Table Of Content

Poverty and Program Participation among Immigrant Children Poverty and Program Participation among Immigrant Children George J. Borjas Summary Researchers have long known that poverty in childhood is linked with a range of negative adult socioeconomic outcomes, from lower educational achievement and behavioral problems to lower earnings in the labor market. But few researchers have explored whether exposure to a disadvantaged background affects immigrant children and native children differently. George Borjas uses Current Population Survey (CPS) data on two specific indicators of poverty—the poverty rate and the rate of participation in public assistance programs—to begin answering that question. He finds that immigrant children have significantly higher rates both of poverty and of program participation than do native children. Nearly half of immigrant children are being raised in households that receive some type of public assistance, compared with roughly one-third of native children. Although the shares of immigrant and native children living in poverty are lower, the rate for immigrant children is nonetheless about 15 percentage points higher than that for native children—about the same as the gap in public assistance. Poverty and program participation rates among different groups of immigrant children also vary widely, depending in part on place of birth (foreign- or U.S.-born), parents (immigrant or native), and national origin. According to the CPS data, these native-immigrant differences persist into young adulthood. In particular, the program participation and poverty status of immigrant children is strongly correlated with their program participation and poverty status when they become young adults. But it is not possible, says Borjas, to tell whether the link results from a set of permanent factors associ- ated with specific individuals or groups that tends to lead to “good” or “bad” outcomes systemati- cally over time or from exposure during childhood to adverse socioeconomic outcomes, such as poverty or welfare dependency. Future research must explore the causal impact of childhood poverty on immigrant adult outcomes and why it might differ between immigrant and native families. Developing successful policies to address problems caused by the intergenerational breeding of poverty and program participation in the immigrant population depends on understanding this causal mechanism. www.futureofchildren.org George J. Borjas is the Robert W. Scrivner Professor of Economics and Social Policy at the John F. Kennedy School of Government at Harvard University and a research associate of the National Bureau of Economic Research. VOL. 21 / NO. 1 / SPRING 2011 247 George J. Borjas Poverty in childhood has long grew from 4.7 percent to 12.9 percent—an been recognized as a determi- increase that presages rapid growth in the nant of a wide range of nega- next few decades in the number of chil- tive socioeconomic outcomes dren born in the United States with at least from lower educational achieve- one foreign-born parent.1 In an important ment and behavioral problems to lower sense, the close link between the skills of earnings in the labor market. But few parents and those of their children suggests researchers have explored whether childhood that current immigration policy has already poverty affects native and immigrant children determined the skill endowment of the work- differently. In this article, I use data on two force for the next two or three generations. specific indicators of poverty—the poverty Therefore, understanding both the impact rate and the rate of participation in public of immigration and the likely future trends assistance programs—to begin answering that in socioeconomic conditions for a large and question. The data suggest that the program growing segment of our population requires a participation rate is significantly higher for careful study of “the coming of age” of immi- immigrant children than for native children. grant children. Nearly half of immigrant children—a remarkably large fraction—are being raised in Filling a Gap in the Research households that receive some type of public Much of the immigration literature in the assistance, compared with roughly one-third social sciences, however, focuses on trends of native children. Although the shares of in the relative skills of immigrants or deter- immigrant and native children living in mining how immigration alters the economic poverty are lower, the rate for immigrant opportunities available to the native-born children is nonetheless about 15 percentage population. Some immigration studies exam- points higher than that for native children— ine the social mobility of immigrant house- the same as the gap for public assistance. The holds.2 The notion that social, cultural, and evidence also suggests that these native- economic differences between immigrants immigrant differences persist into young and natives fade over the course of a few adulthood. In particular, the program partici- generations is the essence of the melting- pation and poverty status of immigrant pot hypothesis. Over time, the children children is strongly correlated with their and grandchildren of immigrants tend to program participation and poverty status a move out of ethnic enclaves, discard their decade later when they become young adults. social and cultural background, and become It is not possible, however, to tell whether indistinguishable from the native population. this link results from a long-term persistence Estimates of the rate of intergenerational in socioeconomic outcomes or is a causal convergence across the many national origin effect of the adverse exposure that occurs groups suggests that although the melting pot during the childhood years. operates, the economic differences observed among the various groups may not dissolve The exact implications of these findings are for at least two or three generations. not yet completely understood, but they have potentially significant policy and social Although this long-run perspective is insight- ramifications. Over the past four decades, ful, the examination of the well-being of the foreign-born share of the U.S. population immigrant children changes the focus of 248 THE FUTURE OF CHILDREN Poverty and Program Participation among Immigrant Children analysis from the rate of intergenerational Finally, the literature shows that the impact social mobility to a host of short-run concerns of childhood poverty persists into adulthood.6 that can increase our understanding of the A poverty spell during childhood increases experiences of immigrant households. For the probability that the adult will have lower example, how does the background of earnings and greatly increases the probability immigrant families influence the socioeco- that the adult will also experience a poverty nomic outcomes for immigrant children? Do spell. In other words, childhood poverty these background characteristics explain a breeds adult poverty. significant part of the observed differences between native and immigrant children and Much of the literature examining the inci- among the various national origin groups dence of childhood poverty and the link within the immigrant population? between childhood poverty and other socioeconomic outcomes ignores the potential One such background characteristic is differences that may exist between immigrant poverty. A large literature has isolated the and native children. The frequency and the incidence and timing of poverty during length of poverty spells likely differ between childhood as a crucial determinant of a wide immigrant and native children (as well as array of socioeconomic outcomes both in the among the national origin groups that make short and long run.3 For example, evidence up the immigrant population). Moreover, shows that growing up in a poor household child poverty could potentially have different can adversely affect a child’s academic consequences for immigrant and native chil- achievement. Similarly, poverty correlates dren. Put differently, exposure to a disadvan- strongly and negatively to the probability that taged background may imply different things a child graduates from high school. Some for different groups of children, particularly studies attempting to uncover the root causes because the immigrant experience introduces of these adverse outcomes have found distinct factors that native children avoid evidence suggesting that poverty affects (such as a temporary family separation result- social and emotional development, with ing from the vagaries of immigration law). children raised in poverty having a higher incidence of behavioral problems that are The Population of Immigrant likely to mar the school experience and lead Children: A Descriptive Analysis to poorer academic outcomes.4 The U.S. Census Bureau began to collect information on the birthplace of participants That the negative impact of childhood and their parents in the Current Population poverty extends well beyond academic Survey (CPS) in 1994. The Annual achievement is also well known. Poor chil- Demographic Files of the CPS (also known as dren, for instance, experience less favorable the March Supplements) provide detailed health outcomes, including a higher propen- information about whether a family’s total sity for low birth weight and a higher mortal- income is below the poverty threshold and ity rate in the first month of life.5 The whether the household participated in various health-related consequences continue into types of social assistance programs during the adolescence. Poorer children have a greater calendar year before the survey. The evidence risk of experiencing accidents and injuries summarized below for immigrant and native and a higher probability of teen childbearing. households over the past fifteen years is VOL. 21 / NO. 1 / SPRING 2011 249 George J. Borjas Figure 1. Trends in the Share of Immigrant Children, 1994–2009 20 1 immigrant 18 parent 16 2 immigrant parents 14 U.S.-born, 12 2 immigrant parents ent 10 Foreign-born, erc 2 immigrant parents P 8 6 4 2 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 Source: Author’s calculations from the 1994–2009 March Current Population Surveys. Note: The population of children includes all persons aged seventeen or less. drawn from those data in the 1994–2009 CPS throughout the analysis. Finally, the residual March Supplements. The observed trends group is composed of “native” children— during this period reflect the combined U.S.-born children whose parents also were impact of the enactment of welfare reform born in the United States. Figure 1 summa- legislation in 1996, the continuation of a high rizes the trend since 1994 in the relative size volume of legal and illegal immigration into of the various groups of immigrant children the United States, and a lengthy economic aged seventeen or younger, classified accord- boom followed abruptly by a deep recession. ing to the birthplace of the parents and of the children.8 A crucial first step is the definition of “immigrant children.” The definition used in most The fraction of children who have at least one of the other articles in this volume defines immigrant parent has increased substantially, immigrant children as those who are foreign- from 17.5 percent of all children in 1994 to born and migrate to the United States with 23.2 percent in 2009. The fraction of mixed- their foreign-born parents and those who are parent children in the population hovered U.S.-born to one or two immigrant (foreign- around 6 percent throughout the entire born) parents. I place immigrant children sample period, while the fraction of children into three groups: children who have one with two immigrant parents rose from 11.6 to immigrant parent (here called “mixed par- 16.9 percent. The rate of increase in the share ents”);7 foreign-born children who have two of immigrant children is much higher than immigrant parents; and U.S.-born children the corresponding increase in the share of who have two immigrant parents. The differ- foreign-born persons in the total population. ences in socioeconomic outcomes between In 1994, 9.6 percent of the total U.S. popula- these three groups of immigrant children tion was foreign-born; by 2009, the foreign- are important, so they will be differentiated born share had increased to 12.9 percent. 250 THE FUTURE OF CHILDREN Poverty and Program Participation among Immigrant Children Figure 2. Trends in the Poverty Rate of Children, 1994–2009 50 Mixed parents 45 40 Foreign-born, 2 immigrant parents 35 U.S.-born, 30 2 immigrant parents ent 25 Native parents c er P 20 15 10 5 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 Source: Author’s calculations from the 1994–2009 March Current Population Surveys. Note: The poverty rate gives the fraction of households with incomes below the poverty threshold. The vast majority of immigrant children— rates of immigrant children, much less study around 80 percent—are, in fact, born in the the long-term consequences of a disadvan- United States.9 While the fraction of immi- taged childhood in an immigrant household. grant children born abroad has remained Researchers and policy makers can thus view relatively constant (around 4 percent of all this article as a first attempt to document children throughout the period), the fraction issues related to poverty and program partici- of immigrant children born in the United pation among immigrant households in the States rose dramatically, from under 12 past decade and to reveal the trends that may percent of all children in 1994 to almost 17 become important determinants of future percent by 2009. outcomes in this population. Poverty and Program Participation Rates The poverty rate is defined as the fraction The socioeconomic background of the house- of children in a particular group that is holds where immigrant children are raised being raised in households where fam- is likely to have lasting influence on a wide ily income is below the poverty threshold. array of outcomes as these children grow up, Figure 2 illustrates the trends in poverty rates complete their education, and enter the labor among the various groups of children being market. As noted, a crucial variable that may examined. Note, for example, that neither have long-term detriments is the likelihood the level nor the trend in poverty rates differs that the immigrant child grows up in a poor household. Although a large literature docu- much between native and mixed-parent chil- ments the consequences of childhood poverty dren. In 2009, about 17 percent of children on a wide array of socioeconomic outcomes, in both of these groups were being raised the existing studies do not typically examine in households where income fell below the the poverty or public assistance participation poverty threshold. VOL. 21 / NO. 1 / SPRING 2011 251 George J. Borjas In contrast, the poverty rate of children with CPS data, which report whether anyone in two immigrant parents is higher, particularly the household received cash benefits or food for immigrant children born abroad. In 2009, stamps (now known as the Supplemental the poverty rate of U.S.-born children with Nutrition Assistance Program, or SNAP) or two immigrant parents was 28.5 percent, was enrolled in the Medicaid program. The while that for foreign-born children was 31.6 summary definition of program participation percent. The figure also reveals a noticeable that I initially use in the analysis indicates relative decline in the poverty rate of these whether anyone in the household received two groups of children between 1996 and assistance from any of these three programs. 2000 (which may reflect the economic boom The top panel of figure 3 illustrates the trend of the late 1990s or be related to the timing in this measure of the program participation of the welfare reform legislation). Finally, the rate during the sample period for the four figure shows that the poverty rate of these groups of children in the data: native children, children has increased rapidly in the past few mixed-parent children, U.S.-born children years, relative to those of children with native with two immigrant parents, and foreign-born or mixed parents, perhaps reflecting the wors- children with two immigrant parents. ening economic conditions after 2007. For instance, between 2007 and 2009 the pov- Figure 3 reveals a number of interesting erty rate barely rose for native children but results. First, as with the poverty rate, pro- increased by around 5 percentage points for gram participation rates differ little between U.S.-born children with two immigrant par- native children and children of mixed parent- ents and by 6 percentage points for foreign- age. Both the level and trend of participation born children with two immigrant parents. rates in these groups are remarkably similar during 1994–2009. In contrast, whether To what extent do immigrant children live they were U.S.-born or foreign-born, chil- in households that receive public assistance? dren with two immigrant parents live in That question is interesting for two reasons. households that overall have higher rates of First, some of this assistance presumably program participation. In 2009, the program helps to lower the measured poverty rate in participation rate was 51.5 percent for the immigrant households.10 Second, exposure U.S.-born children and 38.6 percent for to the public assistance infrastructure during the foreign-born children. In other words, childhood may itself have long-term conse- slightly over half of all U.S.-born children quences, some harmful and some beneficial. with immigrant parents lived in a household It may, for example, introduce the seeds of a where someone received some type of assis- culture of dependency that may persist into tance. In contrast, the participation rate for adulthood. Or it may, in some forms, such as native or mixed-parent children was around Medicaid, serve as a form of human capital 33 percent. investment, leading to healthier and more favorable health and economic outcomes as The data show that foreign-born children the children grow up.11 have the highest measured poverty rate but that U.S.-born children with immigrant par- To document the extent to which immigrant ents have the highest program participation children are exposed to welfare programs rate. The latter finding is not surprising: it is during their childhood, I turn again to the the citizen children in these households who 252 THE FUTURE OF CHILDREN Poverty and Program Participation among Immigrant Children Figure 3. Trends in Program Participation of Children, 1994–2009 60 Mixed parents Program participation rates, including Medicaid 50 Foreign-born, 2 immigrant parents 40 U.S.-born, 2 immigrant parents ent 30 Native parents c er P 20 10 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 35 Mixed parents Program participation rates, excluding Medicaid 30 Foreign-born, 2 immigrant parents 25 U.S.-born, 2 immigrant parents 20 nt Native parents e c Per 15 10 5 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 Source: Author’s calculations from the 1994–2009 March Current Population Surveys. Note: The program participation rate gives the fraction of children living in households that received cash assistance, SNAP benefits, or Medicaid (in the top panel), or cash assistance and SNAP benefits (in the bottom panel). qualify for various types of public assistance. participation trends, this one between But the differential outcomes in program children with two immigrant parents and participation and poverty between these two other children. Even though children with groups of children hint at the possibility that two immigrant parents have a higher par- some of the public assistance restrictions ticipation rate throughout the entire fifteen- imposed on children born abroad have impor- year period, that rate declines dramatically tant consequences on the socioeconomic sta- immediately after enactment of welfare tus of the households in which they grow up. reform legislation in 1996 (and this decline is noticeably steeper for the foreign-born The top panel of figure 3 reveals another children). The Personal Responsibility and interesting difference in the program Work Opportunity Reconciliation Act, or VOL. 21 / NO. 1 / SPRING 2011 253 George J. Borjas Figure 4. Differences in Poverty Rates by National Origin of Immigrant Children, 1994–2009 70 Mexico 60 El Salvador Dominican Republic 50 China Philippines 40 nt Vietnam e c er 30 India P 20 10 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 Source: Author’s calculations from the 1994–2009 March Current Population Surveys. Note: The population of immigrant children includes all persons aged seventeen or less whose parents were born outside the United States or its possessions. PRWORA, led to a relatively steeper drop household receives some type of public in immigrant participation in welfare pro- assistance. After Congress enacted welfare grams, perhaps because of the “chilling reform, it substantially expanded the State effect” of several provisions in the statute Children’s Health Insurance Program that restricted noncitizen eligibility for these (SCHIP), which covers children who lack programs.12 The trends illustrated in the health insurance but whose family income is figure suggest the presence of this chill- too high to make them eligible for Medicaid. ing effect in the families of children with Because the CPS information on whether a two immigrant parents, particularly in the household receives Medicaid assistance families of foreign-born children (children includes information on whether the house- who are not U.S. citizens and therefore do hold participates in the SCHIP program, not qualify for many types of assistance in many of the trends in Medicaid participation the post-PRWORA period). Note further the revealed by the CPS could reflect the cre- growing divergence in recent years between ation and rapid growth of the SCHIP pro- U.S.-born children with two immigrant par- gram after welfare reform. ents, who have experienced a very rapid rise in participation rates, and all other groups of In fact, as the bottom panel of figure 3 shows, children. In fact, the figure clearly indicates the trends in program participation rates that this group of children has the fastest- across the various types of households are rising rate of program participation among quite different when the definition of program the various groups in the analysis. participation focuses only on whether the household receives cash or SNAP benefits. At Many of the trends revealed in the top panel the beginning of the period, both groups of of figure 3 are driven by the inclusion of immigrant children had higher participation Medicaid in the definition of whether the rates than either native or mixed-parentage 254 THE FUTURE OF CHILDREN Poverty and Program Participation among Immigrant Children Figure 5. Differences in Program Participation by National Origin of Immigrant Children, 1994--2009 Program participation rates, including Medicaid 70 Mexico 60 El Salvador Dominican Republic 50 China Philippines 40 nt Vietnam e c er 30 India P 20 10 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 60 Mexico Program participation rates, excluding Medicaid El Salvador 50 Dominican Republic China 40 Philippines ent 30 Vietnam c er India P 20 10 0 1994 1996 1998 2000 2002 2004 2006 2008 2010 Source: Author’s calculations from the 1994–2009 March Current Population Surveys. Note: The population of immigrant children includes all persons aged seventeen or less whose parents were born outside the United States or its possessions. children. The enactment of PRWORA led to The immigration literature has documented a very rapid decline in the participation rate of substantial differences in a wide array of children with two immigrant parents, particu- socioeconomic outcomes across the various larly that of foreign-born children. By the end national origin groups that compose the of the period, foreign-born children have the entire immigrant population; these outcomes lowest rate of program participation among include educational attainment, wages, labor the four groups examined, while the participa- supply, and participation in public assistance tion rate of U.S.-born children with immigrant programs. Not surprisingly, poverty rates and parents is essentially the same as that of native program participation rates also differ and mixed-parentage children (though rising substantially by national origin groups among very rapidly). children with two immigrant parents.13 VOL. 21 / NO. 1 / SPRING 2011 255 George J. Borjas Because the sample size for many national children from Mexico, and 32.6 percent origin groups is so small when foreign-born of children from the Dominican Republic children are examined separately from those received either cash or SNAP benefits.15 born in the United States, the analysis pools together all children with two immigrant The national origin groups with the largest parents into a single group. The national measured poverty and program participation origin of the foreign-born children is, of rates also tend to be the largest immigrant course, determined by the child’s birthplace. groups. In 2009, for example, 46.9 percent of That of the U.S.-born children is determined all children with two immigrant parents were by parental birthplace as follows. About 90 of Mexican origin. To the extent that poverty percent of these children are being raised in status and program participation among these households where the birthplace of the father children are indicators of a young population and mother are the same. For the remaining at risk, figures 4 and 5 suggest the poten- 10 percent of the children, the immigrant tial for the creation of a large population of mother’s birthplace determines the national disadvantaged persons as these children grow origin of the child.14 into adulthood. In fact, as I show below, the data indicate the presence of persistent eth- As figure 4 illustrates, some of the differ- nic differences in program participation and ences in the poverty rates among some of the poverty status as the children of immigrants largest national origin groups in the data are transition into young adulthood. remarkably large. In 2009, only about 6 or 7 percent of the immigrant children from India Aging and Cohort Influences on Poverty or the Philippines lived in households that and Participation Rates were below the poverty level, compared with Research on immigrant economic perfor- nearly 40 percent of children in households mance has provided two insights that now from Mexico or the Dominican Republic. serve as “stylized facts” in the immigration debate. First, the typical immigrant worker in Figure 5 shows that, as with poverty rates, the United States suffers a sizable earnings the disparity across national origin groups disadvantage (relative to native-born workers) in the two alternative measures of program upon arrival, but some of this disadvantage participation rates (including and excluding disappears with time spent in the United Medicaid) is also large. For example, in 2009, States (an assimilation, or “aging,” effect). the participation rate (including Medicaid) Second, skills differ across immigrant of immigrant children from India was about cohorts, with more recent cohorts being 14.6 percent. In contrast, 21.5 percent of relatively less skilled than earlier cohorts children in Filipino households and more (a “cohort effect”). The question is whether than 60 percent of children from Mexico and aging and cohort effects serve to attenuate or the Dominican Republic received assistance. exacerbate the differences in poverty rates in The disparity among national origin groups the sample of children of immigrants. is equally large in the bottom panel of the figure, which excludes Medicaid from the The top panel of table 1 “tracks” specific definition of public assistance. In 2009, 2.5 age cohorts of U.S.-born children of immi- percent of children from India, 11.5 percent grants across CPS cross-sections to determine of children from Vietnam, 23.2 percent of how the poverty rate changes for different 256 THE FUTURE OF CHILDREN