an evaluation of the Chicago Child-Parent Center (CPC) program, an intervention operating in Chicago public schools in low-income areas. Evidence from this intervention indicates that participation in prekindergarten at ages three or four is associated with significantly higher rates of school completion by age twenty; lower rates of official juvenile arrests, violent arrests, and multiple arrests by age eighteen; and with lower rates of placement in special education and of grade retention. In their recent benefit-cost analysis of the CPC program, Reynolds and others found that “participants had an 11-point lower rate of special education placement (40 percent reduction), a 12-point lower rate of grade retention (40 percent reduction), an 8.2-point lower rate of juvenile arrest (33 percent reduction), and an 11-point higher rate of high school completion by January 2000.”3 They report that the CPC preschool program generated a return to society of $47,759 per participant by age twenty-one. Of this, nearly $26,000 took the form of benefits to the general public (as taxpayers 2. Shonkoff and Phillips (2000). Because our interest is in a universal program, we cite here only evidence from studies of existing large-scale programs. 3. Reynolds and others (2002).

1269-07 CH07 04/18/03 17:26 Page 115

   --

115

and crime victims). The average cost of the program per participant was under $7,000, so benefits far exceeded costs. Another major study, by the National Institute of Child Health and Human Development Early Child Care Research Network, sought to evaluate the impact of child care quality on child functioning.4 Child care quality was measured by observing how and when caregivers interacted with study children. The quality of children’s experiences with child care staff was found to be an important influence on children’s language, preacademic, and social competence, but the family was still the most important determinant of young children’s development. These findings are confirmed by the results of the Cost, Quality, and Child Outcomes study, which examined the influence of typical center-based care on children’s development during their preschool years and as they moved into the formal elementary education system.5 Researchers found, in brief, that children who attended higher-quality centers performed better on tests of cognitive and social skills and that the quality of care continued to predict their performance into elementary school. Moreover, children traditionally at risk of not doing well in school were more sensitive to the negative effects of poor child care and received greater benefits from higher-quality child care than other children. Garces, Thomas, and Currie address the question of gains from Head Start, using data from a national panel survey, the Michigan Panel Study of Income Dynamics (PSID), employing four indicators of economic and social success in adulthood.6 They report that for whites, participation in Head Start was associated with a significantly increased probability of completing high school and attending college as well as with elevated earnings in their early twenties. African Americans who participated in Head Start were less likely to have been convicted or charged with a crime. The authors also report some evidence of a positive association with high school completion among African American males. All of these studies suffer from potential selection bias in terms of which children attended preschool, and there is some question of whether the results are generalizable. Even so, the evidence seems strongly consistent with gains from a high-quality, structured prekindergarten program. These gains include an increase in the probability of school readiness, higher school achievement, reduced grade retention, reduced use of special education, 4. NICHD (2000, 2002). 5. Peisner-Feinberg and others (1999). 6. Garces, Thomas, and Currie (2000).

1269-07 CH07 04/18/03 17:26 Page 116

116

    

lower crime among adolescents, lower rates of teen pregnancy, and better performance in the labor market as young adults. These gains accrue to the individuals in the programs, to their families, their classmates, school mates, and the rest of society. The evidence of gains seem to be strongest, and most studied, among children born to economically disadvantaged families, who are at risk of poor educational outcomes.7 The plan we propose would operate year-round since most jobs are yearround and there appear to be considerable advantages to children from consistent care arrangements. The program would be for a full day, with two components: a part-day education component taught by early education specialists or elementary school teachers, assisted by aides; and a child care component, in which aides could provide care under the supervision of lead teachers. This two-part arrangement would combine the advantages of preschool preparatory activities with less intensive activities appropriate for children of this age. The two-part plan also should make the cost of universal prekindergarten lower than that of a full day of schooling.8 Implementation might have some aspects of state child health insurance programs (S-CHIP), in that states might be given a set of options and requirements regarding the running of prekindergarten programs. They might include using the state’s department of education and the school districts to establish a program, administering the program under existing state-level child care departments, setting up a new independent entity to run the program, or some combination of these options. Currently, a patchwork of programs provides ECE directly or through financial assistance for preschool and child care programs for four-yearolds.9 We review programs that provide ECE or subsidies for ECE in the United States, to illustrate both current efforts and unmet needs. We then present our proposal for financing a universal preschool for four-year-olds in the United States. We also consider state-financed prekindergarten programs and discuss in greater detail the state program that comes closest to universal preschool, the Georgia Voluntary Pre-K (GPK) program. We conclude with a discussion of the political feasibility of such a proposal that briefly adverts to experience with such programs in other developed countries.

7. Lee and Burkam (2002); Pianta and others (2002) 8. In certain cases, it might be possible for the teachers/specialists to provide schooling for two different groups of four-year-olds each school day, one in the morning and one in the afternoon. 9. Olson (2002).

1269-07 CH07 04/18/03 17:26 Page 117

   --

117

The Current Picture of Four-Year-Olds and Preschool Education Using census data, the National Center for Education Statistics estimates that in 2000 there were about 4 million four-year-olds, 65 percent of whom were enrolled in preprimary programs, defined as “any type of graded public, parochial, or other private schools. This includes nursery schools, kindergartens, and elementary schools. Attendance may be on either a full-time or part-time basis.”10 This percentage is down from 69 percent in 1999, but up from 60 percent in 1994. Therefore approximately two-thirds of all fouryear-olds already are in some type of preschool.

Major Programs for Four-Year-Olds The major federal programs are summarized in table 7-1. In addition to Head Start, the Child Care and Development Fund, and Title I of the Elementary and Secondary School Act, they include both federal and state tax policies that subsidize child care through tax credits or tax subsidies.  . The single largest national public preschool program for four-year-olds is Head Start, which has provided early education to millions of children since it began operating in 1965. Most Head Start centers offer part-day programs for the duration of the regular school year, although a number of states provide funds to extend services to include full-day, fullyear programs. Head Start focuses on children from low-income families, and program rules require that 90 percent of the children enrolled be from families that participate in the Temporary Assistance for Needy Families (TANF) program or whose income is below the poverty line, limiting the ability of Head Start to serve larger populations. Total federal funding for Head Start was $6 billion in FY2001, with states contributing an additional $200 million.11 In FY2000, the program served 857,664 children, including about 480,300 four-year-olds, slightly more than 12 percent of all four-year-olds.12 This is about 60 percent of eligible threeand four-year-old children, according to estimates by the Children’s Defense Fund.13 The U.S. Department of Health and Human Services reports that 54 percent of children participating in Head Start were four-year-olds.14 If 10. National Center for Education Statistics (2002). 11. National Center for Children in Poverty (2000). 12. U.S. Department of Health and Human Services (2002). 13. Children’s Defense Fund (2001). 14. U.S. Department of Health and Human Services (2002).

Low-income children Low-income children Educationally and economically disadvantaged children All children of employed parents Varies

Target population

Sources: Authors’ calculations based on data as follows. Head Start: National Center for Children in Poverty (2000); CCDF: U.S. Department of Health and Human Services (2002); Title I: U.S. General Accounting Office (2000); tax policy: Joint Committee on Taxation (2002); state initiatives: Edwards (2002). a. Assumes program expenditures are uniformly distributed by age. b. Does not include approximately $1.0 billion in direct TANF expenditures for child care. c. GAO (2000) reports Title I funding for preschoolers, who, it notes, ranged between ages three and five. We make the conservative assumption that 50 percent of them were four-year-olds. d. Represents forgone tax revenue. e. Assumes that the proportion of CCDF program funding dedicated to four-year-olds applies also to these tax policies. f. Based on data from Schulman, Blank, and Ewen (1999) that describe ages covered by state prekindergarten programs.

— —

500 millione 1.4 billiona, f

3.7 billiond 1.9 billion

Tax policy (DCTC and DCAP) State pre-K initiatives

6,700 4,100 1,380

3.3 billiona 936 milliona 216 milliona, c

6.0 billion 7.2 billionb 8.4 billion

Head Start CCDF Title I of ESEA

Estimated funding per four-year-old

Funding devoted to four-year-olds

Total funding

Program

2001 dollars

Table 7-1. Major Programs That Subsidize Early Childhood Education for Four-Year-Olds

1269-07 CH07 04/18/03 17:26 Page 118

1269-07 CH07 04/18/03 17:26 Page 119

   --

119

we uniformly distribute program expenditures, this suggests that $3.3 billion was spent to provide care and education to this group under Head Start.      . The Child Care and Development Fund (CCDF), authorized under the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, provides child care assistance to families receiving TANF aid or leaving a public assistance program and to other low-income working families. The CCDF provides funds to states, which then must meet federal requirements for state financial contributions. Total funding under CCDF comes from three streams. First, states receive a share of the discretionary funds appropriated to CCDF through the annual congressional budget process. States also qualify for mandatory funds representing the amounts that states received from previous federal child care programs in a base period. Finally, states can receive additional matching funds if they meet maintenance of effort requirements and if they commit additional state funds to draw down the matching funds. States also may transfer up to 30 percent of their TANF block grant to the CCDF program. Federal law requires that recipients of subsidies under CCDF must have incomes below 85 percent of state median income.15 In practice, the maximum income threshold varies substantially across states, from $16,000 for a family of three living in Wyoming to $39,000 for a similar family in Connecticut. In addition to meeting state income requirements, eligible children also must be under the age of thirteen, have parents who are working or in a training program, or be in need of child protective services. States generally cannot use CCDF funds to construct new facilities. In FY2000, CCDF provided $5.1 billion in federal funding and an additional $1.9 billion in state matching and maintenance of effort spending, including transfers from TANF to CCDF.16 Again assuming a uniform distribution of expenditures among age groups, this implies that approximately $1 billion was spent under CCDF for the care of four-year-olds. The U.S. Department of Health and Human Services reports that states also directly provided approximately $1 billion in child care services under TANF and estimates that in FY1999 13 percent of children served under CCDF (229,000) were four-year-olds.17 Given the above enrollment and assuming no overlap with four-yearolds in Head Start, this would mean that CCDF serves about 6 percent of all 15. The U.S. Department of Health and Human Services (2000) noted that in 1999 only nine states set allowable income at the 85 percent of state median income maximum. 16. U.S. Department of Health and Human Services (2002). 17. U.S. Department of Health and Human Services (2002 and 2001).

1269-07 CH07 04/18/03 17:26 Page 120

120

    

four-year-olds and that the programs together serve about 18 percent of all four-year-olds.18         . Under Title I of the Elementary and Secondary Education Act (ESEA), states may allocate a portion of federal funds to preschools that meet Head Start performance standards. School districts or local schools also may choose to use some portion of these funds to serve preschool children. Title I grants are determined by a formula that includes a number of factors, including the average per-pupil expenditure in the state, the child poverty rate, and past distributions to the state and the district. The U.S. Department of Education reports that approximately 90 percent of school districts receive grants. The U.S. General Accounting Office estimates that 17 percent of school districts that received Title I funds used them to provide or augment preschool programs but that the districts used less than 10 percent of their total grant for those purposes.19 In total, states spent approximately $407 million in Title I funding on preschool in 1999–2000. Large districts (40,000 or more students) were four times more likely to provide Title I funding to preschoolers than small districts. Overall, GAO estimates that 8 percent of children entering preschool received some funding from Title I. GAO also reports that almost all children served by Title I grants for preschoolers were between the ages of three and five; more than 90 percent of districts set a minimum age of three or four. Under the assumption that preschool most commonly serves four-year-olds, we make the conservative estimate that 50 percent of Title I funds, or $203 million, went to serve four-year-olds.   . Two elements of the federal tax code, the Dependent Care Tax Credit and Dependent Care Assistance Program, assist families with young children, but their usefulness is greatest for middle- and upper-income families because they provide relief primarily for those with positive income tax liability.20 The Dependent Care Tax Credit provides a maximum credit of $2,100 for employment-related child care; twenty-two states provide similar credits, 18. Since the CCDF participation data provide the monthly average number of children served rather than ever enrolled during the year, this may underestimate the number of children actually served. On the other hand, there may be overlap between those covered by Head Start and those receiving CCDF funds, in which case our approach overstates the number of four-yearolds served. 19. U.S. General Accounting Office (2000). The estimates are based on a GAO survey of 16,000 school districts across the nation. 20. The credit begins to benefit a single working mother with two children when her earnings surpass $10,000 (in 2001 dollars); for earnings above $10,000, the family receives a credit equal to 10 percent of family income up to a maximum of $600 per child.

1269-07 CH07 04/18/03 17:26 Page 121

   --

121

ranging from $25 (Louisiana) to $1,584 (New York).21 The Dependent Care Assistance Program allows employees and employers to exclude child care expenses from federal income, FICA, and unemployment taxes, up to $5,000 per employee. Annual forgone federal revenue resulting from these two tax incentives was approximately $3.7 billion in 2001.22 By applying the same proportion of total CCDF program expenditures dedicated to four-year-olds to the tax credits, which target a similar group of children, we estimate that these policies represent approximately $0.5 billion in forgone revenue.   . States themselves use a variety of tax mechanisms to fund their prekindergarten programs, with many making allocations to ECE directly from general revenues. Some states and localities use tax instruments designed to raise revenue for child care.23 Florida law, for example, allows the creation of special taxing districts and grants them the power to increase property taxes to meet special needs. Under this law, several Florida districts dedicate a portion of new property tax revenues to ECE and child care. Given that property taxes are the major source for K–12 public school funding, this type of expansion also is quite fitting. Although states receive a sizable amount of revenue from sales taxes, only a few localities, most notably Aspen, Colorado, use sales tax revenue for child care or ECE. Colorado also is the only state to have a voluntary child care check-off on its income-tax forms that allows taxpayers to make a contribution to child care programs. In addition, a few states use excise taxes to fund child care and ECE programs. Both California and Indiana, for example, increased the tax on tobacco products and earmarked increased revenues for ECE. States also generate revenue from fees, and several states allow residents to dedicate a portion of their vehicle registration fee to broadly defined programs for children or to purchase special license plates from which a portion of the price is donated to programs for children. Thirteen states use proceeds from state lotteries to fund education, and two states, Georgia and Missouri, use lottery funds to support a prekindergarten program specifically. As we later discuss in the context of Georgia’s universal prekindergarten program, lotteries have a potentially large revenue-generating capacity. Education Week reports that all together, state-financed programs cost approximately $1.9 billion.24 Some of these programs target children older or younger than age four, although we estimate that the total amount spent on

21. Mitchell, Stoney, and Dichter (2001). 22. Joint Committee on Taxation (2002). 23. Mitchell, Stoney, and Dichter (2001). 24. Edwards (2002).

1269-07 CH07 04/18/03 17:26 Page 122

122

    

four-year-olds is approximately $1.4 billion.25 If the expenditures from all these federal and state programs are totaled, approximately $6.4 billion currently is spent on four-year-olds under existing preschool and child care programs. This amount does not include private monies spent on care.    . Mitchell, Stoney, and Dichter estimate that U.S. federal, state, and local governments jointly contribute 39 percent of the revenue used to finance child care and ECE, families contribute approximately 60 percent, and businesses and philanthropic organizations supply the remaining 1 percent.26 We calculate that parents as a group pay 54 percent of the cost of preschool and child care for four-year-olds. We arrived at this number by first estimating child care use and expenditures in 1997, using data from the fourth wave of the 1996 Survey of Income and Program Participation as summarized by Smith.27 Using the reported average weekly cost of $67 for preschoolers receiving some paid child care, we estimate that after excluding assistance received from federal or state governments, annual parent expenditures are approximately $3,726 (2001 dollars). Smith reports that 51 percent of preschoolers receive some paid child care. By applying this percentage to the population of approximately 3.9 million four-year-olds, we estimate that total parent out-of-pocket expenditures for four-year-olds were more than $7.4 billion (2001 dollars). This exceeds the estimated $6.4 billion spent on four-year-olds under the previously discussed federal and state programs.28

State Early Childhood Education Systems In recent years, states have expanded both the availability and scope of services available to four-year-olds, although these programs do not have anything like the scope of ECE systems in many other developed countries. States develop their own programs and strategies for providing services to meet needs that are beyond the reach of currently available federal programs. 25. To arrive at this estimate we used data from Schulman, Blank, and Ewen (1999) that describe the ages served by state prekindergarten programs. For programs that served children of ages other than four, we evenly distributed program costs among the ages served. 26. Mitchell, Stoney, and Dichter (2001) note that the data, from Stoney and Greenberg (1996), are for 1995. They argue that although public expenditures have increased since 1995, so too has the overall size of the industry, likely leaving the proportions unchanged. 27. Smith (2002). 28. The estimate of average weekly child care may overstate the actual costs for four-year-olds if child care for younger children is relatively more expensive. If, on the other hand, a larger proportion of four-year-olds receive paid child care and ECE, our estimate of total expenditures will understate the true amount.

1269-07 CH07 04/18/03 17:26 Page 123

   --

123

In doing so, states typically use one of several models, including universal preschool models, Head Start expansions, community partnership models, or kindergarten for four-year-olds. Although this typology defines the nature of services provided, the scope of state programs varies within and across categories along a number of dimensions. Some universal preschool models (New York’s, for example) aim to make some care available to all children but allow programs to provide services for only a few hours a day. Others offer extended-day programs, but for only a portion of the week or year. Still others offer full-year and full-time programs, but only to a target population, which usually is defined by family income.29 As Education Week notes, in total, thirty-nine states and the District of Columbia offer and financially support a prekindergarten or early childhood education program that targets children ages three to five.30 Although in most cases only a subset of children are eligible to receive such services, these programs represent a marked increase in the availability of ECE since 1980, when approximately ten states offered state-financed programs for their preschoolers. As an example, we describe Georgia’s universal prekindergarten program, one of four programs that attempt to provide ECE to all four-yearolds whose parents want them to participate.31 Georgia’s system closely matches the idea of a universal preschool for four-year-olds. The Georgia Voluntary Pre-K program offers care to all four-year-olds in the state regardless of income and requires care to be available for a minimum of 6.5 hours a day, five days a week, for the school year (180 days).32 The program is financed by revenues from a state lottery established for this purpose. In the first year, 1992, GPK was linked to low-income families and served only 750 four-year-olds. By 1994, the program had grown to serve 15,500 children from low-income families at an average per-child cost of $5,032. Growth in lottery funding spurred this expansion and also eliminated the local matching requirement. In 1995, revenues from the Lottery for Education continued to exceed expectations and the state modified GPK to include all four-year-olds in the state, regardless of family income. By 1999–2000, GPK served 63,000 children, using approximately $225 million in proceeds 29. These differences have been widely discussed in the literature and summaries such as Edwards (2002) and Schulman, Blank, and Ewen (1999), which provide detailed surveys of current and past state policy. 30. Edwards (2002). 31. New York, Oklahoma, and the District of Columbia also are in the process of providing such care. 32. For details see Schumacher, Greenberg, and Lombardi (2001).

1269-07 CH07 04/18/03 17:26 Page 124

124

    

from the lottery. The Finance Project reports that in the year 2000 Georgia enrolled approximately 70 percent of four-year-olds in publicly subsidized preschool, including GPK and Head Start.33 Classes are required to have a staff-to-child ratio of one to ten, with a maximum of twenty children. Curriculum is set at the state level, although localities may choose from a set of state-approved models adopted from national proposals. The state has regulations regarding training: lead teachers must have a college degree in a field related to ECE, certification in early childhood or elementary education, or a technical institute or two-year associate’s degree. Georgia allocates annual funds to providers through ten monthly payments covering September through June. These grants vary by location, teacher credentials, and number of children served, and they may include a one-time payment of $8,000 in start-up funds for new classrooms. Statewide, the average allocation per child was $3,580 in 2000.34 Georgia also uses federal funds available for ECE, including CCDF and Head Start. GPK funds are used to provide extended day care for children in Head Start.

Our Plan for Financing a Universal Preschool To succeed in creating and maintaining a universal prekindergarten program for all four-year-olds, financing must be available. To provide financing, we would like a reliable system that is not too burdensome for lower-income parents; that recognizes preschool as an investment in children’s human capital that has both private and public benefits; that is fairly easy to administer; and that would generate sizable amounts of money. One option to consider seriously would be to expand elementary school education to four-year-olds and add an extended child care component. Such an expansion might build on models similar to those that currently exist in the District of Columbia, Oklahoma, and Texas. This option would be financed by an expansion of existing sources of revenue for public schools, such as property taxes and general revenue. The advantages of this approach are simplicity, little added administrative cost, and the ability to get started quickly.

33. Fisher, Cohen, and Flynn (2000). 34. The state also provides limited funding for transportation to and from the program. However, this is only for low-income children, and the state caps funding at $150 per family per year.

1269-07 CH07 04/18/03 17:26 Page 125

   --

125

There are, however, several disadvantages to this proposal: the potentially higher costs of the program if it is combined with public schools, including union restrictions that may make it difficult to hire aides at salaries lower than those of teachers; restrictions on the use of aides as the sole providers of the child care or wrap-around component; the difficulty of raising more revenue from existing sources; and the failure to address the extent to which parents of four-year-olds should share in the cost.35 Other disadvantages relate to issues of time and space. Making the preschool program a year-round program could be difficult in a school setting. And finding space in current educational facilities may also be a challenge in many communities, especially in light of current initiatives to reduce class size.36 As an alternative to expanding the existing educational system to include four-year-olds, we suggest a financial approach that includes existing public sector subsidies for child care and ECE and adds parent contributions based on ability to pay and collected over time to prevent unnecessary hardship. Our financing approach has two policy goals. First, we seek to design a program that provides the developmental advantages of good-quality ECE to all fouryear-olds, the overall goal of our proposal. Second, recognizing the reliance on general revenues in other proposals, we wish to design a proposal with a specific financing mechanism to support the program. Because of the current budgetary necessity to keep program costs down, we seek to include parents to defray the costs, but in a way that reflects their ability to pay. Our scheme has the advantages that it could be easily modified to increase its revenuegenerating capacity and expanded in states that wished to enhance their preschool program. The basic idea is to finance a major share of the cost of a universal preschool program for four-year-olds by having parents pay according to their ability, as measured by their average federal tax rate over a ten-year period. One might think of this as an increase in the average tax rate paid by parents of each four-year-old for a multiple-year period. If we use the tax system’s progressive rates as the basis of payments, parents with lower

35. As we see it, the issue of teacher pay scale is one defined in terms of paying assistant teachers. Using the existing pay scale for the lead teacher in each prekindergarten classroom is consistent with the desired quality of ECE programs. However, using assistants who have less formal training is likely to be cost-effective and their market wages are likely to be considerably below the existing teacher pay scale. 36. One possible solution to these difficulties involves a trade: offering a program that includes prekindergarten and ends at grade 11 in districts in which adding a year is not possible. Some policymakers might be willing to consider this, but we do not pursue this idea here, though we note it as an option.

1269-07 CH07 04/18/03 17:26 Page 126

126

    

incomes and lower taxes will pay less for preschool than will higher-income parents. For the lowest-income parents, revenue will be raised by reducing the subsidy rate of the earned income credit (EIC). In our plan, we cap the amount that parents would pay so that the program would attract the children of higher-income as well as middle- and lower-income parents. For parents of four-year-olds who are enrolled in the preschool program, we propose a plan with the following parameters: —Increase the family’s average tax rate by 1 percentage point. —Reduce the EIC subsidy by 2.5 percentage points (from 40.2 to 37.7 percent).37 This would lower the maximum EIC benefit from $3,816 to $3,579. We allow EIC benefits to phase out completely at the current amount, but because we reduce the maximum benefit, the new effective phase-out rate decreases to 19.5 percent. Other parameters of the EIC remain as they are. —In each year, calculate a family’s payment according to either their federal average tax rate or a reduction in their EIC payment. —Parents who sign up for this program would not be eligible directly for federal aid from the two major federal programs, the Child Care and Development Fund and the dependent care tax credit, during the year the child is in the preschool program. (As encouragement to enroll in the program, parents who do not enroll their child could claim only costs for child care that are above the average costs of the program under the dependent care tax credit or any other tax subsidy for child care or preschool.) —Payments would be made over ten years; the cap or maximum payment per year would be $800. This cap is designed to limit the amount contributed by families with greater incomes, in order to increase the incentive for them to participate. In addition, parents would pay a one-time annual fee of $150 if their income was less than $35,000, but double that, or $300, if their income was above $35,000. Using a 3 percent real discount rate, this plan would raise $4,624 per child, using the 2000 income distribution.38 So, for example, if 85 percent of all four-year-olds were enrolled, and the probability of enrollment was uniform across income groups, the plan would raise $15.43 billion dollars. If a family’s income increases over the ten years, the amount collected for the child will increase, and if income decreases, the amount collected will decrease. Annual amounts collected per family would range from $125 for those in the 37. Under this scheme, approximately 630,000 parents receiving EIC would shift to face an increase in their federal average tax rate. 38. U.S. House of Representatives, Committee on Ways and Means (2000).

1269-07 CH07 04/18/03 17:26 Page 127

   --

127

lowest income group to the $800 maximum.39 The net present value of the ten years of additional taxes would range from $1,066 to $6,824 per family. Table 7-2 presents the full set of the amounts that would be collected, by family income category. To improve on these estimates of the distribution of employed parents by income class, we simulate the revenue-raising capacity of our approach by using data from the March 1999, 2000, and 2001 Current Population Survey, which allows us to look at income among families of four-year-olds rather than all families.40 CPS data also allow us to estimate the proportion of all four-year-olds who live in families in which one parent, or both parents, work. We find that of a total of 3,926,300 four-year-olds, 85.7 percent (3.33 million) have at least one employed parent and 56 percent (2.20 million) live in families in which both parents work. By incorporating the income distributions for these two groups we can recalculate the total revenue that could be raised and per-child spending capacity. If all four-yearolds from families with any employed parents participated, our proposal would raise $15 billion, or approximately $4,400 per child. If we look only at those four-year-olds from families in which both parents work (the most likely to require child care), we estimate that our proposal would yield approximately $9.7 billion, or $4,392 per child. To these revenues raised from parents, we also would add funds currently devoted to subsidizing child care for four-year-olds. They would include the funds from Head Start, from CCDF, and from Temporary Assistance for Needy Families for this age group and the value of tax expenditures under the various tax credit policies, including those paid to employers. Our estimate is that these funding sources currently provide about $6.4 billion.

39. We also estimate the increase in taxes paid under our plan as a percentage of tax payments made under the existing tax structure. Using effective tax rates provided by the IRS (2001a and 2001b), we estimate that total taxes paid as a percent of income increase by 2 to 11 percent (excluding the change for the $25,000 group, which faces an average tax rate quite close to zero), with higher income individuals facing the smaller increase because of our $800 expenditure cap. Note that the average tax rates reported by the IRS are computed either over all returns (non-EITC cases) or for married couples with two dependents (EITC cases) and may not reflect the actual rates faced by parents of four-year-olds. 40. We use three years’ worth of data to maximize the available sample of four-year-olds. We increase the sample to approximately 5,700 four-year-olds (there are fewer than 2,000 in each individual year). We adjust CPS person weights to reflect the number of four-year-olds reported in U.S. Census Bureau (2002), which is based on the 2000 Census, and divide each weight by three to reflect the fact that we are using three years of data. Like the estimates using data from the U.S. House of Representatives, Committee on Ways and Means (2000), the one-percentage-point increase in the average tax rate is applied to total family income.

9.2 12.4

6.8 7.3

10.6 12.0

250 2,133 –0.6 1 1.6

25,000

11 11.6

350 2,986 8.9 1 1.9

35,000

10.9 10.5

450 3,839 10.0 1 1.7

45,000

23.8 21.3

625 5,311 11.1 1 1.5

62,500

13.1 11.4

800 6,824 13.7 0.9 1.3

87,500

11.5 10.9

800 6,824 17.4 0.5 0.7

150,000

3.1 2.7

800 6,824 24.0 0.4 0.6

200,000

Source: Authors’ calculations. a. Except as otherwise noted, distribution is based on 2000 Green Book data (U.S. House of Representatives, Committee on Ways and Means 2000) and tax data from the IRS. b. Based on CPS data averaged over 1998, 1999, and 2000.

205 1,749 –24 1.4 2.4

125 1,066 –40 3 5

Annual tax increase (dollars) NPV of increases (dollars) Average tax rate (percent) Tax increase as percent of income Tax increase plus fees as percent of income Percent of all 4-year-olds Percent of 4-year-olds with both parents employedb

15,000

5,000

Item

Mean earnings (dollars) a

Table 7-2. Parameters of Wolfe-Scrivner Proposal for Financing Universal Preschool for Four-Year-Olds

1269-07 CH07 04/18/03 17:26 Page 128

1269-07 CH07 04/18/03 17:26 Page 129

   --

129

Would high-income families enroll at a “price” of $6,824? We do not know; however, we note with interest that Sawhill and Thomas estimated that the average cost of child care for affluent families in 1998 was very close to $6,824.41 That amount, of course, could be for more than one child. As noted above, the system could be readily modified to change the revenue stream. For example, if the one-time fees were raised to $400 for those with incomes above $35,000, the amount raised per child would increase to about $4,700. A further increase in the cost, equal to a 0.1 percentage point increase in families’ average tax rate, would raise the amount available per child to $4,921. What about a four-year-old living with a single parent who does not work? Are such children eligible? We suggest that they are, since we are concerned with the well-being of children and there are social benefits from getting such children ready for school. The parent would be required to pay the first-year fee and to agree to the payment schedule noted above. In lieu of other current payments, such parents would be required (subject to their ability) to spend some hours working in activities such as preparing food or substituting as a teacher’s aide. We believe that this approach to providing for universal preschool would cover the average costs of a quality program. Although it is difficult to assess whether the quality would be excellent, good, or adequate, several sources are useful guides on this topic. The Committee on Economic Development uses data from the Cost, Quality, and Outcomes study to examine average costs among the child care centers with Early Childhood Environmental Rating Scale (ECERS-R) ratings of 5 (“good”) and above.42 They find an average cost of care of $3.22 per hour (in 2001 dollars) in these programs and suggest that the cost of a four-hour-day, school-year program would be approximately $2,300 per year per child. Extending this estimate to a full-day, fullyear program increases annual costs, up to $6,700 per child if we assume that a full 2,080 hours of care is needed. Marshall and others investigated the quality and cost of child care in Massachusetts using data from a random sample of eighty-eight community-based child care centers that served preschool children on a full-time, full-year basis.43 The authors find that process quality in Massachusetts, as measured by the ECERS-R, is “good,” with an average score across categories of 4.94. Comparing that score to those in four other states (which averaged 41. Sawhill and Thomas (2001). 42. Committee on Economic Development (2002). 43. Marshall and others (2001).

1269-07 CH07 04/18/03 17:26 Page 130

130

    

from 3.82 to 4.49) as observed by Helburn, they note that Massachusetts provides relatively high-quality child care.44 The authors also collected data on center revenues and expenditures and found that average revenues per hour of child care were $3.71 and expenditures were $3.41. By defining full-time, full-year child care as care provided for forty-five hours a week and fiftytwo weeks a year, they estimate average annual per-child center revenues and expenditures of $8,681 and $7,979. Other sources suggest that care may be less costly. Schulman estimates that the cost of full-day, full-year child care for four-year-olds ranges between $4,000 and $6,000.45 Helburn and Bergmann present cost estimates that suggest that average per-child costs for four-year-olds under a free, universal child care program would be approximately $5,000.46 Georgia provides at least 6.5 hours of quality universal care to its four-year-olds for the duration of the school year at a cost of approximately $3,500 per child. Extending such a program to meet a full-day, full-year schedule would likely result in costs in the same $4,000–$6,000 range. Our proposal’s revenue-raising capacity falls in the middle of this range, although if the savings from the current CCDF, Head Start, tax credit, and state funds currently used to meet demand for care are made available to this program, an additional $1,630 per child could be used to increase quality or meet other possible needs, which we discuss below.47 Under existing law, funds from CCDF must be used for children from lower-income families, which, research has suggested, have the most to gain from attending a highquality prekindergarten. Additional resources might be required, for example, if new facilities are needed, extensive changes are made to existing facilities, or transportation costs must be covered. A portion of the savings from reduced use of other programs could be used to subsidize the program in order to provide for new facilities, as well as to improve the quality of care. These funds may be required for some of the operating costs as well, if a smaller proportion of higher-income than lower-income families sign up for this program. 44. Helburn (1995). 45. Schulman (2000). 46. Helburn and Bergmann (2002) present a model of reforming the nation’s child care system that includes several scenarios, ranging from expanding funding so that it is sufficient to serve all children eligible under the current CCDF model, to providing universal, free, higher-quality care to all children. The estimated per-child costs reflect the latter scenario and the additional assumption that approximately 25 percent of care would be provided in settings of higher-than-average quality (ECERS score of 5 or greater). 47. Given an 85 percent participation rate, we estimate the total savings available for reallocation at $5.4 billion. Dividing by the number of participating four-year-olds yields the per-child amount.

1269-07 CH07 04/18/03 17:26 Page 131

   --

131

To this point we have not specifically discussed the collection of the parent’s share of funding for this preschool program. We believe that the agency that can collect the fees and payments at the lowest administrative cost should do so, whether it is the Department of Education, the Department of Health and Human Services, or even the IRS. An argument in favor of the IRS is that the ability-to-pay system is tied to the tax system. In this arrangement, the IRS administrative system could be modified to track participation, distribute payments for child care, and later track and collect parent payments.48 As a second option, if the tax-based system faced political or bureaucratic opposition, the program might be presented instead as a very low (or zero) interest loan program, perhaps called the Ability to Pay Loan Plan (APLP). Under this variation, the underlying mechanics and parameters would not change, although the program would be explained to parents in the context of loans and repayment (set as a fixed percentage of income) rather than as a change in tax policy. This recasting would be transparent to administrators, but it might increase the option’s attractiveness to policymakers and parents.49 Our suggested preschool financing program would replace the patchwork of federal and state programs for funding child care and provide the resources for a universal program of preschool for all four-year-olds. As detailed in the next section, such a program is not unusual in developed countries.

International Perspectives: Early Childhood Education in Other Developed Nations The U.S. system of early childhood education available to four-year-olds falls short of that offered in other developed countries and, in most cases, does so substantially. Although an in-depth investigation of international early childhood education systems is beyond the scope of this chapter, a comparison of the U.S. system with those in other developed nations is suggestive, and its most important lessons can be quite succinctly laid out. Kamerman reports that international systems of ECE for children between the ages of three and five years increasingly are becoming universal and often 48. The IRS rarely serves such an administrative function. An exception was the collection of supplementary taxes calculated on federal income tax liability under the Medicare Catastrophic Coverage Act (P.L. 100-360), which was passed in 1988 but repealed one year later after higherincome Medicare beneficiaries protested new premiums. 49. Another variation, which is based on the underlying idea that the program plays both educational and work-support roles, might assign financial responsibility for the child care portion of the day to parents, while the three to four hours per day of instruction would be paid for by the school system, which could use a lottery, special tax, or other financing instrument to finance it.

1269-07 CH07 04/18/03 17:26 Page 132

132

    

provide full-day, full-year services.50 Countries such as Belgium, France, and Italy provide care to preschool children through education-based programs that serve between 95 and 99 percent of children ages three to six. These programs run between seven and eight hours a day at no cost to parents, offering extended-day services to parents who choose to pay the fees, which are determined by their income. The Scandinavian countries of Denmark, Finland, and Sweden provide ECE programs that more closely resemble a support system for working parents; care covers the full workday and runs year round. Although these Scandinavian countries tend to focus on providing services to children from families with working adults, they nevertheless provide coverage for the majority of three- to five-year-olds because of the high rate of labor force participation in these countries. The programs currently cover between 73 and 83 percent of preschool children, and the Scandinavian countries all have recently committed to providing universal coverage. Kamerman notes that the European countries that do not currently offer coverage view the provision of care as a priority and have begun initiatives to transform their programs. Take-up is quite high (at or only slightly below 100 percent of eligible families) in these countries, and where care is not universally available waiting lists usually exist. In its broad survey of international ECE practices, the Organization for Economic Cooperation and Development (OECD) also finds that most European countries currently provide or are committed to moving toward providing universal care on a scale that outstrips current U.S. efforts.51 ECE programs in most OECD countries are publicly delivered and are financed through public sources. Most often services are provided in publicly funded facilities built for the purpose and often are located near or in primary schools.

Conclusions Both candidates in the last presidential election favored education and suggested support for beginning schooling at an earlier age or increasing federal commitments to ECE, including Gore’s proposed $50 billion commitment to universal prekindergarten and Bush’s emphasis on expanded Head Start and improved ECE quality. Providing universal preschool for four-year-olds is an idea whose time has come. Most schooling is delivered and paid for at the local level in the United States and in time the public may insist that preschool 50. Kamerman (2000). 51. Organization for Economic Cooperation and Development (2001).

1269-07 CH07 04/18/03 17:26 Page 133

   --

133

be financed in the same way. In the interim, greater federal financial assistance combined with voluntary parent contributions based on ability to pay could encourage more states to provide such schooling and help to demonstrate the value of providing all children the skills they need to succeed.

References Blau, David. 2001. The Child Care Problem: An Economic Analysis. Russell Sage. Children’s Defense Fund. 2001. State of America’s Children: 2001. Washington. Committee for Economic Development. 2002. Preschool for All: Investing in a Productive and Just Society. Washington: Research and Policy Committee. Currie, Janet. 2001. “Early Childhood Education Programs.” Journal of Economic Perspectives 15 (2): 213–38 (www.econ.ucla.edu/people/papers/Currie/Currie149.pdf [January 7, 2003]). Edwards, Virginia B. 2002. “Building Blocks for Success: State Efforts in Early Childhood Education.” Education Week report (January 10, 2002). Fisher, Hansine, Carol Cohen, and Margaret Flynn. 2000. Federal Funding for Early Childhood Supports and Services: A Guide to Sources and Strategies. Washington: The Finance Project. Garces, Eliana, Duncan Thomas, and Janet Currie. 2000. “Longer-Term Effects of Head Start.” Working Paper 8054. Cambridge, Mass.: National Bureau of Economic Research (December) (http://papers.nber.org/papers/w8054.pdf [January 7, 2003]). Heckman, James J. 2000. “Policies to Foster Human Capital.” Research in Economics 54 (1): 3–56. Helburn, Suzanne, ed. 1995. Cost, Quality, and Child Outcomes in Child Care Centers, 2d ed. Technical Report. Economics Department, University of Colorado at Denver. Helburn, Suzanne, and Barbara Bergmann. 2002. America’s Childcare Problem: The Way Out. New York: Palgrave. IRS. 2001a. Individual Income Data Returns: 1999. Washington: Internal Revenue Service, Statistics of Income Division. ———. 2001b. Statistics of Income Bulletin. Publication 1136, table 3. Washington: Internal Revenue Service, Statistics of Income Division. Joint Committee on Taxation. 2002. Estimates of Federal Tax Expenditures for Fiscal Years 2002–2006. Washington: U.S. Government Printing Office (January). Kamerman, Sheila. 2000. “Early Childhood Education and Care: An Overview of the Developments in the OECD Countries.” International Journal of Educational Research 33 (1): 7–30. Karoly, Lynn, and others. 1998. Investing in Our Children: What We Know and Don’t Know about the Costs and Benefits of Early Childhood Interventions. Report MR-898TCWF. Santa Monica, Calif.: RAND. Lee, Valerie E., and David T. Burkam. 2002. Inequality at the Starting Gate: Social Background Differences in Achievement as Children Begin School. Washington: Economic Policy Institute. Marshall, Nancy, and others. 2001. The Cost and Quality of Full-Day, Year-Round Early Care and Education in Massachusetts: Preschool Classrooms. Wellesley Centers for Women and Abt Associates Inc.

1269-07 CH07 04/18/03 17:26 Page 134

134

    

Mitchell, Anne, Louise Stoney, and Harriet Dichter. 2001. “Financing Child Care in the United States.” Kansas City, Mo.: Ewing Marion Kauffman Foundation. National Center for Education Statistics. 2002. Digest of Education Statistics: 2001. NCES 2002-130. Washington: U.S. Department of Education, Office of Educational Research. National Center for Children in Poverty. 2000. Map and Track: State Initiatives for Young Children and Families. New York (December). NICHD Early Child Care Research Network. 2000. “The Relation of Child Care to Cognitive and Language Development.” Child Development 71 (4): 960–80. ———. 2002. “Early Child Care and Children’s Development Prior to School Entry: Results from the NICHD Study of Early Child Care.” American Educational Research Journal 39 (Spring): 133–64. Olson, Lynn 2002. “Quality Counts Finds Uneven Early-Childhood Policies.” Education Week 21 (January 9, 2002): 3 (www.edweek.org/ew/newstory.cfm?slug=16qc.h21 [January 7, 2003]). Organization for Economic Cooperation and Development. 2001. Starting Strong: Early Childhood Education and Care (www1.oecd.org/publications/e-book/9101011e.pdf [January 7, 2003]). Peisner-Feinberg, Ellen, and Margaret Burchinal. 1997. “Relations between Preschool Children’s Child-Care Experiences and Concurrent Development: The Cost, Quality, and Outcomes Study,” Merrill-Palmer Quarterly 43 (3): 451–77. Peisner-Feinberg, Ellen, and others. 1999. The Children of the Cost, Quality, and Outcomes Study Go to School: Executive Summary. University of North Carolina at Chapel Hill, Frank Porter Graham Child Development Center. Pianta, Robert C., and others. 2002. “The Relation of Kindergarten Classroom Environment to Teacher, Family, and School Characteristics and Child Outcomes,” Elementary School Journal 102 (3): 225–38. Reynolds, Arthur J., and others. 2002. “Age 21 Cost-Benefit Analysis of the Title I Chicago Child-Parent Centers” (www.ssc.wisc.edu/irp/pubs/dp124502.pdf [January 7, 2003]). Sawhill, Isabel, and Adam Thomas. 2001. “A Hand Up for the Bottom Third: Toward a New Agenda for Low-Income Working Families.” Paper. Brookings (May). Schulman, Karen. 2000. The High Cost of Child Care Puts Quality Care out of Reach for Many Families. Report. Washington: Children’s Defense Fund. Schulman, Karen, Helen Blank, and Danielle Ewen. 1999. Seeds of Success: State Prekindergarten Initiatives 1998–1999. Washington: Children’s Defense Fund. Schumacher, Rachel, Mark Greenberg, and Joan Lombardi. 2001. State Initiatives to Promote Early Learning: Next Steps in Coordinating Subsidized Child Care, Head Start, and State Prekindergarten. Washington: Center for Law and Social Policy (April). Shonkoff, Jack P., and Deborah A. Phillips, eds. 2000. From Neurons to Neighborhoods: The Science of Early Childhood Development. Washington: National Academy Press. Smith, Kristin. 2002. Who’s Minding the Kids? Child Care Arrangements: Spring 1997. Current Population Reports P70-86. U.S. Census Bureau (July). Stoney, Louise, and Mark Greenberg. 1996. “The Financing of Child Care: Current and Emerging Trends.” Future of Children 6 (2) (www.futureofchildren.org/information2826/information_show.htm?doc_id=73284 [January 7, 2003]). U.S. Census Bureau. 2002. Summary File 1: 2000 Census of Population and Housing SF1/05 (RV).

1269-07 CH07 04/18/03 17:26 Page 135

   --

135

U.S. Department of Health and Human Services. 2000. “Helping Families Achieve SelfSufficiency: A Guide on Funding Services for Children and Families through the TANF Program, Administration for Children and Families.” (www.acf.dhhs.gov/programs/ofa/funds2.htm [January 7, 2003]). ———. 2001. FY1999 CCDF Data Tables and Charts. Washington: Administration for Children and Families (www.acf.dhhs.gov/programs/ccb/research/99acf800/ages.htm [January 7, 2003]). ———. 2002. FY2000 CCDF State Expenditures. Washington: Administration for Children and Families (www.acf.dhhs.gov/programs/ccb/research/00acf696/ overview.htm [January 7, 2003]). U.S. General Accounting Office. 2000. “Title I Preschool Education: More Children Served, but Gauging Effect on School Readiness Difficult.” GAO/HEHS-00-171. U.S. House of Representatives, Committee on Ways and Means. 2000. 2000 Green Book. Washington: U.S. Government Printing Office.

1269-08 CH08 04/18/03 17:26 Page 136

8 Improving Neighborhoods for Poor Children JENS LUDWIG

I

n 1990 approximately 8 million Americans were living in highpoverty neighborhoods, almost double the number in 1970.1 New evidence suggesting that “place matters,” at least in some circumstances for some families, implies that policymakers concerned about the life chances of children living in these highpoverty areas should think hard about helping their families move to more affluent neighborhoods. Specifically, short-term findings from a randomized housing-voucher experiment known as Moving to Opportunity (MTO) suggest that moving poor children from high- to low-poverty areas may substantially improve the mental and physical health of their parents (as well as the quality of parenting), improve the health of the children themselves, reduce their involvement in antisocial behavior, and increase their academic achievement. These results are far from the last word on the question of neighborhood effects—long-term results from the experiment are not yet in, and most of the findings so far

1. High-poverty neighborhoods are defined here as census tracts with poverty rates of 40 percent or more (Jargowsky 1997).

136

1269-08 CH08 04/18/03 17:26 Page 137

    

137

come from studying only a few of the demonstration’s five sites. Nevertheless, even these preliminary findings suggest that improving the life prospects of children may require at least some attention to the neighborhoods in which they reside. In this chapter I review the different ways in which neighborhoods—and by extension place-based interventions—may affect the outcomes of both poor and nonpoor children, presenting the best available evidence. I outline a variety of policies that might help improve the neighborhoods and life chances of poor children and include a discussion of their practical and political limitations.

Why Might Neighborhoods Matter? Sociologists, psychologists, and economists have developed a variety of arguments for why neighborhood attributes may affect individual behavior. Most of these emphasize the importance of the interactions among neighbors; some types of interactions, for example, are believed to yield prosocial outcomes. However, it also is possible that place matters because of specific attributes of the neighborhood rather than of its residents, including the quality of local public schools, distance to places where jobs are available, or the effectiveness and fairness of local law enforcement—attributes that in principle could be changed without moving families.

Peer Behaviors Casual empirical observation suggests that children and adults are influenced by the behaviors of their peers. Social scientists have offered a number of explanations for why this might be so.2 One possibility is that children simply may not have thought of certain behaviors (such as body piercing or sniffing glue) before those behaviors were modeled for them by their friends or on television. Rigorous empirical evidence on the existence of the modeling behavioral mechanism is scarce, but anecdotes are plentiful. For example, during the World Cup in 1998, when concern about soccer hooliganism was running high in France, officials in one town canceled a medieval jousting festival for fear that it “might give people ideas.” After a “humanbarbecue” stunt on MTV’s extreme stunt show Jackass, newspapers were

2. Cook and Goss (1996).

1269-08 CH08 04/18/03 17:26 Page 138

138

 

filled with stories of young men who voluntarily had themselves set on fire (and videotaped) by their friends.3 Second, children may also conclude that some behaviors are acceptable or desirable simply because others within their neighborhood or school engage in them. A child might, for example, observe a schoolmate shoplifting from a local store and, assuming that the shoplifter has better information than he himself does, infer that shoplifting must yield high returns at low risk. If the child then chooses to shoplift as a result, it sends a signal to other local children, engendering more shoplifting, more signaling, and so forth. The result may be an “information cascade,” in which almost everyone in a neighborhood, school, or firm engages in some behavior even though almost no one understands the actual risks and rewards. A third possibility is that the payoffs to a given behavior change with its prevalence. For example, if police resources within a given community are relatively fixed, at least in the short run, an increase in criminal behavior by a given individual’s neighbors may decrease the chance that the individual will be caught because the increase in crime reduces the police-hours available per crime (“resource swamping”). The prevalence of a behavior also could affect the social returns to the activity. In the 1980s, when “Madison” was the 539th most popular girl’s name in the United States, most parents who wished to avoid complicated or awkward introductions at the local playground chose more common names. Of course, the nonconformists who named their daughters Madison in the 1980s presumably would have been less likely to do so if their children had been born instead in 2001, when Madison was the second-most-popular girl’s name in America.4 There also is across-area as well as over-time variation in the social rewards to selecting certain children’s names (and perhaps other behaviors as well): more Arkansans probably share first and middle names with local son Billy Bob Thornton than they do with U.S. Representative Martin Olav Sabo of Minneapolis, while the reverse is probably true in the Twin Cities.5 3. See, for example, “News Roundup,” San Antonio Express-News, October 5, 2002, p. B2; Marisa Taylor, “Four Plead Guilty in Mess-Hall Mess,” San Diego Union-Tribune, May 11, 2002, p. B7; and “Boy Seriously Hurt in Fire Stunt,” The Guardian (London), January 9, 2002, p. 8. 4. Virginia Postrel, “Economic Scene: How Can the Marketplace Gauge Fashions? Consider What to Name the Baby,” New York Times, May 23, 2002, p. C2. 5. A famous laboratory experiment demonstrates the existence of either information cascades or the desire of individuals to conform to the social expectations of others. In the experiment, subjects were willing to claim that the shorter of two lines was in fact longer when enough of the investigator’s confederates (who posed as other experimental subjects) made the same claim (Deutsch and Gerard 1955).

1269-08 CH08 04/18/03 17:26 Page 139

    

139

These four mechanisms—modeling, information cascades, resource swamping, and social norms—emphasize the role played by the attitudes and actions of one’s peers and raise the possibility that small changes in a neighborhood environment could lead to large changes in the behaviors of local children. Both the resource-swamping and information-cascade mechanisms have a self-reinforcing character, so that the initial actions of a small number of youth can “tip” the neighborhood from being a stable one with few social problems to being one in which everyone susceptible to engaging in antisocial behavior does so. It also is possible that the social norms associated with some behavior also tip when the behavior exceeds some threshold prevalence. In schools where social drinkers are a small minority, drinking may be frowned upon; once drinkers become a critical share of the student body, however, the social climate may shift from dry to wet. Alternatively it may be the attitudes and behaviors of a community’s adults rather than its children that matter. Sociologists emphasize the importance of the strength and breadth of social ties among the adults within a community and the degree to which they are willing to work together to support shared prosocial norms.6 Children in such neighborhoods may feel encouraged to stay in school, or they may be concerned that a neighbor will report their misbehavior to their parents or the police. Finally, the processes described above all suggest that moving from a high- to a low-poverty area should improve children’s life chances; however, in principle at least, moving could have a detrimental effect on their chances instead. For example, Christopher Jencks and Susan Mayer observe that in many settings children and adults compete for resources and rewards.7 Poor children who are above-average students on the South Side of Chicago may be below average in academically elite suburban high schools, which might dampen their enthusiasm for trying hard in school.

Neighbors’ Characteristics Neighbors may matter for a child not because of what they do to or for the child as much as who they are or what they represent as role models. For example, William Julius Wilson has argued that children growing up in high-poverty neighborhoods drop out of school in part because the lack of educated adults within their communities makes it hard for teens to understand the advantages of more schooling.8 Similarly, minority children who 6. Sampson, Raudenbush, and Earls (1997); Sampson, Morenoff, and Gannon-Rowley (2002). 7. Jencks and Mayer (1990). 8. Wilson (1987); see also Ludwig (1999).

1269-08 CH08 04/18/03 17:26 Page 140

140

 

grow up in segregated neighborhoods may conclude that the institutions of American society are racist and that any returns they get from prosocial behavior will be minor.

Neighborhood Institutions Finally, it may be that the behaviors or characteristics of a neighborhood’s residents are not as important as the institutional features of the neighborhood itself.9 The distance of the suburbs, with their better job opportunities, from inner-city neighborhoods may be an important impediment to improving the earnings of less skilled urban workers.10 And on average homebuyers believe that the quality of a variety of local public services (particularly schools) vary systematically across neighborhoods, as evidenced by the systematic relationship between house prices and a variety of school attributes.11 What are the implications of this discussion for efforts to improve the life chances of poor children by improving their neighborhoods? First, knowing the specific mechanism(s) responsible for neighborhood effects on children is important in developing policies that might help children. For example, if the behaviors of neighbors and peers matter most, then investments in housing or schooling that have no effect on those behaviors will have little impact. On the other hand, if inadequate policing or poor schools are the problem, then more direct efforts to improve them could make a difference. Second, small changes could in principle matter a lot for children if their propensity to engage in a particular pro- or antisocial behavior “tips” as the prevalence of certain neighborhood behaviors or characteristics rises. This observation has important implications for the societywide effects of any mobility program. If, for example, a 10 percent increase in neighborhood poverty rates always yields a 10 percent increase in school dropout rates, then moving poor children from high- to low-poverty neighborhoods will reduce the dropout rate among movers, increase dropouts among residents of the destination neighborhood, and leave the overall dropout rate in the society as a whole unchanged. On the other hand, if the dropout rate in a community has a much larger impact on a child’s likelihood of dropping out when the neighborhood prevalence exceeds some cutoff value, then reallocating poor children across neighborhoods could reduce the dropout rate among movers and in society as a whole. 9. Manski (1993). 10. Kain (1968); Holzer (1991); Raphael (1998); Weinberg (2000). 11. Black (1999); Figlio and Lucas (2002).

1269-08 CH08 04/18/03 17:26 Page 141

    

141

Evidence on Neighborhood Effects Despite the large empirical literature on how neighborhoods affect families, very little reliable evidence has been available to date because most families have at least some degree of choice in where they live and with whom they associate. As a result, when we observe that poor children in the leafy North Shore suburbs of Chicago have superior educational outcomes compared with those of poor children on the city’s high-poverty South Side, we inevitably must wonder whether the differences are due to the effects of the local neighborhood or instead to hard-to-measure family attributes associated with both children’s outcomes and their parents’ residential choices. Fortunately there is now a growing body of evidence from randomized or “natural” housing-mobility experiments that break the link between family preferences and neighborhood context, thereby overcoming the self-selection problem that has plagued earlier research. In this section I review findings from the Gautreaux and Moving to Opportunity housing-voucher programs, which suggest that offering poor families the chance to move from highpoverty public housing communities into low-poverty areas may improve their children’s well-being and developmental outcomes.

The Gautreaux Program In 1966 Dorothy Gautreaux filed suit against the Chicago Housing Authority (CHA) for racial discrimination in how the agency placed families in public housing units. In 1976 the United States Supreme Court ordered the CHA to relocate thousands of African American families from public housing on the city’s West and South Sides to more economically and racially diverse neighborhoods in the metropolitan area (Hills v. Gautreaux, 425 U.S. 284, 306 [1976]). Sociologist James Rosenbaum of Northwestern University collected data on families that were moved to the Chicago suburbs as well as to other parts of the city and compared the educational and other outcomes of the children in these households.12 Rosenbaum’s comparisons of city and suburban movers highlighted substantial differences in the educational outcomes of the two groups. For example, suburban movers appeared to have substantially 12. Families in the Gautreaux program were placed on a waiting list, and those at the top of the list were offered apartments located by program administrators. By some accounts the apartments were identified independently of who was next on the waiting list, and families accepted the first apartment made available to them. If so, then Gautreaux breaks the link between family preferences and residential outcomes that complicates previous studies of neighborhood effects.

1269-08 CH08 04/18/03 17:26 Page 142

142

 

lower dropout rates than city movers (5 percent versus 20 percent), and they were more likely to attend college (54 versus 21 percent).13 Some lingering questions inevitably remain about whether assignment to apartments was in fact “quasi-random” and whether those families that participated in follow-up surveys were representative of the larger program population. Nevertheless, the findings from Gautreaux were sufficiently impressive to motivate the U.S. Department of Housing and Urban Development (HUD) to sponsor a true randomized housing-voucher experiment known as Moving to Opportunity.

Moving to Opportunity The MTO demonstration has been in operation since 1994 in five U.S. cities: Baltimore, Boston, Chicago, Los Angeles, and New York. Eligibility at each site is limited to very low-income families with children that live in public housing or Section 8 project-based housing (publicly financed but privately operated) located within specific high-poverty census tracts. Families that apply for the program are randomly assigned to one of three treatment groups. The experimental group is offered the opportunity to relocate to privatemarket apartments using housing vouchers. These families also are offered housing-search and other assistance from a local nonprofit, although under the MTO program’s design they can use the housing vouchers only in lowpoverty neighborhoods (defined as census tracts with 1990 poverty rates below 10 percent). The comparison group also is offered housing vouchers to relocate to private-market apartments. The relocation choices of families in the comparison group, however, are not constrained by the MTO program design. Families assigned to this group do not receive any additional counseling or other services beyond those provided to other families that participate in the government’s standard housing-voucher program. Families assigned to the control group do not receive any assistance under MTO, although these families do not lose access to the services that they already were receiving. On average, 48 percent of families assigned to the experimental group relocated to low-poverty areas through the MTO program. Since comparison group families were not constrained in their choice of a new neighborhood, it is perhaps not surprising that their relocation rate was higher 13. Rubinowitz and Rosenbaum (2000).

1269-08 CH08 04/18/03 17:26 Page 143

    

143

(60 percent).14 Interestingly, only a small share of the families assigned to the comparison group relocated to the lowest-poverty census tracts (under 10 percent), although compared with the control group this group nevertheless did wind up in more economically and racially diverse neighborhoods. The MTO program thus produced substantial average differences across groups in their post-program neighborhood environments, with both the experimental and comparison groups winding up in neighborhoods with greater economic and racial diversity than those of the controls (although the overall changes were more dramatic for the experimental group). Because families are randomly assigned to each of the three groups, simple comparisons of average outcomes across groups are sufficient to identify the causal effects of offering housing vouchers (and thus the chance to move) to MTO program participants.15 Comparisons of outcomes across MTO treatment groups suggests that offering public housing families housing vouchers may at the very least indirectly affect children by improving the quality of their home environments. Survey data from the Boston MTO site analyzed by Katz, Kling, and Liebman reveal better self-reported physical and mental health among the heads of household in the experimental and comparison groups than among the controls (table 8-1).16 The magnitude of the changes is quite substantial for both groups, suggesting increases on the order of 20 to 30 percent in the proportion of household heads who report good physical or mental health. Survey data from the New York MTO site reveal qualitatively similar changes.17 It is important to remember that by pooling all families assigned to the experimental and comparison groups, regardless of whether a family relocates, we capture the causal effects of offering families the chance to move, which understates the effects of moving on those who actually move through the MTO program. If assignment to the experimental and comparison groups has no impact on those who do not move, the effects of moving can be obtained by dividing the effects of the offer to move by the fraction of families who move. For example, the first row of table 8-1, which comes from 14. While the take-up rates themselves vary across sites, as does the magnitude of the differential between the experimental and comparison group take-up rates, in every city the proportion of families that move is at least somewhat higher in the comparison group compared with the experimentals. 15. In practice most of the published analyses of MTO’s impacts use regression methods to adjust for chance differences across groups in baseline characteristics. As we would expect with random assignment, the estimated “intent to treat” effects typically are not very sensitive to whether this regression adjustment is made. 16. Katz, Kling, and Liebman (2001). 17. Leventhal and Brooks-Gunn (2002b).

1269-08 CH08 04/18/03 17:26 Page 144

 

144

Table 8-1. Effects of MTO Demonstration on Parent Outcomes a Outcome Physical/mental health Overall health good or better Calm and peaceful a good bit of time or more over last 4 weeks Parenting Got angry with childb Criticized/made fun of childb Child obeys school night curfew c Labor marketd On welfare Employed

Control mean .578 .465

2.79 1.52

Experimental group minus controls

Comparison group minus controls

.115** (.048) .107** (.050)

.162** (.056) .138** (.063)

-0.37** -0.14

-0.22 -0.21

.765

.092*

.031

.441

-.070** (.028) -.008 (.027)

-.030 (.034) -.021 (.032)

.460

Sources: Authors’ compilations from sources as follows. Physical/mental health: Katz, Kling, and Liebman (2001) (Boston demonstration site); parenting: Leventhal and Brooks-Gunn (2002) (New York site); labor market: Ludwig, Duncan, and Pinkston (2000) (Baltimore site). *Statistically significant at the 10 percent level; ** statistically significant at the 5 percent level. a. The figures in the table generally reflect the fraction of MTO household heads who report “yes” to a given outcome defined. Standard errors are in parentheses. b. Results are from parents’ reports about frequency of event on a four-point scale, with 1 indicating “not at all” and 4 indicating “almost every day.” c. Fraction of parents who enforce curfew at least a few times a week. d. Fraction of households employed or on welfare on average during a postprogram quarter.

the Boston MTO site, shows that the difference between household heads in the experimental group and those in the control group in the proportion who report their overall health to be good or better equals 11.5 percent. Because just under half (48 percent) of the families in Boston assigned to the MTO experimental group relocated through the program, the effect of moving on the experimental group families who move is about twice as large as the overall difference between the two groups (23.9 percent).18 While the improved mental and physical health of MTO adults does not consistently translate into improved labor market outcomes, at least in the 18. Katz, Kling, and Liebman (2001).

1269-08 CH08 04/18/03 17:26 Page 145

    

145

short run, the changes do seem to improve the quality of parenting that MTO children receive. Table 8-1 shows some signs that rates of welfare receipt may have declined in the Baltimore MTO site, at least among experimental group adults compared with controls. But administrative data from Maryland show no signs of changes in employment or earnings in jobs that are part of the state’s unemployment insurance system,19 and survey data from the other MTO sites do not reveal much change in either welfare receipt or labor market employment or earnings.20 On the other hand, survey data from the New York site suggest some reduction in the frequency with which parents get angry or criticize their children—“harsh parenting”—and some increase (at least for the experimental group) in the fraction of parents who make their children obey a school-night curfew—“structured parenting.”21 For the MTO children, the change in neighborhoods also translates into substantial improvements in mental and physical health (table 8-2). Survey data from the Boston MTO site show that experimental group children experienced about half as many injuries or asthma attacks that required medical attention as control children. Experimental group children in Boston’s MTO program also experienced about half as many criminal victimizations as the controls. Survey data from New York show that experimental and comparison group children were around 20 percent less likely than controls to be depressed or anxious. Given these changes in children’s mental and physical health, perhaps it is not surprising that the MTO program also appears to reduce children’s involvement in antisocial behavior. Official juvenile justice records (“rap sheets”) for the Baltimore MTO site show that teenagers in the experimental and comparison groups were arrested for violent crimes perhaps half as often or less compared with the control group teens (table 8-2). While the results are largest in absolute terms and in relation to their standard errors for boys, they still are proportionally large for girls compared with the control group’s mean. Pooling boys and girls together yields a statistically significant (p <.10) relative decline in violent arrests for the comparison group as well as for the experimental group. The data in table 8-2 also suggest that experimental group teens may be arrested at higher rates for property offenses compared with controls. In 19. See Ludwig, Duncan, and Pinkston (2001). These findings are not necessarily inconsistent, since overall in Maryland only about half of those who leave welfare are employed in a job covered by the state unemployment insurance system (the source of the Baltimore employment data) during the quarter after their exit (Born 1999). 20. Katz, Kling, and Liebman (2001); Leventhal and Brooks-Gunn (2000a); Goering, Feins, and Richardson (2002). 21. Leventhal and Brooks-Gunn (2000c).

1269-08 CH08 04/18/03 17:26 Page 146

 

146

Table 8-2. Effects of MTO Demonstration on Children’s Mental and Physical Health and Antisocial Behaviors a

Outcome Mental/physical health Injury/accident in last 6 months requiring medical attentionb Asthma attack in last 6 months requiring medical attentionb Fraction unhappy, sad, or depressedc Fraction too fearful or anxiousc Problem behavior Arrests for violent crimed Boys Girls Arrests for property crimesd Boys Girls Fraction of 7 problem behaviorsb Boys Girls

Control mean .105 .098 .42 .48

4.3 1.8

3.3 0.9

.326 .193

Experimental group minus controls

Comparison group minus controls

-.059** (.027) -.051* (.029) -.10 (.05) -.16* (.05)

-.037 (.033) -.004 (.037) -.15* (.05) -.08 (.06)

-2.9** (1.4) -0.7 (0.7)

-1.9 (1.2) -0.4 (0.9)

2.7 (1.7) 1.0 (0.6)

-0.5 (1.2) 1.6 (1.1)

-.090** (.041) -.023 (.030)

-.113** (.053) -.050 (.034)

Sources: Authors’ compilations from sources as follows. Boston demonstration site: Katz, Kling, and Liebman (2001); New York site: Leventhal and Brooks-Gunn (2002); Baltimore site: Ludwig, Duncan, and Hirschfield (2001). *Statistically significant at the 10 percent level; **statistically significant at the 5 percent level. a. Standard errors in parentheses. b. Children aged five to sixteen at random assignment from the Boston MTO site. The seven problem behaviors reported by mothers are trouble getting along with teachers, disobedience at home, disobedience at school, hanging around with troublemakers, bullying others, inability to sit still, and depression. c. Children aged eight to thirteen from New York MTO site. d. Arrest rates per 100 teens per quarter, using the sample of Baltimore MTO teens who were aged eleven to sixteen at random assignment.

1269-08 CH08 04/18/03 17:26 Page 147

    

147

theory the property-arrest data could suggest that teens act out when their families move or take advantage of the more lucrative opportunities for theft in low-poverty neighborhoods. But an equally plausible hypothesis is that low-income minority children are more likely to be arrested when they commit a crime in a low-poverty neighborhood than in a high-poverty neighborhood. In that case comparison of arrest rates across groups will understate the effects of relocation on criminal behavior and overstate any increases in offending. Put differently, we can be more confident that the reduction in violent arrests represents a real change in teen behavior than does the increase in property arrests. Consistent with the hypothesis that MTO improves children’s behavior, surveys of household heads in Boston show that experimental and comparison group boys exhibited around onethird fewer problem behaviors such as bullying or disobedience than boys in the control group. Data from MTO also reveal changes in the measure of perhaps greatest interest to policymakers—children’s educational outcomes (table 8-3). In the Baltimore MTO site, young children (ages five to twelve at randomization) assigned to the experimental and comparison groups experienced gains in reading achievement scores on the CTBS, a standard commercially developed test, of around one-quarter of a standard deviation.22 Children in the experimental group also experienced gains in math scores of about the same magnitude, although those in the comparison group did not. To put the magnitude of the effects into perspective, changes in test scores of about this magnitude are similar to those obtained by moving children in grades 1 through 3 from classes of twenty-two to fifteen students.23 Compared with controls, the fraction of children who achieved a passing score on the Maryland Functional Test (MFT) in reading increased by around one-quarter for those in the comparison group and nearly doubled for the experimental group, although only the latter change is statistically significant. None of the other outcome measures show statistically significant changes across treatment groups for young children. Schools in Maryland do not administer the same types of CTBS reading and math scores to older children in middle and high school, so we have only limited information about how the MTO program affects their academic achievement. Table 8-3 shows that the fraction of experimental group teens who passed the MFT reading test was about one-quarter higher than the fraction of controls, although the standard error around this estimate is also quite 22. Ludwig, Ladd, and Duncan (2001). 23. Krueger (1999).

1269-08 CH08 04/18/03 17:26 Page 148

 

148

Table 8-3. Effects of MTO Demonstration on Children’s Educational Outcomes a

Outcome

Control mean

Young children CTBS reading scores (percentiles)

25.13

CTBS math scores (percentiles)

28.77

Fraction passing Maryland Functional Test in reading Fraction in special education

.192

Fraction retained in grade

.075

Adolescents Fraction passing Maryland Functional Test in reading Absences

.162

.268 29.85

Fraction in special education

.202

Fraction retained in grade

.092

Fraction subject to disciplinary action

.126

Experimental group minus controls

Comparison group minus controls

7.34** (2.75) 7.48** (3.67) .178** (.083) .043 (.032) -.013 (.013)

6.39** (3.23) 1.48 (3.83) .059 (.093) .052 (.037) .002 (.015)

.073 (.072) 2.06 (2.92) -.017 (.050) .065** (.032) .086* (.050)

-.009 (.071) 1.78 (3.45) .013 (.059) .109** (.052) .088 (.069)

Source: Ludwig, Ladd, and Duncan (2001). *Statistically significant at the 10 percent level; **statistically significant at the 5 percent level. a. Findings are from the Baltimore demonstration site. Standard errors in parentheses. Young children are aged five to twelve; adolescents are aged twelve and over.

large and the difference is not statistically significant. Some additional mixed support for the idea that academic achievement may have improved among the experimental and comparison group teens comes from standardized achievement tests administered by the research team in the New York site, which suggest gains among eleven- to eighteen-year-old boys but not girls.24 The data on nonacademic outcomes for MTO teens in Baltimore echo those for property arrests, in that grade retention and disciplinary action seem 24. Leventhal and Brooks-Gunn (2002c).

1269-08 CH08 04/18/03 17:26 Page 149

    

149

to have increased in the experimental and comparison groups compared with controls (table 8-3). But as with the arrest data for property offenses, there is some ambiguity about whether the nonacademic school outcomes reflect a deterioration in the behavior of MTO adolescents or instead enforcement of more rigorous standards in schools that are located in lower-poverty areas. Suggestive evidence that changing standards rather than behavior may be responsible comes from the fact that we do not observe similar deteriorations in the two measures of teen behavior that are less susceptible to variation across schools in enforcement and definitions—pass rates on the Maryland MFT test and school absences. What do the MTO results imply about the importance of steering poor families into the lowest-poverty neighborhoods? We observe statistically significant effects of assignment to the MTO experimental but not the comparison group for several of the outcomes shown in tables 8-1 through 8-3 (such as adult welfare receipt, asthma attacks among children, whether children obey a school-night curfew, and children’s math scores). For many of the other outcome measures we observe roughly similar effects of assignment to the experimental and comparison groups.25 The bottom line is that while steering poor families to the lowest-poverty areas enhances the effects of moving, even the somewhat less dramatic neighborhood changes experienced by the comparison group yielded gains (at least in the short term) in outcomes such as the mental and physical health of parents and children, problem behavior, and children’s reading scores.

What Next? For the past several decades social scientists have debated about whether “place matters” for the life chances of poor children. Despite a large theoretical literature about why place might matter,26 most research has found that after controlling for family background, the variation in children’s outcomes across different types of neighborhoods is surprisingly modest.27 Some observers also argue that only the most skilled and motivated of poor families 25. However, since a smaller proportion of experimental than comparison group families move through the MTO program, one effect of steering families to the lowest-poverty areas is to increase the benefits to families that move but reduce the proportion of those who experience any benefits at all. 26. Wilson (1987, 1996); Brooks-Gunn, Duncan, and Aber (1997). 27. Jencks and Mayer (1990); Brooks-Gunn, Duncan, and Aber (1997); Leventhal and BrooksGunn (2000).

1269-08 CH08 04/18/03 17:26 Page 150

150

 

are likely to successfully navigate the move into mixed-income areas, and that as a result even the modest effects found in previous studies may overstate the gains from living in more affluent areas.28 The new findings from HUD’s MTO housing-voucher experiment seem to paint a different picture about the potential role of neighborhoods, at least for those children living in the nation’s most distressed areas. While the randomized experimental design of MTO overcomes the main concern with previous studies—self-selection—even the current MTO evidence is limited in important ways. One limitation is the difficulty of comparing findings across sites, given that common outcome measures were not defined in advance of the decentralized study of the program’s short-term effects. The results presented here also speak only of the impacts on families through the first two to four years of the program. Sustaining the MTO demonstration and evaluation over the long term remains an important research priority. As we await longer-term findings from the MTO experiment, policymakers will confront a series of policy choices about how to provide housing assistance to low-income families. Based on the preliminary findings from MTO, policymakers should seriously consider strategies that improve the opportunities for mobility for poor parents and their children. What elements might be included in a strategy to improve the quality of neighborhoods for poor families? As noted by Anthony Downs, three general approaches have the potential to help poor families move from high- to low-poverty areas: housing vouchers; subsidized housing developments in suburban, or at least economically integrated, areas; and requirements that private developers in such areas set aside housing in new developments for low-income families.29 Each strategy presents different potential advantages and limitations, and each may play some role in a larger effort to improve the neighborhoods of poor children. The first possibility is to offer housing vouchers to public housing residents, particularly those living in the highest-poverty urban areas. This idea enjoys a long history within economics, based in part on the beliefs that housing vouchers are more cost-effective than housing projects and that landlords ultimately will supply suitable new rental units in response to new housing vouchers.30 Counseling services may be a useful element of such a plan, particularly if one goal is to avoid reconcentrating poor families in their 28. Evans, Oates, and Schwab (1992). 29. Downs (1999). 30. See, for example, Aaron (1972) and Olsen (2001) and the sources cited therein. Conceptual arguments in favor of housing vouchers include the reliance on older (and therefore less expensive) housing stock and the effects of market competition on private landlords, which are absent for public housing managers and private organizations that run housing projects for the government

1269-08 CH08 04/18/03 17:26 Page 151

    

151

new neighborhoods, although whether public housing agencies are the best organizations to carry out these efforts remains unclear. Also unclear is whether families should be steered or subsidized to move into neighborhoods with particular socioeconomic or demographic characteristics. MTO suggests that steering may reduce the proportion of families who move but increase the gains from moving for those who do move. Longer-term research from MTO may provide better information about this trade-off. An alternative to relying on the private housing market is to continue to provide government support for the construction of new housing projects but to locate new housing units in more mixed-income communities and earmark these units for families currently living in the highest-poverty public housing complexes. This strategy may appeal to those who are concerned about the supply of housing in the private rental market.31 The disadvantages of new scattered-site construction relative to vouchers may include cost, as well as the greater visibility—and perhaps political opposition—that may be associated with the former. The final possibility suggested by Downs is to require new private developments in economically or racially mixed areas to include some proportion of affordable rental units. One advantage of such a requirement from the perspective of policymakers is that the costs do not show up as a line item in the government’s budget, and, as Downs notes, developers can be compensated in part by granting them zoning waivers that allow a project to be built at a higher density than otherwise would be allowed. One key to implementing any of these mobility strategies may be to move only those families that wish to move. Programs that move public housing

(Aaron 1972; Olsen 2000, 2001). Some policymakers may argue that because a stock of public housing already exists there may be short-term budgetary advantages in keeping families in housing projects. But much of the nation’ s housing stock is now fairly old (Quigley 2000), and almost all of the empirical research suggests that when the costs of rehabilitating and maintaining the public housing stock are taken into account and the quality of housing services is held constant, housing vouchers are more cost-effective than public housing (Olsen 2000, 2001; U.S. Department of Housing and Urban Development 2000; Shroder and Reiger 2000; U.S. General Accounting Office 2001). As a result, the costs of gradually shifting toward vouchers as existing housing projects reach the end of their life cycle may be modest and even may save the government money. 31. Those who are concerned about a shift to housing vouchers cite the inability of many poor families to locate and lease suitable private-market apartments within the allotted search period; see, for example, Frederick Kunkle, “Housing Vouchers No Magic Key: Rising Rents, Dwindling Choices Thwart Low-Income Residents,” Washington Post, August 5, 2002, p. A1. On the other hand, past government spending on vouchers does seem to have increased the availability of suitable housing in the private market (Olsen 2001). While almost all of the vouchers available to local housing agencies wind up being used by someone, the failure of some families to lease housing may still be of policy concern if those are the families in greatest need of housing assistance.

1269-08 CH08 04/18/03 17:26 Page 152

152

 

residents without their consent—by, for example, tearing down their housing projects—appear to generate only modest changes in neighborhood characteristics, which in turn lead to modest if any improvements in children’s educational outcomes.32 Also important, at least with the voucher program, may be to grant subsidies to those families currently living in public rather than private-market housing, since previous research suggests that residents of private apartments experience few moves or neighborhood changes in response to rental subsidies.33 A reasonable question may arise at this point: if neighborhoods matter for poor children in high-poverty urban neighborhoods, why not simply improve the neighborhoods directly? One way to accomplish that goal would be to develop mixed-income housing projects that offer subsidies to working- or middle-class families to live there. While this strategy seems similar in many respects to residential mobility programs, mixed-income housing need not improve the life chances of poor children if neighborhood institutions rather than neighbors are the driving force behind neighborhood effects. For example, if schools are what matter most about neighborhoods for poor children and a new mixed-income development does not improve the quality of the local public schools, then neighborhood children may benefit little. One could try to improve the schools, but examples of successful school reform are hard to find.34 There also are few documented successes among previous efforts to improve urban neighborhoods through empowerment zones or community development programs.35 The most important unresolved issue for any mobility strategy comes from the effects on nonparticipants. We cannot rule out the possibility that relocating poor families may have negative consequences for families living in both the initial and destination neighborhoods. It is even possible that those consequences entirely offset any gains to movers, so that the overall amount of pro- and antisocial behavior within the society remains unchanged. These societywide impacts depend in part on the question of whether neighborhood effects on individuals are linear (that is, whether they affect individuals in a constant way) with respect to neighborhood attributes. The MTO evidence on this point so far is unclear.36 32. Jacob (2001). 33. Struyk and Bendick (1981). 34. While there is no shortage of schools that can point to dramatic year-to-year improvements in average test scores, Kane and Staiger (2002) note that disentangling schools’ effectiveness from normal year-to-year variability in scores is difficult. 35. Ladd (1995); Downs (1999); Ferguson and Dickens (1999); Rusk (1999). 36. Johnson, Ladd, and Ludwig (2002).

1269-08 CH08 04/18/03 17:26 Page 153

    

153

Policymakers thus are faced with something of a gamble: the MTO findings suggest that relocation programs may improve the life chances of at least some poor families, but there is some risk that these moves will impose costs on other children. Moreover, even the possibility of negative effects on other children, particularly those in destination neighborhoods, may produce some political opposition. If political opposition is proportional to the scale and public visibility of any mobility effort, policymakers might improve the political viability of any place-based programs by emphasizing vouchers over other types of housing programs, proceeding gradually, and limiting such efforts to residents of the highest-poverty public housing areas.37 Policymakers and voters who believe that society has an obligation to help the nation’s most disadvantaged children—particularly those who reside in distressed neighborhoods in part because of racial discrimination and past housing policies—may conclude that this is a chance worth taking.

References Aaron, Henry. 1972. Shelter and Subsidies: Who Benefits from Federal Housing Policies? Brookings. Black, Sandra E. 1999. “Do Better Schools Matter? Parental Valuation of Elementary Education.” Quarterly Journal of Economics. Brooks-Gunn, Jeanne, Greg J. Duncan, and J. Lawrence Aber, eds. 1997. Neighborhood Poverty. New York: Russell Sage. Cook, Philip J., and Kristin A. Goss. 1996. “A Selective Review of the Social-Contagion Literature.” Working Paper. Duke University, Sanford Institute of Public Policy. Deutsch, Morton, and Harold B. Gerard. 1955. “A Study of Normative and Informational Social Influences upon Individual Judgment.” Journal of Abnormal and Social Psychology 51 (3): 629–36. Downs, Anthony. 1999. “Some Realities about Sprawl and Urban Decline.” Housing Policy Debate 10 (4): 955–74. Evans, William N., Wallace E. Oates, and Robert M. Schwab. 1992.“Measuring Peer Group Effects: A Study of Teenage Behavior.” Journal of Political Economy 100 (5): 966–91. Ferguson, Ronald F., and William T. Dickens. 1999. Urban Problems and Community Development. Brookings. Figlio, David, and Mel Lucas. 2002. “What’s in a Grade? School Report Cards and House Prices.” Working Paper 8019. Cambridge, Mass.: National Bureau of Economic Research. Goering, John. 2003. “Closing Thoughts: Future Evaluation Research and Housing Policy Questions.” In Choosing a Better Life, edited by John Goering and Judith D. Feins. Washington: Urban Institute Press (forthcoming).

37. See also Goering, Feins, and Richardson (2002) and Goering (2002).

1269-08 CH08 04/18/03 17:26 Page 154

154

 

Goering, John, Judith D. Feins, and Todd M. Richardson. 2002. “A Cross-Site Analysis of Initial Moving to Opportunity Demonstration Results.” Journal of Housing Research 13 (1). Holzer, Harry J. 1991. “The Spatial Mismatch Hypothesis: What Has the Evidence Shown?” Urban Studies 28 (1): 105–22. Jacob, Brian A. 2001.“The Impact of Public Housing Demolitions on Student Achievement in Chicago.” Working Paper. Harvard University, Kennedy School of Government. Jargowsky, Paul A. 1997. Poverty and Place: Ghettos, Barrios, and the American City. New York: Russell Sage. Jencks, Christopher, and Susan E. Mayer. 1990. “The Social Consequences of Growing Up in a Poor Neighborhood.” In Inner-City Poverty in the United States, edited by Laurence Lynn and Michael McGeary, 111–87. Washington: National Academy Press. Johnson, Michael P., Helen F. Ladd, and Jens Ludwig. 2002. “The Benefits and Costs of Residential Mobility Programmes for the Poor.” Housing Studies 17 (1): 125–38. Kain, John F. 1968. “Housing Segregation, Negro Employment, and Metropolitan Decentralization.” Quarterly Journal of Economics 82 (2): 175–97. Kane, Thomas J., and Douglas O. Staiger. 2002. “Volatility in School Test Scores: Implications for Test-Based Accountability Systems.” In Brookings Papers on Education Policy, edited by Diane Ravitch, 235–84. Brookings. Katz, Lawrence F., Jeffrey Kling, and Jeffrey B. Liebman. 2001. “Moving to Opportunity in Boston: Early Results of a Randomized Mobility Experiment.” Quarterly Journal of Economics 116 (2): 607–54. Krueger, Alan B. 1999. “Experimental Estimates of Education Production Functions.” Quarterly Journal of Economics 114 (2): 497–532. Ladd, Helen F. 1995. “Spatially Targeted Economic Development Strategies: Do They Work?” Cityscape 1 (1): 193–18. Leventhal, Tama, and Jeanne Brooks-Gunn. 2000. “The Neighborhoods They Live In: The Effects of Neighborhood Residence on Child and Adolescent Outcomes.” Psychological Bulletin 126 (2): 309–37. ———. 2002a. “The Early Impacts of Moving to Opportunity on Children and Youth in New York City.” Working Paper. Columbia University, Teachers College, Center for Children and Families . ———. 2002b. “Moving to Opportunity: An Experimental Study of Neighborhood Effects on Mental Health.” Working Paper. Columbia University, Teachers College, Center for Children and Families. ———. 2002c. “A Randomized Study of Neighborhood Effects on Low-Income Children’s Educational Outcomes.” Working Paper. Columbia University, Teachers College, Center for Children and Families. Ludwig, Jens. 1999. “Information and Inner-City Educational Attainment.” Economics of Education Review 18 (1): 17–30. Ludwig, Jens, Greg J. Duncan, and Paul Hirschfield. 2001. “Urban Poverty and Juvenile Crime: Evidence from a Randomized Housing-Mobility Experiment.” Quarterly Journal of Economics 116 (2): 655–80. Ludwig, Jens, Helen F. Ladd, and Greg J. Duncan. 2001. “Urban Poverty and Educational Outcomes.” In Brookings-Wharton Papers on Urban Affairs, edited by William G. Gale and Janet Rothenberg Pack, 147–201. Brookings. Ludwig, Jens, Greg J. Duncan, and Joshua C. Pinkston. 2001. “Neighborhood Effects on Economic Self-Sufficiency: Evidence from a Randomized Housing-Mobility Experi-

1269-08 CH08 04/18/03 17:26 Page 155

    

155

ment.”Working Paper 159. Northwestern University/University of Chicago Joint Center for Poverty Research. Manski, Charles F. 1993. “Identification of Endogenous Social Effects: The Reflection Problem.” Review of Economic Studies 60: 531–42. Olsen, Edgar O. 2000.“The Cost-Effectiveness of Alternative Methods of Delivering Housing Subsidies.” Working Paper. University of Virginia, Department of Economics. ———. 2001. “Housing Programs for Low-Income Households.” Working Paper 8208. Cambridge, Mass.: National Bureau of Economic Research. Quigley, John M. 2000. “A Decent Home: Housing Policy in Perspective.” In BrookingsWharton Papers on Urban Affairs, edited by William G. Gale and Janet Rothenberg Pack, 53–100. Brookings. Raphael, Steven. 1998.“The Spatial Mismatch Hypothesis of Black Youth Joblessness: Evidence from the San Francisco Bay Area.” Journal of Urban Economics 43 (1): 79–111. Rubinowitz, Leonard S., and James E. Rosenbaum. 2000. Crossing the Class and Color Lines: From Public Housing to White Suburbia. University of Chicago Press. Rusk, David. 1999. Inside Game, Outside Game. Brookings. Sampson, Robert J., Steven W. Raudenbush, and Felton Earls. 1997. “Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy.” Science 277: 918–24. Sampson, Robert J., Jeffrey D. Morenoff, and Thomas Gannon-Rowley. 2002. “Assessing ‘Neighborhood Effects’: Social Processes and New Directions in Research.” Annual Review of Sociology 28. Shroder, Mark, and Arthur Reiger. 2000. “Vouchers versus Production Revisited.” Journal of Housing Research 11 (1): 91–108. Struyk, R. J., and M. Bendick, eds. 1981. Housing Vouchers for the Poor: Lessons from a National Experiment. Washington: Urban Institute. U.S. Department of Housing and Urban Development. 2000. Economic Cost Analysis of Different Forms of Assisted Housing. Issue Brief 11. Office of Policy Development and Research. U.S. General Accounting Office. 2001. Federal Housing Assistance Programs: Costs and Housing Characteristics. GAO-01-901R. Weinberg, Bruce A. 2000. “Black Residential Centralization and the Spatial Mismatch Hypothesis.” Journal of Urban Economics 48: 110–34. Wilson, William J. 1987. The Truly Disadvantaged. University of Chicago Press. ———. 1996. When Work Disappears: The World of the New Urban Poor. Alfred A. Knopf.

1269-09 CH09 04/18/03 17:28 Page 156

9 The Blair Government and Child Poverty: An Extra One Percent for Children in the United Kingdom JOHN HILLS

I

n March 1999, just before the second anniversary of his installation as prime minister of the United Kingdom, Tony Blair surprised the audience to his lecture in memory of William Beveridge, one of the founding fathers of the U.K. postwar welfare state, by making the following pledge: “Our historic aim will be for ours to be the first generation to end child poverty. It is a 20-year mission, but I believe it can be done.”1 His statement was remarkable in several ways. First, of course, for its boldness, even foolhardiness—eliminating any aspect of poverty is seldom the aim of any set of reform proposals. Second, setting a twenty-year period for this ambition might be expected to make life easier. After all, average incomes could increase by a half in that time, so simply maintaining the income growth of the poor at the same rate would take many of them across an absolute poverty line of the kind used in the United States. However, in the United Kingdom and Europe, poverty generally is understood to be relative to contemporary income. If the poverty line is seen in

1. Blair (1999).

156

1269-09 CH09 04/18/03 17:28 Page 157

    

157

relative terms, the objective is to make progress against a moving target, which is far more difficult. But perhaps what is most remarkable is that U.K. policymakers have been taking this objective seriously. The long-term goal has been broken down into shorter-term objectives: cutting the child poverty rate by a quarter within six years and halving it within eleven. Both the Treasury and the Department for Work and Pensions (which administers cash benefit and employment policies) have been given, as one of a small number of key objectives, instructions to reduce by a quarter by 2004 the number of children living in families with incomes of less than 60 percent of contemporary median income. And a considerable amount of policy reform and new public spending has been aimed at reducing child poverty and improving children’s life chances. As one measure of the scale of this undertaking, the Treasury calculates that on average each family with children will have gained £1,200 a year from its reforms to the tax and benefit system (comparing the actual 2003–04 system with that of 1996–97 if its rates and brackets simply had been adjusted for price inflation).2 The gain will be twice as much for those in the poorest fifth of the population. In aggregate, this represents a total of nearly £9 billion per year, or 0.9 percent of U.K. gross domestic product.3 A similar commitment by the U.S. government in relation to national income would mean annual spending of around $90 billion. On top of this, there are significant new programs, including those for preschoolers (Sure Start); expanded child care provisions; more places in nursery schools for three- and four-year-olds; grants for low-income families whose sixteen- and seventeen-year-olds stay in school; support for adolescents; and proposals for “baby bonds,” grants to establish trust funds when children are born. This chapter examines the background to this commitment in terms of trends in poverty and inequality in the United Kingdom and public attitudes toward them, as well as the constraints that public attitudes may place on policy responses. It describes the policy reforms that have been introduced, both those involving cash incomes and wider policies regarding investment in children and what has become known as “social exclusion.” It discusses the evidence so far on the outcomes of the reforms, both short-term effects and

2. Her Majesty’s Treasury (2002a, p. 89). 3. Independent analysis suggests that the real increase in assistance channeled through the tax and benefit system that is specifically linked to children will amount to £7.5 billion between 1998–99 and 2003–04 (see figure 9-5).

1269-09 CH09 04/18/03 17:28 Page 158

158

 

the prospects for improving disadvantaged children’s life chances. Finally, the chapter considers whether any potential implications exist for U.S. policy. As will be seen, many of the measures currently being introduced in the United Kingdom match recommendations for U.S. policy made by other contributors to this volume.4 If nothing else, over the next few years the United Kingdom will offer something of an experimental test-bed for this kind of agenda.

The Problem and Constraints on Solutions The high priority that the “New Labour” government under Tony Blair has given to this area is partly explained by the sheer scale of the growth in inequality and relative poverty in the United Kingdom in the preceding two decades. But also important is that the general public in the United Kingdom sees that growth as a problem—perhaps more so than the U.S. public would. This has meant that up to a point the public shares the government’s priorities. But public attitudes also influence the choice of instruments used to achieve poverty reductions—or at least, the policy package adopted shows every sign of having been carefully designed to go with the grain of public attitudes rather than cut across it. Figure 9-1 combines two indications of this. The bars show the proportion of the population with income of below half the average amount.5 From a rate of around 10 percent in the 1960s and a low point of 6 percent in 1977, the proportion grew rapidly in the 1980s, to above 20 percent by the early 1990s. As relative poverty grew in the United Kingdom, so (with a lag) did the proportion of the population agreeing that “the gap between those with low incomes and those with high incomes is too large.” By the mid-1990s a remarkable 87 percent of respondents to the annual British Social Attitudes (BSA) survey agreed with that statement. Since then the percentage agreeing has fallen back, but it is still more than 80 percent. Furthermore, a large 4. See, for example, chapter 2 in this volume. 5. Incomes are measured after benefits and direct taxes and are adjusted for family size. This series does not adjust for changes in housing costs. As a result of the operation of the housing benefit (housing allowance) system at times when general rental subsidies were withdrawn (and housing allowances, which are included in some definitions of income, rose without there being any change in living standards), this series can be argued to understate the growth in relative poverty. The poverty threshold used here is 50 percent of mean income in order to give a consistent year-byyear trend, rather than the threshold of 60 percent of median income, which is now more commonly used in the United Kingdom and across the European Union. However, in the 1990s the two thresholds had very similar values.

1269-09 CH09 04/18/03 17:28 Page 159

    

159

Figure 9-1. Relative Poverty and British Public Attitudes, 1983–2000 Percent

Percent

90 80

20

70 60

15

50 40

10

30 20

5

87 19 88 19 89 19 19 90 91 / 1 9 92 92 1 9 /93 93 1 9 /94 94 1 9 /95 95 / 1 9 96 96 1 9 /97 97 1 9 /98 98 1 9 /99 99 / 2 0 00 00 /0 1

86

19

84

85

19

19

19

19

83

10

Population with income of less than half the average (left axis)

Respondents agreeing that the gap between rich and poor is too large (right axis)

Source: Goodman and Webb (1994); Department for Work and Pensions (2002); Hills (2001).

proportion of the population (73 percent of respondents to the 1998 BSA survey) believed that it is definitely or probably government’s responsibility to reduce such differences.6 This is an obvious area of difference with U.S. attitudes. In a comparable survey in the late 1980s, only 56 percent of U.S. respondents thought that income differences in their country were too large; even for them, doing something about it may not have been seen as the government’s responsibility.7 The United Kingdom does not have an official poverty line; figures for trends in numbers of people with relative low income are the most commonly

6. Hills and Lelkes (1999). 7. Evans (1993).

1269-09 CH09 04/18/03 17:28 Page 160

160

 

used official measure. Measured against an absolute poverty line, the trends do not look so bad, although progress in the 1980s was still slow.8 However, U.K. public perceptions appear to be linked to trends in relative, not absolute, poverty measures.9 The explicit targets given to government departments in operationalizing the child poverty commitment over the period to 2004 reflect this. However, the government also draws attention to statistics measuring progress in absolute terms and to a wide range of other indicators of well-being, published in its annual Opportunities for All report on poverty and social exclusion. The government was still consulting more than three years after Blair made his pledge on how best to measure child poverty. But since the government has criticized its predecessors because relative poverty rose under them, it would be surprising if relative income measures did not remain at the core of its assessments. Figure 9-2 shows that the rise in poverty among children was particularly fast, with relative poverty (against a line of 60 percent of median income) rising from 12 percent in 1979 to around 25 percent in the 1990s.10 Even against an absolute standard, the second line shows that despite economic growth, child poverty was little lower in 1992 than it had been in 1979, although it has fallen substantially since then, particularly since 1996.11 Among countries in the Organization for Economic Cooperation and Development (OECD), the United Kingdom went from having one of the more equal income distributions in the 1970s to having one of the highest relative poverty rates in the mid-1990s.12 As figure 9-3 shows, UNICEF’s analysis of relative child poverty rates in the mid-1990s (using a lower threshold of 50 percent of median income) gives a rate in the United Kingdom that was far higher than in most other countries in Northern Europe and almost as high as in the United States. And if an absolute international threshold based on the U.S. poverty line is used—converted using purchasing power parities and consistent income definitions—the child poverty rate

8. For instance, against an absolute line of 60 percent of median income as it was in 1994–95, adjusted only for inflation, 27 percent of the population were in poverty in 1979, 21 percent in 1991/92 (averaging two financial years), and 16 percent in 1996–97, but only 11 percent in 2000–01 (Department for Work and Pensions, 2002, table H5; before housing cost figures). 9. Hills (2001). 10. This is despite the use of an equivalence scale that gives less weight to children’s needs than the modified OECD scale more commonly used in other countries. 11. The fixed real line used in the figure of 60 percent of median income, as it was in 1994–95, is a somewhat lower one than would be obtained by converting the U.S. poverty line at purchasing power parities. 12. Smeeding, Rainwater, and Burtless (2001, figure 5.1).

1269-09 CH09 04/18/03 17:28 Page 161

    

161

Figure 9-2. U.K. Child Poverty in Relative and Absolute Terms Percent Children below 60% of 1994/5 median income in real terms

35 30 25 20

Children below 60% of contemporary median income

15 10 5

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

Source: Department for Work and Pensions (2002), tables H2 and H6; income is calculated before deducting housing costs.

of 29.1 percent in the United Kingdom remains much higher than in most of Europe and twice the 13.9 percent figure in the United States.13 So the United Kingdom had become a far more unequal society by the 1990s than before; that was seen by the public as a problem; and child poverty had grown particularly fast. But political choices of policy instruments are constrained by their public acceptability, and the Blair government in particular used to be accused of paying too close attention to opinion polls and focus groups. In this case, what is clear is that an across-the-board rise in cash benefits would not be the most popular way of tackling poverty.14 First, the public’s priorities are for more public spending on health care (the universal, tax-financed National Health Service) and education, not for cash benefits as a whole. Second, with regard to cash benefits, majorities do favor higher spending on what are in fact the major items—pensions for the elderly, disabled people, low-paid working families with children—but not on 13. UNICEF (2000, figure 2). The U.S. figure is lower than that conventionally given, as the income definitions used by UNICEF allow for income items such as food stamps and the EITC. 14. For further details of the analysis on which this is based, see Hills (2001, 2002).

1269-09 CH09 04/18/03 17:28 Page 162

 

162

Figure 9-3. International Comparison of Child Poverty Rates a Sweden Norway Belgium Denmark Netherlands France Germany Japan Greece Australia Canada Ireland United Kingdom Italy United States 5

10

15

20

Percent Source: UNICEF (2000). a. Percent of children living in households with incomes of less than half of the national median income, mid-1990s.

increases in unemployment benefits. This is not because many people see unemployment benefits themselves as overgenerous, but because they are worried about disincentives and abuse of the system. Only barely more people now favor statements about “redistribution” from rich to poor or increases in “welfare benefits for the poor” than disagree with them, although such statements drew clear majority support as recently as the early 1990s. On the other hand, there is very strong support for the idea that government should guarantee a job for anyone who wants to work. Putting all this together, there seems to be clear public backing for a strategy based on improvements—even at the cost of higher taxes—in universal public services such as health and education; selective increases in cash benefits; help for the working poor with children; and a work-based strategy for the unemployed. In many ways, that is exactly what the Blair government is delivering and drawing attention to (although it is making other less publicized changes as well).

Policies since 1997 The policy agenda since New Labour was elected in 1997 can be divided into two parts: tax and benefit reform and work promotion; and tackling longer-

1269-09 CH09 04/18/03 17:28 Page 163

    

163

term sources of disadvantage. The latter has not focused just on children, but on social exclusion more generally, taken to cover population groups at special risk—particularly adolescents—and low-income neighborhoods. As Chancellor [Treasury Secretary] Gordon Brown has put it, “Tackling childhood poverty and disadvantage is not about providing either more money or better public services: it is of necessity about both.”15

Policies toward Income There has been no across-the-board increase in cash benefit rates. As a default, they have remained indexed to prices only rather than raised in real terms or in line with some measure of income or earnings, as was general practice in the 1960s and 1970s. Ministers have strenuously avoided use of the word “redistribution,” even though, as shown in figure 9-6 below, that could be a reasonable description of the impact of the tax and benefit changes that they have announced, particularly if one takes account of the likely impact of some of the revenue measures that have contributed to funding the changes. While a great deal of attention has been paid to poverty and to the gap between low and average incomes, there has been little sign of concern about inequalities in income at the top of the distribution. Both the 1997 and 2001 Labour election manifestos committed the government not to increase income tax rates (and indeed the main income tax rate was cut, although various tax allowances have been abolished and social insurance contribution rates rose in April 2003). There has been strong emphasis on “work as the best form of welfare,” with the “New Deals” forming the center of a work-based strategy characterized by the sharp distinction embodied in the slogan “Work for those who can, security for those who cannot.” The largest program has been the New Deal for Young People, with mandatory participation for unemployed people under the age of twenty-five receiving benefits for six months or more. The program involves initial intensive support in job search, followed by compulsory participation in one of four options: subsidized employment, full-time education and training, voluntary activity, or work for an “environmental task force.” For older unemployed people and their partners, the New Deal for the Long-Term Unemployed has a similar structure, but it applies only after two years of unemployment. The New Deal for Lone Parents is an essentially voluntary scheme, in contrast to U.S. equivalents. Lone parents receiving income support (the means-tested state safety net) are 15. Her Majesty’s Treasury (2001, p. iii).

1269-09 CH09 04/18/03 17:28 Page 164

164

 

invited in for an interview with a personal adviser to discuss job search, available training options, and the in-work benefits and tax credits available. Attending such interviews has now been made compulsory, but taking up training or work is not. Under “making work pay,” emphasis has been put on the working families tax credit (WFTC), a refundable tax credit for couples with children or single parents working sixteen or more hours per week (with extra if they work thirty hours or more). Unlike the U.S. earned income tax credit (EITC), U.K. tax credits are paid regularly through the year, either through the pay packet or, in certain cases, as a direct payment to the mother in a couple. An additional tax credit pays up to 70 percent of approved child care costs of up to £200 per week for more than one child. However, benefits for workless parents with children also have been increased, so that while net gains from working have increased, they have not done so nearly as markedly as under the EITC reforms of the Clinton administration in the United States. In April 1999 New Labour also brought in the United Kingdom’s first national minimum wage; previous sectoral minimums had been abolished by the Conservatives in the 1980s. Its rate at introduction was equivalent to 44 percent of median hourly full-time earnings; in comparison, the U.S. equivalent was 38 percent of the median.16 At the same time, various reforms to income tax and national insurance contributions have reduced the direct tax burden on both low-paid workers and their employers. The previous jumps—or notches—in tax liability as particular thresholds of low earnings are reached have been removed. As well as measures to make work pay, the government has made substantial changes to the tax and benefit system as a whole. Key features affecting families with children include the following: —Significant real increases in the value of the universal child allowance, Child Benefit, and introduction of a new tax credit benefiting all taxpayers with children except those with the highest incomes (partly paid for by abolishing the previous additional income tax allowance for married couples). The rate of the additional tax credit is doubled in the first year of a child’s life, and the maternity allowance paid to mothers with some previous work record has been made more generous and easier to claim; in addition, from 2003 on it is paid for twenty-six rather than eighteen weeks. —As well as the more generous in-work support through the WFTC, there also have been much less publicized but substantial increases in 16. Metcalf (2000, table 10.1).

1269-09 CH09 04/18/03 17:28 Page 165

    

165

allowances for younger children in nonworking families receiving income support.17 By April 2003, the rates for two children under age eleven were more than 80 percent higher in real terms than they had been in 1997.18 —In its first year in office, Labour implemented—at some cost to its reputation on the left—its predecessor’s plans to phase out special additional benefits for lone parents. Losses from these changes generally have been more than compensated for by the increases in allowances for children in all families, so that the overall effect has been one of “leveling up” the support for children in one- and two-parent families. In further reforms from April 2003, the system of tax credits was simplified and extended further. The various benefits and tax credits for children described above (apart from Child Benefit) have been amalgamated into a single child tax credit of equal value to those out of work and in low-paid work but of less value to those with higher incomes. Payments now usually go directly to the mother rather than through the paycheck. Figure 9-4 illustrates what the level of support is at different levels of household income for a couple with one child. If out of work or working less than sixteen hours on a very low income, the couple would be entitled to the adult part of income support on top of the child tax credit, but this is reduced at a rate of 100 percent as earnings rise. If working more than sixteen hours, they could claim the new working tax credit (essentially the adult part of the old WFTC), which is withdrawn at a rate of 55 percent as earnings rise. Comparing the system in April 2003 with what it was six years earlier, just before Labour came to power, there have been large—even spectacular—increases in particular elements of the cash benefit and tax credit system, even after allowing for inflation: 27 percent in the rate of the universal Child Benefit for the first child; 84 percent in the maximum amount of inwork support for a family with two children under the age of eleven;19 17 percent in the total income support for a nonworking lone mother with one young child; and 29 percent in the total income support for a couple with two young children. 17. In contrast to programs in the United States such as TANF, the U.K. income safety net created by income support applies to virtually all household types, including two-parent as well as oneparent families, childless couples, and single people. 18. Her Majesty’s Treasury (2002c, p.13). 19. Comparing maximum support through the former family credit system in April 1997 (for those working sixteen to twenty-nine hours per week) with the support that will come through the new child and working tax credits. Tenants in low-paid work also can receive means-tested allowances for rent and local taxes. In this case, the increased generosity of the tax credits may lead to a reduction in these allowances, offsetting nearly all of the gain.

1269-09 CH09 04/18/03 17:28 Page 166

 

166

Figure 9-4. Support for Children through the U.K. Tax and Benefit System from April 2003 a Support (pounds/week) 140 120 100 80 60 40 20 250

500

750

1000

Gross weekly family income (pounds/week) Child benefit

Income support

Child tax credit

Working tax credit

Source: Her Majesty’s Treasury (2002c). a. Figures are for a couple with one child.

Figure 9-5 shows the impact of these increases, by comparison with past trends, on the total support specifically linked to having children channeled in one way or another through the tax and benefit system. The total annual cost to government had remained at around £10 billion (at 2002 prices) from 1975 to the start of the 1990s, rising to around £13 billion between 1993 and 1998. The measures described above result in a dramatic increase by almost half, to nearly £21 billion in 2003.20 Averaged across all children, the real increase between 1998 and 2003 is from £20 to £31 per child per week (but with much larger amounts for those with lower incomes).21

Investment in Children and Measures against Social Exclusion In terms of volume of spending, other measures have been of smaller scale, but the sheer number of different initiatives is striking. For young children, 20. Allowing for the rise in national income over the same period, this is an increase from 1.4 percent of GDP in 1998–99 to 1.9 percent of GDP in 2003–04. 21. Adam, Brewer, and Reed (2002, figure 4.21) suggest that the increase in support for children for the poorest quarter of couples with one child is from about £30 to £50 per week and for those with two children, from about £50 to £80 per week.

1269-09 CH09 04/18/03 17:28 Page 167

    

167

Figure 9-5. U.K. Government Tax and Benefit Support for Children, 1975–2003 a Billions of 2002 pounds

2002 pounds

30

30

Weekly per child (right axis)

20

20

10

10 Annual total (left axis)

1980

1985

1990

1995

2000

Financial year (by start date) Source: Adam, Brewer, and Reed (2002). a. Projected figures for 2003–04.

these include the new community-based Sure Start program, similar to Early Head Start. It is aimed at mothers, babies, and children up to age three living in low-income areas. The plan is to have 500 programs across the country by 2004, reaching 400,000 young children, including one-third of all poor children under age four.22 Local programs are intended to provide home visits for newborns, integrated health care, early education, part-time child care, and other services. The design of the program varies from area to area, depending on local preferences. It aims to reduce smoking during pregnancy and the number of young children placed on the child protection register, the number with speech and language problems, and the number in households with no employed adult. More generally, one use of the increased public spending announced in July 2002 was to expand child care, particularly to support the government’s goal of raising lone mothers’ employment rate to 70 percent by 2010 from about 44 percent in 1997. Total spending on child care, early education, and Sure Start is planned to reach £1.5 billion by 2005, with most of the amount representing new funding. 22. Her Majesty’s Treasury (2002b, paragraph 3.23).

1269-09 CH09 04/18/03 17:28 Page 168

168

 

The government also has been considering an “asset-based welfare” system—or “baby bonds”—for all newborns. It would, through matched savings, produce an asset that could be accessed when a child reaches adulthood. Details are still to be agreed, but annual spending would be several hundred million pounds. At the other end of childhood, a pilot system of educational maintenance allowances has been implemented and will be extended nationally from September 2004. The allowances will pay up to £1,500 a year for sixteen- and seventeen-year-olds from low-income families if they stay in school. The aim is to increase educational participation after the age of sixteen and indirectly to reduce social differentials in access to education after the age of eighteen. Spending on pilots in some of the poorest areas will reach more than £200 million in 2003–04; the national scheme will cost considerably more.23 Strikingly, most of the initiatives outlined so far have been led by the Treasury, under Gordon Brown, which now takes a very active role in social policy. Outside this, an early action of the New Labour government was to establish a “social exclusion unit” (SEU), initially reporting directly to Tony Blair but now coming under the deputy prime minister. The unit has concentrated on a series of issues that both affect groups at particular risk of later disadvantage and come under several different government departments. A recurrent theme in the unit’s recommendations has been the need for “joined up government” where problems have fallen between agencies. Thus it has looked at teenage pregnancy, street homelessness, ex-prisoners, and young people outside education, training, or work. This last study has led, for instance, to a new “connexions” service, which offers support and advice to all young people ages thirteen to nineteen. Like many other recent initiatives, it started with pilot schemes in low-income areas; it now has been extended across the whole of England, with a budget of nearly £500 million in 2003–04. A feature of New Labour’s first four years in office was the development of a large number of programs targeting separate areas—or “zones”—that include additional resources for services in particular areas, such as education, health, employment, and crime reduction. For instance, a Treasury-led New Deal for Communities is spending £2 billion over ten years in thirty-nine very deprived neighborhoods to improve local employment opportunities as well as housing and the physical environment. The SEU’s largest initiative has been 23. At present, the universal child allowance, Child Benefit, is paid only to the parents of sixteento seventeen-year-olds who are in full-time education. It is unclear at present whether the expansion of the more valuable educational maintenance allowances to all low-income families whose children stay in school will partly be funded by the withdrawal of Child Benefit from higher-income families with such children.

1269-09 CH09 04/18/03 17:28 Page 169

    

169

to design a more integrated national strategy for poor neighborhoods, looking in particular at the poorest tenth of electoral wards across the country. Its National Strategy for Neighbourhood Renewal aims to reduce the differentials in employment rates, crime, educational attainment, health, and housing conditions that exist between such areas and the average ward.24 “Floor targets”—of some significance in the target-dominated culture of the public sector under New Labour—have been set for the relevant departments, with the overall aim of improving the performance of mainstream public services such as local schools in deprived areas. They have been coupled with significant increases in general public spending on health and education, in particular since 1999, reflecting voters’ highest priorities. The increases have allowed, for instance, reductions in class size in primary grades and special initiatives to improve literacy and numeracy standards in primary schools. Together with the floor targets, the increased spending is intended to increase substantially the resources for and the performance of mainstream services in deprived neighborhoods.

Impacts and Prospects It is, of course, too early to make an overall judgment about the impact of these measures on child poverty even in the short run, let alone on life chances of deprived children in the long run. Figures 9-1 and 9-2 include indications of overall and child poverty up to Labour’s fourth year in office, 2000–01. Compared with rates the year before Labour came to power, overall relative poverty (using the threshold of 60 percent of median income) was down by only 1 percentage point, although for children it was down by between 3 and 5 percentage points, depending on how housing costs are treated. As figure 9-2 shows, the number of children below an absolute threshold, of a kind more similar to the U.S. poverty line, fell substantially, from 23 percent in 1996–97 to 13 percent in 2000–01. If one takes the starting point of the year Blair made his child poverty pledge, 1998–99, the rate of reduction in two years is just about fast enough, if continued, to meet the targets of cutting child poverty by a quarter in six years and halving it in eleven, even in relative terms. However, the reduction has not been as impressive as some had expected, given government claims at the time of the 2001 election that its measures already had taken more than 1 million children out of poverty. 24. Social Exclusion Unit (2001).

1269-09 CH09 04/18/03 17:28 Page 170

170

 

Two crucial provisos should have been added to such claims for them to have been valid. First, the claims related to measures announced by then, some of which were implemented only in 2002 or 2003 and were not in force by 2000–01. Second, they compare, using a fixed poverty line and microsimulation modeling, the actual systems as they will be introduced against a counterfactual of the tax and benefit system before Labour came to office, simply with its rates and brackets adjusted for price inflation. By doing this, official modeling suggests, for instance, that taken together the personal tax and benefit changes announced since 1997 mean that 1.5 million fewer children will be below the 60 percent of median income line in 2003 than would have been in the counterfactual case, a reduction of 12 percentage points.25 Independent modeling confirms this kind of finding. For instance, figure 9-6 shows the modeled impact of tax and benefit changes announced up to March 2001 (which therefore do not include the most recent changes introducing the new tax credit system from April 2003). Taking all of the measures together—including the impact of some changes not listed above—they amount to the equivalent of a gain of 17 percent in net income for the poorest tenth of the population, falling to under 1 percent for the richest tenth. For families with children in the poorest tenth, the average gain is more than 20 percent. If this was all that was going on, the measures would cut the child poverty rate by 10 percentage points, or by 1.3 million.26 However, this is not all that has been going on. Changes in the macroeconomy, such as falling unemployment—for some of which the government can take credit—have boosted the impact of tax and transfer reforms. But the “moving target” inherent in a relative measure for poverty rates has the opposite effect. After all, between 1997 and 2001, average personal disposable income per capita grew by about 14 percent in real terms. In other words, for low-income groups—many of whom are dependent on cash benefits—quite a large proportion of the gains shown in figure 9-6 would be needed just to hold their place in relative terms. Unless there were very large impacts from other factors, such as falling rates of worklessness, one would not expect the actual fall in relative poverty, even when we have the out-turn figures for 2003, to be as large as the modeling results. Piachaud and Sutherland suggest that, in the absence of any further improvements resulting from employment rates or other changes in the economy, the tax and benefit measures are large enough to reduce the relative child poverty rate by a further 2 percentage points between 2000–01 (for which we have actual fig25. Her Majesty’s Treasury (2002a, p. 90). 26. Piachaud and Sutherland (2002a, table 9.5).

1269-09 CH09 04/18/03 17:28 Page 171

    

171

Figure 9-6. Proportional Impact of U.K. Tax and Benefit Changes Announced 1997–2001 Percent

25

20

15

10

5

1

2

3

4

5

6

7

8

9

10

Income group All

With children

With pensioners

Source: Sutherland (2001).

ures) and 2003–04.27 This would be almost, but not by itself quite enough to meet the target of a one-quarter cut in relative child poverty by 2004. The objective of cutting child poverty has two aspects: one is the straightforward reduction of poverty in the short term; the other is the belief that this will have long-term effects. Tackling childhood poverty is seen as both a protective strategy for this generation and a preventive one for the next: “Tackling childhood disadvantage is particularly important because childhood experience lays the foundations for later life. Children growing up in low-income households are more likely to have poor health, to do badly at school, become teenage mothers or come into early contact with the police, to be unemployed as adults or to earn lower wages.”28 27. Piachaud and Sutherland (2002b). If it was not for the “moving target” effect of using a relative poverty definition, the projected impact of the tax and benefit changes over the three years would be much larger against an absolute poverty line. Indeed, figures published as this book goes to press suggest that the relative child poverty rate fell by 0–1 percentage point from 2000–01 to 2001–02, but that the absolute rate fell by up to 4 percentage points in this single year. 28. Her Majesty’s Treasury (2001, p. v).

1269-09 CH09 04/18/03 17:28 Page 172

172

 

In support of this kind of statement the government cites a number of recent U.K. research studies, several of them using detailed evidence drawn from the birth cohort studies and longitudinal sources in which the United Kingdom is now comparatively rich.29 However, as other contributions to this volume (see, for instance, Greg Duncan and Katherine Magnuson’s chapter) have discussed, much of this evidence relates to associations between childhood poverty and later outcomes. Even when the studies control for other childhood and background factors, it remains possible that these associations reflect unobserved factors and that the causal links between income and poverty are much weaker. What is being done in the United Kingdom is not a random assignment experiment, with only some children benefiting from the policy reforms. Nonetheless, the scale of what is happening— particularly by contrast with changes in the opposite direction in the 1980s— should generate rich data to test some of these associations, as should the local variations in some programs and evaluations of pilots of others.

Conclusions: Are There Lessons for the United States? Using the tax and benefit system directly to raise the income of poor families with children is, it hardly needs saying, the most obvious way to tackle childhood poverty. It is often, however, discounted in the United States on grounds of political feasibility. Yet that is precisely what is being done in the United Kingdom at present, despite some strong similarities in political constraints and economic perspectives. Several features of the strategy that have allowed this are worth noting. Some may be particular to the United Kingdom, but others may have resonance in the United States. —First, the measures outlined above have taken six years to evolve. They were not announced immediately, and implementation has taken longer. The Blair government had a commanding majority in Parliament after the 1997 election and again after its reelection in 2001—following which it has acted with greater self-confidence on tax and spending issues. It has not had to worry about mid-term elections or a potentially hostile Congress: when it makes tax and benefit proposals, virtually all are implemented. —The antipoverty strategy has been presented as an attack on child poverty, together with some action on pensioner poverty. But children’s incomes are determined by the income of the family they are part of. Given 29. For example, Ermisch, Francesconi, and Pelavin (2001); Gregg and Machin (1999, 2000); Hobcraft (1998).

1269-09 CH09 04/18/03 17:28 Page 173

    

173

family composition, for each four children taken out of poverty, an average of three adults also will be taken out of poverty. Reducing child poverty therefore means reducing family poverty. Most lone parents, as well as couples, have gained from the measures, whether working or not, but that has not been stressed as its main characteristic; indeed, many people believe that the government actually has penalized lone parents. —The tax and benefit reforms have been combined with a work-based strategy for the unemployed and a raft of other measures targeting public services, low-income neighborhoods, child care, early education, and vulnerable groups. The combination is important: without the wider measures, the prospects for long-term impacts would be weaker; without the immediate income measures, the “social exclusion” agenda would look as if it were a way to run away from the biggest—and most expensive—issues. —Perhaps important is that the stress on “work as the best form of welfare” preceded the benefit and tax credit increases. The welfare-to-work measures themselves have been far less punitive for those without work than U.S. reforms. There has been more emphasis on the carrots of assistance with finding work and improvements to in-work support than on the sticks of compulsion and sanctions (although these do increasingly apply in some degree to unemployed claimants, apart from single parents). —In contrast to U.S. improvements to the EITC, benefits have been improved for the nonworking as well as the working poor. Although government rhetoric has stressed the “making work pay” aspects of the reforms, they might more accurately be described as raising family incomes for the poor without damaging work incentives in the way incentives might have been damaged if an unreformed cash benefit system simply had been made more generous. —The end result is a system of tax credits for children that is identical for parents who are out of work and those in low-paid or low-hours work. One hope is that this will ease the transition back into the labor market. It also may lead to a reduction in work hours if parents choose to spend more time at home with their children than at work earning more income. The balance of these two effects on labor supply remains to be seen. —As another aspect of equalization, particularly important in the U.S. context, the outcome of the reforms to family benefits can be seen as something of a leveling upward in the treatment of one- and two-parent families, with some of the special additional allowances for lone parents withdrawn as family benefits in general have risen. —Another part of equalization has been the leveling up of credits and benefits for children under eleven years of age to match those for children

1269-09 CH09 04/18/03 17:28 Page 174

174

 

over eleven. This partly reflects research evidence on the lack of a substantial difference between family spending on younger and older children, but it also has had the effect of concentrating gains on parents with younger children, who face the greatest competing pressures from work and child care. This has been reinforced by additional tax credits in the first year of a baby’s life and by improvements (from a low base in European terms) in maternity pay. —On the other hand, the reforms have continued the trend toward a system that gives more assistance for the first child than for subsequent children. This can be justified in terms of the impact on work opportunities of having even one child—and it certainly was politically expedient in ensuring that virtually all families with children gained from the reforms—but it does run against some of the evidence on the concentration of poverty in larger families. —The child tax credit system from April 2003 is based on regular payments made through the year, usually to the mother. This contrasts with the generally annual lump-sum payments from EITC. It also marks a quiet reversal of policy from the first stage of the reforms, when paying tax credits through the paycheck was seen as an important part of reinforcing the “work pays” message and of destigmatizing the payments. Previous evidence suggests that for some families, direct payments to the mother result in spending patterns that are more favorable to children. Again, this is an area where the results of the reforms will be of great interest. —Running through the reforms has been a strong degree of what one might call “selective universalism.” The universal child allowance has been significantly increased, and the largest increases in spending are planned for the universal National Health Service and state education. Health and education spending now is planned to absorb 3.1 percentage points more of national income in 2005–06, the probable end of Labour’s second term, than it did in the year before Labour took office, an even greater increase than the cost of the tax and benefit changes discussed here.30 In all of this, three points stand out. First, while what has been going on has at no point been described as a “war on poverty,” what has emerged in

30. However, public spending as a whole will have risen only by less than 1 percentage point of GDP. The room to allow this has come partly from a very tight squeeze on all public spending in Labour’s first three years in office, partly from falling debt interest payments, and also from falling overall spending on cash benefits—despite the child poverty measures—reflecting both lower unemployment and the way in which other, generally price-linked, benefits have fallen in value relative to incomes.

1269-09 CH09 04/18/03 17:28 Page 175

    

175

a piecemeal way does contain a whole series of different measures with both short- and long-term effects that together do constitute at least a major assault on it. Given the multiple drivers of poverty and exclusion and the variety of circumstances of those affected, this seems eminently sensible. Second, and perhaps the feature that would most surprise U.S. commentators, the package implemented has involved increases in cash transfers not just to the working poor, but also to the nonworking poor, provided that they have children. The end result is an equalization and merger of the support systems for children (but not adults) in the two groups. At the same time, the welfare-to-work measures for lone mothers, which involve some improvements in the cash returns to part-time work and improved child care, remain essentially voluntary. This partly reflects U.K. norms about working mothers.31 The treatment of nonworking families also is crucial for the potential success of the strategy. For instance, Piachaud and Sutherland explored the impact on child poverty of a hypothetical—and unlikely—scenario in which all parents with children over the age of five gained work of sixteen hours a week at the minimum wage.32 The reduction in child poverty that this created was only a quarter of that from the tax and benefit changes announced up to April 2001. The increased cash transfers to nonworking families probably have been politically possible only because they have come behind a strategy that stresses the aim of increasing work opportunities and incentives. For the credibility of this rhetoric, as well as to avoid negative employment effects, it has been important for low-paid working families to gain somewhat more than those without work. But the prospects of seeing continuing reductions in child poverty in the next few years are possible only because nonworking families have gained significantly as well. The third and perhaps the most important lesson is that the U.K. experience since 1997 demonstrates that is possible for a government—even one alive to the constraints of the opinions of a public that is skeptical about handouts to the unemployed—to bring in reforms that have a noticeable direct impact on the incomes of all poor families, working and nonworking, one-

31. When offered a choice between three statements about a single mother’s responsibilities to her children under school age, 21 percent of respondents to the 1998 BSA survey said that she “has a special duty to stay at home and look after the child,” 51 percent that “she should do as she chooses, like everyone else,” and only 17 percent that she has “a special duty to go out to work to support the child.” For a school age child, however, 44 percent choose the “go out to work” option, 45 percent the “do as she chooses” option, and only 5 percent the “stay at home” option (Hills and Lelkes, 1999). 32. Piachaud and Sutherland (2002a).

1269-09 CH09 04/18/03 17:28 Page 176

176

 

and two-parent. The results of those reforms as they unfold over the next few years—in terms of both their impact on children’s life chances and their political sustainability—could be of great interest to U.S. reformers.

References Adam, Stuart, Michael Brewer, and Howard Reed. 2002. The Benefits of Parenting: Government Financial Support for Families with Children since 1975. IFS Commentary 91. London: Institute for Fiscal Studies. Blair, Tony. 1999. “Beveridge Revisited: A Welfare State for the 21st Century.” In Ending Child Poverty: Popular Welfare for the 21st Century, edited by Robert Walker. Bristol, U.K.: Policy Press. Department for Work and Pensions. 2002. Households below Average Income 1994/5 –2000/01. Leeds, U.K.: Corporate Document Services. Ermisch, John, Marco Francesconi, and David Pelavin. 2001. Outcomes for Children of Poverty. DWP Research Report 158. Leeds, U.K.: Corporate Document Services. Evans, Geoffrey. 1993. “Class Conflict and Inequality.” In International Social Attitudes: The 10th BSA Report, edited by Roger Jowell, Lindsay Brook, and Lizanne Dowds. Aldershot, U.K.: Dartmouth. Goodman, Alissa, and Steven Webb. 1994. For Richer, for Poorer: The Changing Distribution of Income in the United Kingdom 1961/91. IFS Commentary 42. London: Institute for Fiscal Studies. Gregg, Paul, and Stephen Machin. 1999. Child Development and Family Income. York, U.K: Joseph Rowntree Foundation. ———. 2000. “Childhood Experiences, Educational Attainment, and Adult Labour Market Performance.” In Child Well-Being, Child Poverty, and Child Policy in Modern Nations, edited by Koen Vleminckx and Timothy Smeeding. Bristol, U.K.: Policy Press. Her Majesty’s Treasury. 2001. Tackling Child Poverty: Giving Every Child the Best Possible Start in Life. Pre–budget report document. London. ———. 2002a. Budget 2002: The Strength to Make Long-Term Decisions. HC 592. London: Stationery Office. ———. 2002b. 2002 Spending Review: Opportunity and Security for All. Cm. 5570. London: Stationery Office. ———. 2002c. The Child and Working Tax Credits. The Modernisation of Britain’s Tax and Benefit System, No. 10. London: Her Majesty’s Treasury. Hills, John. 2001. “Poverty and Social Security: What Rights? Whose Responsibilities?” In British Social Attitudes: The 18th Report—Public Policy, Social Ties, edited by Alison Park and others. London: Sage. ———. 2002. “Following or Leading Public Opinion? Social Security Policy and Public Attitudes since 1997.” Fiscal Studies 23 (4): 539–58. Hills, John, and Orsolya Lelkes. 1999. “Social Security, Selective Universalism, and Patchwork Redistribution.” In British Social Attitudes: The 16th Report—Who Shares New Labour Values? edited by Roger Jowell and others. Aldershot, U.K.: Ashgate.

1269-09 CH09 04/18/03 17:28 Page 177

    

177

Hobcraft, John. 1998. Intergenerational and Life Course Transmission of Social Exclusion: Influences of Childhood Poverty, Family Disruption, and Contact with the Police. CASE Paper 15. London School of Economics. Metcalf, David. 2000. “Unions, the National Minimum Wage and the Distribution of Pay.” In Public Policy for the 21st Century: Social and Economic Essays in Memory of Henry Neuburger, edited by Neil Fraser and John Hills. Bristol, U.K.: Policy Press. Piachaud, David, and Holly Sutherland. 2002a. “Child Poverty.” In Understanding Social Exclusion, edited by John Hills, Julian Le Grand, and David Piachaud. Oxford University Press. ———. 2002b. Changing Poverty Post-1997. CASE Paper 63. London School of Economics. Smeeding, Timothy, Lee Rainwater, and Gary Burtless. 2001. “U.S. Poverty in a CrossNational Context.” In Understanding Poverty, edited by Sheldon Danziger and Robert Haveman. Harvard University Press. Social Exclusion Unit. 2001. A New Commitment to Neighbourhood Renewal: National Strategy Action Plan. London: Cabinet Office. Sutherland, Holly. 2001. Five Labour Budgets (1997–2001): Impacts on the Distribution of Household Incomes and on Child Poverty. Microsimulation Unit Research Note 41. Cambridge University, Department of Applied Economics. UNICEF. 2000. A League Table of Child Poverty in Rich Nations. Innocenti Report Card 1. Florence, Italy: Innocenti Research Centre (June).

1269-09 CH09 04/18/03 17:28 Page 178

1269-10 Contributors 04/18/03 17:29 Page 179

Contributors

David J. Armor is a professor in the School of Public Policy at George Mason University. He has conducted research and written widely on education and education reform, school desegregation and related civil rights issues, and academic achievement. He is the author of Forced Justice: School Desegregation and the Law (Oxford Press, 1995). His most recent book, Maximizing Intelligence (Transaction, 2003), is about the effects of family environment on early childhood cognitive development. Greg J. Duncan is a professor in the School of Education and Social Policy at Northwestern University. He also is a faculty fellow with the Institute for Policy Research at Northwestern University and the director of the Northwestern University/University of Chicago Joint Center for Poverty Research. He has published extensively on income distribution, child poverty, and welfare dependence. He is the co-editor of For Better and For Worse: Welfare Reform and the Well-Being of Children and Families (Russell Sage, 2001) and From Neurons to Neighborhoods: The Science of Early Childhood Development (National Academy Press, 2000).

179

1269-10 Contributors 04/18/03 17:29 Page 180

180



Irwin Garfinkel is a professor at the Columbia University School of Social Work. His current research addresses fragile families and child well-being, child support reform, and welfare reform and social well-being in New York City. His publications include Assuring Child Support (Russell Sage Foundation, 1992) and, with Sara McLanahan, Single Mothers and Their Children: A New American Dilemma? (Urban Institute, 1986). More recently, he coedited Fathers under Fire: The Revolution in Child Support Enforcement (Russell Sage, 1998) and Social Policies for Children (Brookings, 1996). Robert Haveman is a professor in the Department of Economics and the Lafollette School of Public Affairs at the University of Wisconsin–Madison. His research interests are poverty and social policy, environmental policy, and tax policy. His publications include Earnings Inequality: The Influence of Changing Opportunities and Choice (American Enterprise Institute Press, 1996) and, with Barbara Wolfe, Succeeding Generations: On the Effects of Investments in Children (Russell Sage, 1995). John Hills is professor of social policy and director of the Centre for Analysis of Social Exclusion at the London School of Economics. His areas of specialization include social exclusion, the welfare state, social security, and income distribution. He is the author of New Inequalities: The Changing Distribution of Income and Wealth in the UK (Cambridge University Press, 1996) and the co-author of Understanding Social Exclusion (Oxford University Press, 2002) and Paying for Health, Education, and Housing: How Does the Centre Pull the Purse Strings? (Oxford University Press, 2000). Andrea Kane is outreach director for the Welfare Reform and Beyond initiative at the Brookings Institution and director of public policy at the National Campaign to Prevent Teen Pregnancy. Previously she was special assistant to President Clinton with the White House Domestic Policy Council and director of the welfare reform program with the National Governors’ Association. She recently co-edited Welfare Reform and Beyond: The Future of the Safety Net (Brookings, 2002) with Isabel V. Sawhill, R. Kent Weaver, and Ron Haskins. Jens Ludwig is an associate professor with the Georgetown University Public Policy Institute. His research focuses on social policy issues, including urban poverty, education, crime, and gun control. He is co-author of Gun Violence: The Real Costs (Oxford University Press, 2000) and co-editor of Evaluating Gun Policy (Brookings, 2003), and he has written widely on the effectiveness of residential mobility programs and improving inner-city education.

1269-10 Contributors 04/18/03 17:29 Page 181



181

Katherine Magnuson is a postdoctoral fellow at the Columbia University School of Social Work. Her research interests include child well-being and school readiness. Sara McLanahan is a professor of sociology and public affairs at Princeton University. She is also director of the Bendheim-Thomas Center for Research on Child Well-Being at Princeton University. She has researched and written widely on child well-being, marriage, single mothers, and child support. Her recent publications include, with Irwin Garfinkel, Social Policies for Children (Brookings, 1996) and Child Support and Child Well-Being (Urban Institute, 1994). Isabel V. Sawhill is a senior fellow in the Economic Studies program at the Brookings Institution and director of the Brookings Roundtable on Children. She also is co-director of the Welfare Reform and Beyond project at Brookings. Her recent publications include, with Ron Haskins, R. Kent Weaver, and Andrea Kane, Welfare Reform and Beyond (Brookings, 2002) and, with Rudolph Penner and Timothy Taylor, Updating the Social Contract: Growth and Opportunity in the New Century (W. W. Norton, 2000). Scott Scrivner is an assistant research associate with Public/Private Ventures in Philadelphia. His primary research interests include work force development, disability, and poverty and health policy. As a member of P/PV’s work force development department, he currently is involved in Fathers at Work, a work-based responsible fatherhood demonstration. His contribution to this volume was written while he was employed at the University of Wisconsin–Madison. Barbara Wolfe is a professor in the Department of Economics, the Department of Population Health Sciences, and the Lafollette School of Public Affairs at the University of Wisconsin–Madison. Her areas of expertise include poverty and health issues, disability, child well-being, and teen pregnancy. Her publications include, with Robert Haveman, Succeeding Generations: On the Effects of Investments in Children (Russell Sage, 1995) and Making Schools Work: Improving Performance and Controlling Costs (Brookings, 1994).

1269-11 BM 04/18/03 17:29 Page 182

Brookings Roundtable on Children

Policy briefs Currie, Janet. 2001. “A Fresh Start for Head Start?” Children’s Roundtable Policy Brief 5 (March). Duncan, Greg J., and Jens Ludwig. 2000. “Can Housing Vouchers Help Poor Children?” Children’s Roundtable Policy Brief 3 (July). Ellwood, David T. 1999. “The Plight of the Working Poor.” Children’s Roundtable Policy Brief 2 (November). Ladner, Joyce. 2000. “Children in Out-of-Home Placements.” Children’s Roundtable Policy Brief 4 (September). Sawhill, Isabel V. 1999. “Investing in Children.” Children’s Roundtable Policy Brief 1 (April). Sawhill, Isabel V., and Laura Chadwick. 1999. “Children in Cities: Uncertain Futures.” Brookings Survey Series (December).

Discussion papers Currie, Janet. 2000. “Early Childhood Intervention Programs: What Do We Know?” (April). Ellwood, David T., and Isabel V. Sawhill. 2000. “Fixing the Marriage Penalty in the EITC” (June).

182

1269-11 BM 04/18/03 17:29 Page 183

   

183

Hollister, Robinson. 2003. “The Growth in After-School Programs and Their Impact” (February). Sawhill, Isabel V., and Adam Thomas. 2001. “A Hand Up for the Bottom Third: Toward a New Agenda for Low-Income Working Families” (May). Schorr, Lisbeth. 2003. “Determining What Works in Social Programs and Social Policies: Toward a More Inclusive Knowledge Base” (February).

Edited volumes Sawhill, Isabel V., ed. 2003. One Percent for the Kids: New Policies, Brighter Futures for America’s Children. Brookings.

1269-11 BM 04/18/03 17:29 Page 184

1269-12 Index 04/18/03 17:29 Page 185

Index

Abecedarian project, 95, 102 Abuse: of children of teen mothers, 60; of mothers, 88 Academic success: after-school programs and, 72; of children of teen mothers, 60; early education and, 95; housing mobility programs and, 147–48; intelligence and, 99–100; interventions for children with behavioral and emotional problems, 22–23. See also Educational attainment; Intelligence Adolescents: benefits of interventions, 5; effects of mothers’ employment, 65–66; effects of welfare reform, 25; positive approaches to youth development, 72; vocational education programs, 61–62; whole family education, 98. See also After-school programs; Teen pregnancy prevention African Americans: academic achievement, 102; academic skills, 93; benefits of Head Start, 115; births to

unmarried parents, 76, 96; housing mobility programs, 141; two-parent families, 96 After-school programs: benefits, 66–67; concerns about, 71–72; content, 62, 66; cost-benefit analysis, 69–70; costs, 58, 68–69; effective programs, 66; federal funding, 57; goals, 57, 66–67, 72; intensity and duration, 67, 71; locations, 67–68; participation, 68, 69; proposal, 57–58, 65–68; research on, 61–65, 70–71; service learning components, 62–63, 66, 70, 71; staffing, 67; supervision, 63, 67; support for, 57, 70; teen pregnancy prevention, 56, 63, 64 Aid to Families with Dependent Children (AFDC), 27, 86 Arkansas: profamily policies, 97 Baltimore: Moving to Opportunity program, 142, 145, 147–49 Behavioral problems, 22–23

185

1269-12 Index 04/18/03 17:29 Page 186

186 Bergmann, Barbara, 130 Beveridge, William, 156 Big Brothers/Big Sisters, 62, 64 Blair, Tony, 156, 161, 168, 172. See also United Kingdom child poverty reduction initiative Blau, David M., 26 Boston: Moving to Opportunity program, 142, 143, 144, 145 Boys and Girls Clubs, 68 Breast feeding and child intelligence, 104, 106–07 Britain. See United Kingdom British Social Attitudes (BSA) survey, 158 Brooks-Gunn, Jeanne, 25 Brown, Gordon, 163, 168 Bush, George W., 12, 57, 132 Bush (G. W.) administration, 76 Canada. See Self-Sufficiency Program CAS-Carrera. See Children’s Aid Society-Carrera CCDF. See Child Care and Development Fund Cherry, Robert, 46 Chicago: Gautreaux Program, 141–42; Moving to Opportunity program, 142 Chicago Child-Parent Center (CPC), 114–15 Child allowances: benefits, 26–27, 29; integration into existing programs, 29; for middle-income families, 26–27; monthly payments, 27; for preschool children, 26–27; proposals, 7, 8, 18, 26–27, 28; research on, 24–26; in U.K., 164–65, 174 Childbearing, out-of-wedlock: efforts to reduce, 76, 96; increase in, 76; welfare benefits and, 85, 93. See also Single mothers; Single parents; Teen pregnancy

 Child care: costs, 122, 129–30; effects on child development, 115; for employed parents, 48; funding sources, 122; proposals, 8; quality of, 21–22, 115, 129–30; for twoparent families, 79, 84; subsidies, 41, 86, 119, 127, 164; in U.K., 164, 167; in universal preschool, 116 Child Care and Development Fund (CCDF), 119–20, 124, 126, 127, 130 Child care tax: credits, 46, 120–21; changes to finance universal preschool, 126, 130; deduction, 18, 28 Children: behavioral and emotional problems, 22–23; developmental needs, 18–19; federal spending on, 1; improving life chances of, 2; mother’s employment, 17, 19, 23; peer influences, 137–39. See also Healthy development of children; Infants; Intelligence; Preschool children Children of the National Longitudinal Study of Youth (CNLSY), 104, 105, 106, 107, 108 Children’s Aid Society-Carrera (CASCarrera), 63–64, 68 Children’s Defense Fund, 117 Child support policies: current, 86–87; effects on marriage rates, 86; enforcement, 86; fathers’ obligations, 79, 86–87; government benefits, 79, 87 Child tax credit (CTC), 46; acceleration of phase-in, 50, 52; replacement by child allowance, 18, 28–29 Child tax credit, U.K., 164, 165, 173, 174 Child Trends, 62 Clinton administration, 164 CNLSY. See Children of the National Longitudinal Study of Youth

1269-12 Index 04/18/03 17:29 Page 187

 Cognitive ability. See Intelligence Committee on Economic Development, 129 Community service, 62–63. See also Service learning Compensatory education programs, 93–94, 95 Cost, Quality, and Child Outcomes study, 115, 129 Cowan, C. P., 80 Cowan, Philip A., 80 CPC. See Chicago Child-Parent Center Crime: early education and, 115; housing mobility programs and, 145–47 CTC. See Child tax credit Currie, Janet, 115 Delaware: profamily policies, 97 Denk, Robert, 46 Dependent Care Assistance Program, 120, 121 Dependent Care Tax Credit. See Child care tax Dichter, Harriet, 122 Disabilities, children with, 22 Downs, Anthony, 150, 151 Duncan, Greg, 25 Early education: adult income and, 115; benefits, 17, 21, 113, 114–16; costs, 95, 115, 129–30; current programs, 117–24; in developed nations, 131–32; effects on IQ and achievement, 95; expanding, 21–22; federal funding, 119–20, 124; for high-risk children, 17, 21–22, 113; intelligence modification through, 94, 95, 102–03; as political issue, 132; private funding, 122; proposals, 8, 17–18; state funding, 117, 121–22, 123–24; state programs, 122–24, 130; Title I grants, 120; in U.K., 167. See also Preschool, universal

187 Early Head Start, 21, 95, 98, 102, 167 Earned income tax credit (EITC), 46; changes to promote marriage, 79; compared to U.K. tax credit, 164, 174; effects on family incomes, 42; effects on marriage rates, 86; expansion proposals, 45–46; increase in, 41; reducing to pay for universal preschool, 126; supplementing, 50–51, 52 Eccles, Jacquelynne S., 61, 62 Education: association with marriage, 83; compensatory programs, 93–94, 95; government spending, 1; housing mobility programs and, 147–49; U.K. spending, 169, 174. See also Academic success; After-school programs; Early education Educational attainment: and adult income, 101; British policies, 168; and career choices, 100; early education and, 114, 115; intelligence and, 100–01; of parents, 105, 108; in poor neighborhoods, 139; of single parents, 84. See also Academic success Education, Department of, 66, 120, 131 Education Week, 121, 123 EITC. See Earned income tax credit Elementary and Secondary Education Act (ESEA), Title I, 93, 102, 106, 120 Emotional problems, 22–23 Emotional support of children, 104, 107, 108 Employment: association with marriage, 83–84; demand-side measures, 48–49; of fathers, 79; housing mobility programs and, 144–45; increasing earnings from, 44–51; minimum wage, 49; public sector, 48; subsidies, 48; vocational education programs, 61–62. See also Mothers, employment of; Work requirements

1269-12 Index 04/18/03 17:29 Page 188

188 ESEA. See Elementary and Secondary Education Act Extended-Service Schools (ESS), 64–65. See also After-school programs Families: child care expenses, 122, 129; importance of family environment, 5, 7, 24, 94, 99, 109–10; number of children, 104, 106, 108; preschool fees paid by, 125–27, 129, 131; structure and children’s intelligence, 104, 106, 108; tax burdens, 47. See also Incomes, family; Parents; Whole family policy proposal Family and Medical Leave Act, 23 Family planning services, 61 Family strengthening demonstration, 77; assessment and referral service, 77–78, 84; components, 77–80; costs, 81, 89; criticisms of, 87–89; experimental design, 80–81; marriage-friendly policies, 78–79; motivation of participants, 81–83; participants, 77, 87; political feasibility, 89; rationale, 81–87; relationship programs, 79–80 Fashola, Olatokunbo S., 61 Fathers: abusive, 88; child support, 79, 86–87; educational levels, 105, 108; employment, 79; human capital, 84; involvement with children, 77, 82; unmarried, 84. See also Parents Federal government: Head Start funding, 117; spending on children, 1; support for low-income working families, 41 Finance Project, 124 Florida: profamily policies, 97; property taxes, 121 Fragile Families and Child Well-Being Study, 81–82, 87, 88 Garces, Eliana, 115 Gautreaux, Dorothy, 141

 Gautreaux Program, 141–42 General Accounting Office, 120 Georgia: Lottery for Education, 121, 123–24 Georgia Voluntary Pre-Kindergarten (GPK) program, 123–24, 130 Girls Inc., 65 Gootman, Jennifer Appleton, 62 Gore, Albert, Jr., 12, 57, 132 GPK. See Georgia Voluntary PreKindergarten program Graefe, Deborah Roempke, 59 Grossman, Jean B., 67 Head Start: behavioral and emotional interventions, 22; benefits, 21, 115; Early, 21, 95, 98, 102, 167; effects on academic performance, 93; effects on IQ, 95, 102–03; expanding, 8, 21–22, 132; federal funding, 124, 127, 130; participants, 117; programs, 117; spending, 117–18 Health and Human Services, Department of, 117, 119, 131 Health care: for children, 16–17; costs of proposals, 20; prenatal, 16, 19–20; proposals, 7–8; U.K. spending, 169, 174; unmet needs of children, 20 Health insurance: for children, 7, 116; children lacking coverage, 20; costs, 20; federal programs, 41; proposals, 8, 20; State Child Health Insurance Program, 7, 116 Healthy development of children: child care and, 115; costs of proposals, 3–4, 18, 22–23, 27–29; distributional effects of proposals, 30, 34; home environments, 24; needs, 18–19; policy issues, 16–17; poverty and, 24–26; proposals, 17–18 Helburn, Suzanne, 130 Hills v. Gautreaux, 141

1269-12 Index 04/18/03 17:29 Page 189



189

Hispanics: academic achievement, 102; academic skills, 93; births to unmarried parents, 76 Hollister, Robinson, 62 HOME scale of parenting behaviors, 107 Housing: affordable, 151; assistance programs, 150–53; subsidies, 79, 84, 86. See also Neighborhoods Housing, public: destruction of projects, 152; in high-poverty neighborhoods, 142; racial discrimination, 141; scattered construction, 151. See also Housing vouchers Housing mobility: approaches, 150; benefits, 136–37; educational outcomes, 147–49; effects on adults, 143–45; effects on children, 145–50; effects on nonparticipants, 152; Gautreaux Program, 141–42; health effects, 143–44, 145; political feasibility, 153; to suburbs, 141–42; voluntary participation, 151–52. See also Moving to Opportunity; Neighborhoods Housing vouchers, 8, 141–42, 150–51, 152. See also Moving to Opportunity

parent families, 108; welfare reform, 40, 41–42. See also Child allowances Individuals with Disabilities Education Act, 22 Infant Health and Development Study, 107 Infants: breast feeding, 104, 106–07; caregiving, 17; educational and family services, 95, 102–03; intellectual development, 94, 102–03; low birth weight, 60, 103–04, 107, 108; and maternal employment, 17; maternal work requirements, 18, 19, 23; teen mothers, 60 Intelligence: and adult income, 101; definition, 99; early education and, 21; and educational attainment, 100–01; 102–03; environmental factors, 102, 103–04, 108, 109–10; family environment and, 7, 24, 94, 99, 109–10; genetic factors, 101–02, 103–04, 108; importance, 99–101; malleability at young ages, 95, 101–03; modification strategies, 94; of parents, 103, 105; risk factors, 99, 103–08; whole family approach to improving, 95, 98. See also Academic success Internal Revenue Service (IRS), 131

Incomes: costs to increase, 51–52; and early education, 115; effects of welfare reform, 40; increasing earnings, 44–51; inequality, 158, 159, 160, 163; and intelligence and educational attainment, 101; supplements, 50–51. See also Poverty; Wages Incomes, family: EITC and, 42; effects on child development, 6, 24–26; effects on child intelligence, 104, 106; proposed support, 18, 26, 41, 42; single mothers, 18, 41–42; two-

Jencks, Christopher, 139 Justice, Department of, 66 Kamerman, Sheila, 131–32 Katz, Lawrence F., 143 Kirby, Douglas, 62 Kling, Jeffrey, 143 Labour Party, 161–63, 164, 168, 172. See also United Kingdom child poverty reduction initiative Learn and Serve America initiative, 63 Lichter, Daniel T., 59

1269-12 Index 04/18/03 17:29 Page 190

190 Liebman, Jeffrey B., 143 Los Angeles, Moving to Opportunity program, 142 Lotteries, state, 121, 123–24 Manpower Demonstration Research Corporation, 65 Markman, Howard J., 80 Marriages: association with education and employment, 83–84; efforts to promote healthy, 76, 77, 78–79, 88, 89, 95; quality of relationships, 81; relationship programs, 79–80, 89; stability, 81, 85; superiority to cohabitation, 88. See also Family strengthening demonstration; Parents Marshall, Nancy, 129–30 Massachusetts: child care, 129–30. See also Boston Maternal leave. See Parental leave Mayer, Susan, 26, 139 Maynard, Rebecca, 60 Mental health: children’s problems, 22–23; housing mobility and, 143–44, 145; problems of poor parents, 24 MFIP. See Minnesota Family Investment Program Milwaukee. See New Hope Project Minimum wage, 44–45, 49, 164 Minnesota: profamily policies, 97 Minnesota Family Investment Program (MFIP), 43, 44 Mitchell, Anne, 122 Moore, Kristin, 6 Morris, Pamela A., 25 Mothers: adolescent, 59–60; ages, 104, 106, 108; educational levels, 105, 108; IQs, 105, 108. See also Parents; Single mothers; Teen pregnancy Mothers, employment of: effects of welfare reform, 40; effects on ado-

 lescent children, 65–66; effects on child’s development, 17, 19, 23; incentives, 45, 46, 49, 50; incomes, 40; measures to increase, 48–49; tax consequences, 45; U.K., 167, 175. See also Work requirements Moving to Opportunity (MTO): benefits, 136–37; cities, 142; outcomes, 143–49, 150, 151; treatment groups, 142–43 MTV, 137–38 National Academy of Sciences, 62, 72, 114 National Assessment of Educational Progress (NAEP), 102, 105, 106 National Campaign to Prevent Teen Pregnancy, 62, 94, 97 National Health Interview Survey, 20 National Institute of Child Health and Human Development Early Child Care Research Network, 115 National Longitudinal Study of Youth (NLSY), 100. See also Children of the National Longitudinal Study of Youth National Survey of America’s Families (NSAF), 69 National Survey of Children, 106 Neighborhoods: characteristics of neighbors, 139–40; evidence on effects, 141; improving, 152; influence on children, 8; institutions, 140, 152; low-poverty, 8, 136–37, 141–42; mechanisms for effects on individuals, 137–40, 152; peer behaviors, 137–39; segregated, 140; social ties, 139. See also Housing mobility Neighborhoods, high-poverty: afterschool programs, 65–66, 69; education levels, 139; population, 136; public housing, 142; U.K. programs in, 168–69

1269-12 Index 04/18/03 17:29 Page 191

 New Deal for Communities, 168 New Deal for Lone Parents, 163–64 New Deal for the Long-Term Unemployed, 163 New Deal for Young People, 163 New Hope Project, 43, 44, 45, 50, 52 New Jersey: profamily policies, 97 New York City: Beacon Program, 68; Moving to Opportunity program, 142, 143, 145, 148 NLSY. See National Longitudinal Study of Youth No Child Left Behind Act of 2001, 57 NSAF. See National Survey of America’s Families Nurturing, 104, 107, 108 Nutrition: assistance programs, 7; effects on child intelligence, 104, 106–07 Oklahoma: profamily policies, 97 Olsen, Darcy, 68 Organization for Economic Cooperation and Development (OECD), 132, 160 PAP. See Parents and Partners (PAP) program Parental leave, 7, 18, 23–24 Parenting behaviors: association with relationship quality, 81; cognitive stimulation, 104, 107, 108; effects of housing mobility programs, 145; effects of poor mental health, 24; emotional support, 104, 107, 108; HOME scale, 107; improving, 7; training in, 22, 96–97, 98 Parents: educational levels, 105, 108; education of, 104; human capital, 83–84; income supports, 18, 26, 41, 42; IQs, 103, 105; mentoring for, 80; quality of relationships, 7, 81. See

191 also Families; Fathers; Marriages; Mothers; Single parents Parents and Partners (PAP) program, 80 Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA), 23, 76, 95, 119 Philips, Meredith, 107 Piachaud, David, 170, 175 Political feasibility of proposals: child allowances, 27; family strengthening demonstration, 89; housing mobility, 153 Poverty: and academic ability, 106; children in U.K., 160–61, 169; children of single parents, 41, 59; effects on child development, 24–26; effects on home environments, 24; effects on preschool children, 17; long-term effects of reducing, 171–72; rates for single mothers, 40–41; relative (U.K.), 158, 159–60, 169; U.K. attitudes, 158–59, 160; U.S. attitudes, 159. See also Incomes; Neighborhoods, high-poverty; United Kingdom child poverty reduction initiative PPV. See Public/Private Ventures Pregnancy. See Prenatal care; Teen pregnancy Prenatal care, 16, 19–20 PREP program, 80 Preschool, universal: argument for, 114–16; costs, 116, 130; current programs, 123–24; in developed nations, 131–32; education and child care components, 116; funding sources, 125–27, 129, 130–31; Georgia program, 123–24, 130; goals, 125; proposal, 113–14, 124–31; year-round operation, 116. See also Early education

1269-12 Index 04/18/03 17:29 Page 192

192 Preschool children: benefits of interventions, 5; child allowance proposals, 26–27; development of, 114; effects of poverty, 17; effects of welfare reform, 25; needs, 16. See also Child care; Early education PRWORA. See Personal Responsibility and Work Opportunity Reconciliation Act Public/Private Ventures (PPV), 63, 65 Quantum Opportunities Program (QOP), 62, 64, 68 Reach for Health and Community Youth Service Learning, 62, 71 Reynolds, Arthur J., 114 Rosenbaum, James, 141 Sabo, Martin Olav, 138 Sawhill, Isabel V., 129 Sawicky, Max B., 46 Schorr, Lisbeth B., 6 Schulman, Karen, 130 Self-Sufficiency Program (SSP; Canada), 43, 45, 50–51 Service learning, 62–63, 66, 70, 71 SEU. See Social exclusion Single mothers: effects of welfare reform, 40–42; human capital, 41, 84; income supports, 18; increase in, 76; intelligence of children, 106; poverty rates, 40–41; wages, 41–42. See also Childbearing, out-ofwedlock; Mothers; Teen pregnancy Single parents: African American, 76, 96; children in poverty, 41, 59; educational attainment, 84; efforts to promote marriage, 78–79, 84–87; housing subsidies, 84, 86; human capital, 41, 84; improving relationships, 7; services for, 77–78, 84–85; U.K. programs for, 163–64, 165,

 173; welfare benefits, 85–86; work requirements, 7, 78–79. See also Childbearing, out-of-wedlock; Family strengthening demonstration; Parents Smith, Kristin, 122 Social exclusion, 163, 168–69, 173 SSP. See Self-Sufficiency Program State Child Health Insurance Program, 7, 116 States: child care subsidies, 119; child care tax credits, 120–21; early education systems, 122–24, 130; Head Start funding, 117; lotteries, 121, 123–24; preschool funding, 121–22, 123–24, 130; profamily policies, 97; taxes, 121 Stoney, Louise, 122 Supreme Court, U.S., 141 Sure Start program, 167 Surgeon general, 94, 97 Survey of Income and Program Participation, 122 Sutherland, Holly, 170, 175 TANF. See Temporary Assistance for Needy Families Taxes: U.K. policies, 164, 165, 173, 174; burden on low-income working families, 47; child care credits, 46, 120–21, 126, 130; child care deduction, 18, 28; dependent exemption, 46; family credit proposal, 46–47, 51–52; incentives for working, 45, 46; payroll, 42, 45, 47; property, 121; reform proposals, 46–47; reforms in U.K child poverty reduction initiative, 157, 163, 164, 165, 170–71; sales, 121; U.K. income, 163. See also Child tax credit; Earned income tax credit Teenagers. See Adolescents; Teen pregnancy

1269-12 Index 04/18/03 17:29 Page 193

 Teen Outreach Program (TOP), 62, 68, 71 Teen pregnancy: effects of welfare policies, 93; effects on children, 59, 60, 106; effects on mothers and fathers, 57, 59–60; rates in U.S., 56; societal costs, 59, 60 Teen pregnancy prevention: approaches, 61; benefits of, 56, 57, 60, 69–70; controversies about, 71–72; current programs, 62, 94, 97, 98; effective programs, 62; proposal, 7; whole family approach, 98. See also After-school programs Templeton, Janice, 61 Temporary Assistance for Needy Families (TANF): background, 27; block grants, 119; changes to promote marriage, 78–79, 84; child care subsidies, 127; effects on marriage rates, 86; eligibility for Head Start, 117; reducing aid amounts with child allowance, 28, 29; work exemptions, 23 Thomas, Adam, 129 Thomas, Duncan, 115 Title I, ESEA, 93, 102, 106, 120 TOP. See Teen Outreach Program 21st Century Community Learning Center, 57, 58, 67, 68, 71 Unemployment insurance, 49, 162 UNICEF, 160 United Kingdom: Child Benefit, 164, 165, 174; child care subsidies, 164; child poverty, 160–61, 169; child tax credits, 164, 165, 173, 174; education spending, 169, 174; health spending, 169, 174; income taxes, 163; low birth weights and intelligence, 107; minimum wage, 164; National Strategy for Neighbourhood Renewal, 169; New Deal for

193 Communities, 168; New Deal for Lone Parents, 163–64; New Deal for the Long-Term Unemployed, 163; New Deal for Young People, 163; New Labour, 161–63, 164, 168, 172; public attitudes toward poverty, 158–59, 160; relative poverty, 158, 159–60, 169; social exclusion unit, 168–69; unemployment benefits, 162; unemployment rates, 170; welfare programs, 161–62; working families tax credit, 164 United Kingdom child poverty reduction initiative: announcement, 156; baby bonds proposal, 168; background, 158–62; benefit system reforms, 157, 163–66, 170–71, 172; goals, 156–57, 171–72; impact, 169–71; investment in children, 166–68; lessons for U.S., 172–76; new programs, 157, 167; policy instruments, 161–62; short-term objectives, 157, 160; spending, 3, 166, 167; tax reforms, 157, 163, 164, 165, 170–71, 172 Wages: minimum, 44–45, 49, 164; for single mothers leaving welfare, 41–42. See also Incomes Walker, Gary, 72 Welfare programs: AFDC, 27, 86; changes to promote marriage, 84–87; for low-income working families, 41; in U.K., 161–62. See also Temporary Assistance for Needy Families; Work requirements Welfare reform: earnings supplementation programs, 42–44; effects on children, 25, 43; effects on family incomes, 41–42; effects on maternal earnings, 40; effects on maternal

1269-12 Index 04/18/03 17:29 Page 194



194 employment, 40; evaluations, 42–44; incentives for working, 49; political support, 27; preferences for single parents, 84–86; profamily aspects, 76, 95, 97; in U.K., 157, 163–66, 170–71, 172, 173, 175; work requirements, 23 WFTC. See Working families tax credit Whites: benefits of Head Start, 115; births to unmarried parents, 76 Whole family policy proposal, 94–99, 110; components, 96–97; costs,

98–99; goals, 94, 95, 98; model program, 97; rationale, 94, 99–110 Wilson, William Julius, 139 Working families tax credit (WFTC), 164 Work requirements: changes in welfare reform, 23; effects on adolescent children, 65–66; exemptions for mothers of infants, 18, 19, 23; single parents, 7, 78–79; two-parent families, 78–79. See also Employment Youth. See Adolescents

1269-13 Trustees 04/18/03 17:30 Page 195

THE BROOKINGS INSTITUTION

The Brookings Institution is an independent organization devoted to nonpartisan research, education, and publication in economics, governance, foreign policy, and the social sciences generally. Its principal purposes are to aid in the development of sound public policies and to promote public understanding of issues of national importance. The Institution was founded on December 8, 1927, to merge the activities of the Institute for Government Research, founded in 1916, the Institute of Economics, founded in 1922, and the Robert Brookings Graduate School of Economics and Government, founded in 1924. The Institution maintains a position of neutrality on issues of public policy. Interpretations or conclusions in Brookings publications should be understood to be solely those of the authors.

Board of Trustees James A. Johnson Chairman Strobe Talbott President Elizabeth E. Bailey Zoë Baird Alan R. Batkin James W. Cicconi Arthur B. Culvahouse Jr. Alan M. Dachs Robert A. Day Kenneth M. Duberstein Lawrence K. Fish Richard W. Fisher Cyrus F. Freidheim Jr.

Bart Friedman David Friend Ann M. Fudge Jeffrey W. Greenberg Brian L. Greenspun Lee H. Hamilton William A. Haseltine Teresa Heinz Samuel Hellman Joel Z. Hyatt Shirley Ann Jackson Robert L. Johnson Ann Dibble Jordan Michael H. Jordan Marie L. Knowles Mario M. Morino

William A. Owens Frank H. Pearl John Edward Porter Steven Rattner Rozanne L. Ridgway Judith Rodin Warren B. Rudman Haim Saban Leonard D. Schaeffer Joan E. Spero Lawrence H. Summers John L. Thornton Vincent J. Trosino Beatrice W. Welters Stephen M. Wolf Daniel Yergin

Andrew Heiskell F. Warren Hellman Robert A. Helman Roy M. Huffington Vernon E. Jordan Jr. Breene M. Kerr James T. Lynn Jessica Tuchman Mathews David O. Maxwell Donald F. McHenry Robert S. McNamara Mary Patterson McPherson Arjay Miller

Constance Berry Newman Maconda Brown O’Connor Samuel Pisar J. Woodward Redmond Charles W. Robinson James D. Robinson III B. Francis Saul II Ralph S. Saul Henry B. Schacht Michael P. Schulhof John C. Whitehead James D. Wolfensohn Ezra K. Zilkha

Honorary Trustees Leonard Abramson Rex J. Bates Louis W. Cabot A. W. Clausen William T. Coleman Jr. Lloyd N. Cutler D. Ronald Daniel Bruce B. Dayton Charles W. Duncan Jr. Walter Y. Elisha Robert F. Erburu Henry Louis Gates Jr. Robert D. Haas