PRB Discuss Online: The Long-Term Effects of Childhood Poverty in the United States

(October 2010) Most poor children achieve less, exhibit more problem behaviors, and are less healthy than children raised in more-affluent families. Looking beyond these well-known correlations between poverty and negative outcomes in childhood, recent studies have assessed the effects of childhood poverty in the United States on later attainment and health. During a PRB Discuss Online, Greg Duncan, professor of education at the University of California, Irvine, and the current president of the Society for Research in Child Development, answered questions from participants about the effects of childhood poverty in the United States on later attainment and health.


Oct. 28, 2010, 1 p.m. (EDT)


Transcript of Questions and Answers


Richard Cincotta: Are there racial, or locational (rural-urban, ethnic vs. mixed-race neighborhood, high-rise vs. house, etc.) differences that show up in findings? I know that you may not have the data to fully answer that question—how about: Are there any interesting socio-environmental conditions that mediate your most general findings?
Greg Duncan: Good question. We checked to see if the relationship between early income and later productivity and health differed by gender or race. We found no significant differences at all. Recent research by Brian Jacob and Jens Ludwig found evidence that generous housing vouchers helped most the education outcomes for poor children in Chicago’s worst neighborhoods. In our research, the story was about main effects of low income and not the interactions.


William O’Hare: I thought I read somewhere that must poor adults did not grow up in poverty. The odds of being a poor adult are higher but the number of non-poor children is so much larger that in shear numbers persons not growing up in poverty overwhelm the higher odds for poor children. If that is true, what does in mean about where we should focus our attention to prevent or limit adult poverty?
Greg Duncan: It is indeed true that most children who grow up experiencing poverty do not end up as poor adults. In fact, that study used the same data base—the Panel Study of Income Dynamics—as did our current set of studies. Our recent studies are based on the idea that poverty has selective effects on life chances—that the timing of poverty in childhood matters a lot and that poverty affects some kind of outcomes more than others. Abundant research has suggested that poverty early in childhood is most harmful, and that achievement is more sensitive to economic deprivation than is problem behavior (e.g., delinquency, teen births). Our recent studies are the first to use data that span a long enough time to link poverty as early as the prenatal year to adult outcomes as late as age 37. We again find that early poverty—perhaps even poverty during pregnancy and infancy—matters the most. But we also find that the early poverty impacts on labor market success may have as much to do with health as anything else. This complements recent epidemiological research showing that very stressful conditions during pregnancy and infancy can “program” the body in a way that leads to bad health outcomes later on.


J. Herman Blake: I have several interests and will put them in separate questions. What are the long-term effects of childhood poverty on oral health of adolescents and adults? Are there racial or ethnic variations? I am interested in both preventive approaches to oral health and later treatments for oral health problems.
Greg Duncan: Regrettably, we do not have data on oral health for adolescents and adults. Our adult health measures are, among others, self-reported weight and height, and diagnoses of arthritis, hypertension and diabetes. An interesting recent study of socioeconomic disparities and the oral health of children is: W. Thomas Boyce, Pamela K. Den Besten, Juliet Stamperdahl, Ling Zhan, Yebin Jiang, Nancy E. Adler, John D. Featherstone, Social inequalities in childhood dental caries: The convergent roles of stress, bacteria and disadvantage, Social Science & Medicine, Volume 71, Issue 9, November 2010, Pages 1644-1652.

Jane Cover: For which adult outcomes of child poverty is the evidence quite strong, and what are the areas or outcomes for which the evidence is more mixed, i.e., what do we know definitively (or as definitively as social science permits), and what is still up in the air?
Greg Duncan: Good question. The evidence linking low income to achievement in childhood is very solid, based on several different data sets and uses a number of experimental and longitudinal methods. But you ask about adulthood. Almost all of the existing childhood poverty/adult outcome studies are based on data that begins in adolescence. There are reasons to believe that poverty early in childhood is what matters the most, but almost no studies have followed individuals from conception to adulthood. The Panel Study of Income Dynamics (PSID) is one of them. Several Scandinavian countries have administrative data that could serve this purpose as well. There are several wonderful cohort studies in the UK but they did not measure income early in life. Our results are based on the PSID and we are in the process of replicating our results using data from the Norwegian data registers. Both provide terrific income histories across all of childhood but only limited information on adult outcomes. The PSID is the best here, with information on earnings, criminal behavior and health. The PSID is a public use data set, so we hope that people will attempt to replicate and extend what we have done.

Juanita Tamayo Lott: My research concern is along the lines of Bill O’Hare’s. Are long-term findings based on cross-sectional or longitudinal data and do they consider context such as economic depression, an era of affluence, civil or international war, natural calamaties, or persistent structural inequities? Does such research delineate intergenerational sets of populations by those who are born and raised poor and continue to be poor in adulthood; those who are poor as children but not as adults; and children who are not born or raised poor but become so as adults. What is their distribution? To what extent do researchers, service providers, and policy makers assume, even if unconsciously, that the poor will continue to be poor?
Greg Duncan: Our study, by the way, has been published, so you and other readers can see the details for yourselves. The reference is Duncan, G. J., Ziol-Guest, K. M. and Kalil, A. (2010), Early-Childhood Poverty and Adult Attainment, Behavior, and Health. Child Development, 81: 306–325. Briefly, we are able to use 38-year longitudinal data on a national sample of U.S. children born between 1968 and 1975. The data provide a year by year accounting of family income from the prenatal year through age 15. Our analyses related adult outcomes to income very early in childhood (the prenatal year through age 5), but we controlled statistically for differences in income between ages 6 and 15, as well as a host of demographic characteristics at birth such as family structure and parental education and test scores. Since the data are from the U.S., there is a limited range of disruptive historical events and our results should not be generalized to other countries in different stages of development. We approached the research with as few preconceptions as possible. As author of the 1984 book Years of Poverty, Years of Plenty and the former principal investigator for the data used in the study, I am keenly aware of the surprising amount of economic mobility, both within and across generations, that has been documented with these and other data. The study’s website ( details many of the studies that provide answers to some of your other questions.

J. Herman Blake: How does your study of the long-term effects of childhood poverty help us understand the issues reflected in the question of a “culture of poverty?” Can we analyze this issue without becoming enmeshed in social controversy?
Greg Duncan: Culture of poverty arguments have cycled through poverty debates for a long time. Did you see the recent New York Times article about them? It was written by Patricia Cohen and appeared on October 18, 2010 ( Our studies attempt to isolate the role of income rather than culture in affecting children’s life chances. In finding that income early in childhood seems to matter, we are not saying that it matters more than culture or parenting or neighborhood conditions. Income is important to understand because it is one of the few things that policy CAN change. Of course, cultural arguments are invoked every time some anti-poverty policy is being debated, but the work supports and welfare reforms of the 1990s proved that Americans are willing to spend tens of billions more dollars on helping to boost the incomes of poor families, provided that the help led to more labor market involvement and less welfare reliance. So that I think that it is possible to fashion helpful policies that are consistent with the values of most Americans. Even if, given the budget deficit, we can’t spend more money on low-income families, our research suggests that existing programs such as the Earned Income Tax Credit but be more effective if families with young children received relatively more and families with older children received relatively less.

Okunrinmeta Olamide Adenike: What is the effect of Long-Term Childhood Poverty on the critical thinking and Intellectual performance of affected children?
Greg Duncan: Evidence from a number of studies suggests that increasing the family incomes of poor families will lead to gains in young children’s cognitive and academic performance. Many correlational studies have shown that. But so too did the Negative Income Tax experiment, which assigned families at random to get higher incomes or not. Quasi-experimental studies of the mid-1990s expansions of the Earned Income Tax Credit found links to higher child test scores, as did a study of Canada’s expansion of its child benefits. We know less about why. Several studies point to what money can buy in terms of enriching children’s early learning experiences, both inside the home (e.g., more books) and outside (e.g., better-quality child care).

Allan Hill: Are there intervention studies that reducing household poverty improves childhood and post-childhood outcomes net of other changes (e.g. secular health improvements)? Are there panel or intervention studies in LDCs that address this question?
Greg Duncan: The income-based intervention studies (some random assignment and some strong quasi-experimental design) compare treatment and control groups in the same cohort, so secular changes are affecting both groups equally. In LDCs there are a number of conditional cash transfer programs with positive results. Here again treatment and control groups in the same cohort are being compared.


Lisa: There is much research available on the impact of long-term poverty,or generational poverty, on outcomes for children. What, if any, are the differences in outcomes for children of sudden, possibly shorter-term poverty for children? I would like to be able to talk about the impact of poverty caused by the current economic downturn on children in some of our publications coming out in the next year or two.
Greg Duncan: There haven’t been good studies of very short term (several months to one or two years) poverty spells. If the spells are several years in duration, and occur during early childhood, then our results would indicate that there are likely negative impacts. There is a separate literature on the impact of unemployment on children. Ariel Kalil at the University of Chicago, who was a coauthor on my poverty studies, has contributed to this literature and knows it well.


Meskerem Bekele, Ethiopia: It is good to discuss with a professional like you. And I want to thank PRB for their facilitation. The issue is interesting and surprising to me. I do not think like this about poverty before. . . .to me, childhood poverty makes someone strong. Even most of our country’s “successful” persons talk about themselves likes this. When the research shows that “…childhood poverty have a substantial negative association with adult earnings, work hours, and certain health conditions…”—what do you mean by “work hours”?
Greg Duncan: These sessions are enjoyable for us as well. Families experience a surprising amount of income fluctuation from one year to the next. Our research indicates that timing is everything—if it occurs early in childhood then it is likely to be harmful. Poverty spells at other times, perhaps because families are better able to cope with them, appear to have benign impacts. Our look at work hours used information on the children’s report of their adult work hours between ages 25 and 37. We average annual work hours over that time and tried to see if they were sensitive to early spells of poverty. And, yes, they appeared to be.


Usha Natampalli: Do children who face multiple deprivation owing to poverty and hunger have shorter life expectations?
Greg Duncan: We know that there are strong simple associations between family background and longevity—individuals raised in lower socioeconomic status families do not live as long as individuals raised in more affluent families. Our study took a more detailed look at the timing of poverty in childhood and found that early childhood was the most sensitive. We could only follow the children into their 30s so we couldn’t observe mortality. But with early poverty linked to obesity and hypertension, the chances are that these effect might well carry over into life expectancy.


Luis Uribe: This kind of problems in the adulthood-related to child poverty, affect in the same way to the Latino community?
Greg Duncan: The data we used followed children born in the United States between 1968 and 1975, so there were very few Latinos in the sample. We did not observe differences in poverty effects between Blacks and Whites, but we can’t say if Latinos are affected in the same way. As you may know, other studies have indicated that Latinos seem to have better health than their income levels would lead us to expect.


J Kishore: Dear Greg Duncan, it is well known that poverty is causing long term effects of childhood poverty. In India we say it is caste which is linked with poverty and such poor children belong to native castes remain poor performers in every field. International bodies are not recognizing caste in India a problem to a significant amount of children suffering from malnutrition and subsequent poor skill development. What will you suggest in such situation?
Greg Duncan: My research is US based, and I know little about the Indian context. A good source of information about ideas for developing countries is the MIT Poverty Action Lab:


Beth Mattingly: What are the factors that promote resiliency among children exposed to poverty at a very young age? How much do these vary by timing and duration of childhood poverty?
Greg Duncan: Our study was focused on income itself, and we found that poverty in the very early years was the most likely to scar later development. Low income later in childhood mattered much less. So the “timing” issue as it relates to the age of the child appears to be very important. As to promoting resilience, I can think of both parent and child factors. Ensuring that the pregnancy and the child’s early years are not severely stressed involved adequate prenatal care and, after the birth, genuine choice over when the mother returns to the labor force. Workplace leave policies and state welfare programs can work against the needed degree of choice. As to child-based factors, I would put most of my money into early education programs, since a child starting school ready to learn has the greatest chance of later academic successes.


Mil Duncan: What do we know about the individual characteristics, family circumstances and neighborhood conditions of those who escape poverty?
Greg Duncan: Hi Mil. Good (and tough) questions! Despite our focus on intergenerational poverty, research does show that most children who grow up poor or in welfare dependent households are not themselves persistently poor or welfare dependent when they become adults. That said, there is certainly a substantial correlation between the economic standing of parents and grown children. In looking for characteristics that matter the most for escaping, the child’s completed schooling level is by far the most important, with each additional year of education producing close to a 10% increase in lifetime earnings. Evidence is mounting that neighborhood conditions matter much less than family conditions—once you account for the powerful role parent education plays and the selective role that income appears to play, it is hard to see many neighborhood effects. And recent mobility experiments have confirmed that most child outcomes (test scores, for example) fail to improve when children move with their families from very poor to somewhat more affluent neighborhoods.


Dr. Josephine Alumanah: How can the parents/Guardians be empowered? How can poverty be eradicated? Prevention is better than cure, don’t we think?
Greg Duncan: The answer to these questions would fill volumes! My research is US-based, where income supports are closely tied to work, and public education is available to all. If I were to pick one intervention it would be education—beginning in the preschool years with a high-quality, curriculum-based program, and continuing through college. As you probably know, the quality of education has declined with the recession and even before the recession the US had fallen behind many countries in its production of college graduates. This is most worrisome. As to parent empowerment, I spent several years evaluating a highly effective effective work-support program in Milwaukee called New Hope. It offered a bundle of supports—earnings supplements, childcare subsidies, health insurance subsidies and a temporary community service job—for adults willing to commit to full time work. A random-assignment evaluation showed that the children in New Hope families—particularly the boys—did much better than their control-group counterparts. The website documents the program and its impacts (, as does the book Higher Ground: New Hope for Low Income Families and Their Children (


Diana Lavery: Hi Dr. Duncan – I recently read a few studies about the positive effects of social networks on health and well-being. I’m wondering if there were differences in children who grow up in poverty and have strong social networks (such as more “collective” communities, e.g. Hispanic and Asian) vs. children who grow up in poverty and have weak social networks (such as more “individualistic” communities, e.g. white and black). I know that there weren’t statistically significant racial differences in your analysis, but were there differences by strength of social networks? If there were, perhaps this would be useful when considering types of preventative programs.
Greg Duncan: Great question. There is a body of research on the role of neighborhood “collective efficacy”—the degree to which neighbors look out for one another—that is relevant. Rob Sampson, Steve Raudenbush and Tony Earls published a paper in Science that linked collective efficacy to child outcomes (Science 15 August 1997: Vol. 277. no. 5328, pp. 918 – 924). On the other hand, in the mobility experiment I mentioned to Mil Duncan, moving to more affluent neighborhoods increased collective efficacy but did not improve most child outcomes. The website has collected all of the studies associated with this experiment. Some research studies on child outcomes find that, relative to whites and controlling for family circumstances, Asian children do better, but it is hard to infer what it might be about Asian communities or families that makes the difference.


Diana Lavery: Since early childhood poverty is associated with lower health and lower productivity in adulthood, what do you think the best intervention would be to assist children in poverty today? Should interventions be more education oriented, more health/nutrition oriented, or something else?
Greg Duncan: The research I summarized concerned income-based poverty in the United States (Duncan, G. J., Ziol-Guest, K. M. and Kalil, A. (2010), Early-Childhood Poverty and Adult Attainment, Behavior, and Health. Child Development, 81: 306–325). There are several U.S. programs that have the potential for transferring more income to low-income families with young children—the Earned Income Tax Credit and the Child Tax Credit are two prominent examples. Each could be geared to pay relative larger benefits to families with younger children and less to families with older children. Other direct service programs such as Early Childhood Education (e.g., pre-kindergarten) programs could help as well. Whatever we choose, we should support programs with proven track records of providing more benefits than costs.