Child Poverty, Program Participation, and Intergenerational Coresidence: Evidence from the Survey of Income and Program Participation

Child Poverty, Program Participation, and Intergenerational Coresidence: Evidence from the Survey of Income and Program Participation The purpose of this study is to provide an assessment of poverty and program participation among children who reside in households that include elderly adults aged 65 and over. Children and the elderly draw on different pools of public support, fueling recent discussions of the intergenerational tensions regarding the appropriate resource share due each age group. Yet within family households, the “fortunes” of children may be closely tied to that of their elderly relatives. While a sizable literature documents the considerable kinship support systems that link the generations, the implications of these systems for poverty and for program participation have not been extensively considered. Given the trends and debates surrounding public support for these two broad age groups, such a consideration is of considerable policy importance. This study is based on data from the 1990 panel of the Survey of Income and Program Participation (SIPP). Estimates of poverty status and participation in means-tested resource programs are generated for children, and the extent to which these outcomes are associated with elderly coresidence is considered. The results suggest that children who have at least one elderly individual living with them have lower rates of poverty and economic hardship, but that this benefit occurs only for children living with single parents. Social Security and Supplemental Security Income represent the most significant sources of elderly income for this group. As a result of these contributions, children of single parents living with at least one elderly person have a rate of poverty that is a third lower than their counterparts with no elderly coresidents. Living with an elderly individual also is associated with higher chances of living in an owned home as opposed to either public or private rental housing, and, for children living with both parents only, a higher rate of participation in the food stamp program. Child Poverty, Program Participation, and Intergenerational Coresidence: Evidence from the Survey of Income and Program Participation Introduction High poverty rates among children in the United States are regarded as both a personal and national tragedy (Haveman and Wolfe, 1994). The alarming trends in child poverty have prompted a considerable amount of effort in recent years toward developing an understanding of the multiple bases of child well-being. Various sources have attributed declining child well-being to social changes loosely characterized as the disintegration of the nuclear family, especially through rising levels of divorce and out-of-wedlock childbearing (Duncan and Rodgers, 1991; Eggebeen and Lichter, 1991; Hogan and Lichter 1995). As well, limited work opportunities among parents of young children, and welfare and economic support policies that have undermined the cohesiveness of two-parent families have been cited (Moffitt, 1992; Murray, 1993). The associated risks and costs to children in both the shortand long-term have been documented extensively (Haverman and Wolfe, 1994; McLanahan and Sandefur, 1994). An additional element of the child welfare debate surrounds the shifting generational balance in the United States, and its implications for generational claims on public resources. In relative terms, the older population is growing in size while the childhood population is in decline. Observers have argued that due to the shifting age composition, policies and public sentiments have increasingly favored support of the elderly (Axinn and Stern, 1985; Preston, 1984; Treas and Torrecilha, 1995). Pressures placed on limited public resources by the expanding older population have required a renegotiation of public support for other constituencies and, according to this argument, children are often the losers in this calculus. Indeed, poverty rates among children have been rising in recent decades, while during the same time period poverty rates among the elderly have declined (Council of Economic Advisors, 1985; Preston, 1984). Political actions that reduce or threaten public support of children, such as the recent federal welfare legislation, undertaken while protecting programs oriented toward the elderly, such as Social Security and Medicare, are often interpreted as a reflection of these political trends. In a "backlash" to this pattern, the elderly have been characterized as “greedy,” believing themselves entitled to an old age of “subsidized leisure”—apparently willing to ignore the well-being of their grandchildren to ensure their own benefit (Peterson, 1996; and see related discussion in Hirshorn, 1991; Kotlikoff, 1992; Quadagno, 1989). Yet economic, sociological, and demographic evidence suggests that, at least with reference to their own kin, the elderly are far from selfish. Many studies demonstrate the substantial contributions to child well-being provided by older relatives, ranging from occasional child-care and gifts to long-term foster care (Aldous, 1995; Bass and Caro, 1996; Cherlin and Furstenberg, 1986; Hogan, Hao, and Parish, 1990). To a large extent, the elderly appear to be transferring at least some of the economic “benefits” associated with old age to their younger relatives. This process is not limited to the wealthy who bestow upon their young kin ample trust funds and private school tuition. Transfers are particularly critical among families in marginal economic circumstances, who depend on extensive family networks to help meet basic needs. And, while intergenerational transfers among the wealthy are likely asset-based and the result of life-long savings patterns (Kotlikoff, 1988), contributions to children from economically marginal relatives may themselves be in part transfer-based. As a result, at least among economically struggling families, the interests of children and the elderly are far from “in competition”— indeed, they are closely intertwined. The purpose of this study is to provide an assessment of poverty and program participation occurring among residents of households that include both older adults and children, age groups which draw on different pools of public support, but whose lives are often intertwined economically as well as socially. While a sizable literature documents the considerable kinship support systems that are in place linking the generations, the implications of these systems for poverty and for program participation have not been extensively considered. Given the trends and debates surrounding public support for these two broad age groups, such a consideration is long overdue. Background Family structure and employment status of parents are among the most consistently documented determinants of childhood poverty. For example, Eggebeen and Lichter (1991) suggest that 50% of the increase in child poverty between 1980 and 1988 can be accounted for by changes in family structure. Hogan and Lichter (1995) explain further that poverty is quite rare for children living in two-parent families, with both parents employed. The highest rates of poverty are reported for children in non-working mother-only families, followed by other nonworking parent configurations. In short, family status and work status of parents are both critically important—separately and in combination—in determining the chances that a child will experience poverty. These associations underlie some portion of the higher poverty rates observed among minority children. Black and Latino children experience poverty rates that are roughly triple that of white children (Haveman and Wolfe, 1994; Hogan and Lichter, 1995). Inasmuch as minority children are more likely than their white, non-Hispanic counterparts to live in single-parent households, and their parents are less likely to experience stable employment, some of the poverty differential is accounted for. However, even controlling for family status and employment of parents, minority children experience disproportionately high rates of poverty (Hogan and Lichter, 1995; Eggebeen and Lichter, 1991). In general, minority children also experience much less favorable health care, health outcomes, and educational outcomes (Angel and Angel, 1993; Haveman and Wolfe, 1994). Extended family networks and child well-being. One response to economic marginality is the instrumental use of extended family networks. Economic as well as in-kind assistance within extended family networks represent important supplements to some low-income family budgets (Edin and Lein, 1997). Extended family members may contribute directly to the economic support of a child's family, or they may offer in-kind help with clothing, food, or childcare. In addition, extended family members may contribute to a child's "social capital," through providing supervision and guidance for children (Coleman, 1990). The role of grandparents in particular has been examined in generating or supporting child well-being (Bass and Caro, 1996; Cherlin and Furstenberg, 1986). Indeed, some grandparents are seen as serving in “substitute” or “supplemental” parenting roles (Burton, 1992; Courtney, 1996; Stack, 1974). Intergenerational coresidence and child well-being. In addition to providing extra-household support, extended families may choose shared housing as vehicles for coping with economic distress (Edin and Lein, 1997). Coresidence among family members is especially common among African Americans and Latinos (Angel and Angel, 1993; Hernandez, 1993). This pattern appears to be related to greater economic need among these minority populations (Angel and Tienda, 1982; Burr and Mutchler, 1993; Mutchler and Burr, 1991; Tienda and Angel, 1982), although cultural factors may also play a role (Burr and Mutchler, 1993; 1998). Grandparents, especially grandmothers, are particularly likely to participate in coresidence arrangements. Recent data from the Current Population Survey suggest that almost 3.4 million children lived in their grandparents’ households in 1993; in one-third of these cases, no parent was present (Saluter, 1994). Although coresidence is not the only way in which a grandparent may contribute to the well-being of a grandchild, it does facilitate a variety of intergenerational exchanges that potentially contribute to child well-being. Of course, not all elderly family members are able to make positive contributions to child well-being. In many families, grandparents or other aging relatives may be poorly positioned to provide the kind of assistance that “substitute” or “supplemental” parenting may imply. Older individuals who relinquish headship and join the households of younger relatives are disproportionately likely to have few economic resources and be in poor health—conditions that would not appear to promote substantial economic or in-kind contributions on the part of the elderly individual (Mutchler and Burr, 1991). However, older household heads who include other relatives in their home appear to be less needy, and likely respond in large part to the needs of the extended members, who are often adult offspring and their children (Mutchler and Burr, 1991). Contributions of the elderly head to their young kin would be expected to be considerably more substantial in these cases. In one of the few assessments of this question based on SIPP data, Angel and Angel (1993), considering black and white grandmothers in multi-generational households, find that “entitlement” based income (Social Security) was higher among older grandmother household heads than among grandmother non-heads. In contrast, means-tested income, including AFDC and SSI, was substantially higher when grandmothers were not heads of household but lived with their children and grandchildren. Coresidence may facilitate child well-being through indirect means as well. Households including extended family members may facilitate the work participation of younger members, through providing childcare and assistance with housework activities. For example, Hao and Brinton (1997) observed that single mothers who lived with kin were significantly more likely than other single mothers to enter the labor market or return to school. Added adults in the household may also provide more reliable supervision of children (see McLanahan and Sandefur, 1994, with reference to grandmothers). Coresidence may also result in increases in available time for the adults in such a household, which may translate to improved access to and utilization of social service supports. Summary. This review suggests that older relatives, particularly grandparents, are important sources of support to children, especially among minority populations. Multigenerational households are one vehicle by which such support may be provided. Indeed, the resources of economically marginal families—those at greatest risk of poverty—may be quite limited across multiple generations. Yet, those offered by members of the older generations, especially Social Security, may be more reliable and possibly more extensive than the resources of their younger relatives, some of whom rely on sporadic and low-wage labor, means-tested transfers, and other intermittent sources. As a result, households containing an older person may provide some economic stability to the children who live in them. Inasmuch as children and the aged represent generational extremes that provide access to the largest potential number of “pools” for program participation and income support, such households may experience benefits in terms of poverty risk. The research questions guiding the current study are as follows: (1) Is poverty status among children associated with the presence of an elderly person (aged 65 or over) in the family household? (2) Do rates of program participation differ for children living in households including an elderly person, as compared to children living in households containing only younger persons? and (3) To what extent do households containing children receive income from various sources directed to the older population (e.g., Social Security) and children (e.g., AFDC)? These research goals reflect the need for basic information on the sources and amounts of economic transfers occurring within multiple-generation households that include both children and elderly, particularly public-based transfers. My expectation is that elderly individuals make positive contributions to the economic well-being of coresident children, through their economic contributions and through access to transfers and services. In particular, low-income children may be indirect beneficiaries—and perhaps indirectly dependent upon—transfer programs designed to benefit the elderly population. This study can help us assess the extent to which economically distressed children rely on their older coresident elders, and also for anticipating the ways in which challenges to the economic security of the older population may result in the further erosion of child well-being. Data and Methods In this study, the level of economic well-being and rates of program participation are examined for unmarried children under the age of 18 residing in family households. The child is the unit of analysis. Comparisons are drawn among children who live in households that contain older individuals (age 65 and over) and children who live only with household members younger than 65. Inasmuch as economic status of children is strongly determined by presence and number of parents, calculations are usually performed within parental status groups. Additional analyses summarize the sources of income (e.g., Social Security) and program participation (e.g., Food Stamps, subsidized housing) experienced by children with and without older coresidents. When sample size permits, comparisons are generated separately for African American, Latino, and non-Hispanic White children. The data for this project are drawn from the 1990 panel of the Survey of Income and Program Participation (see Shea, 1995, for a description of the data). The benefits of the SIPP data for this project include large sample size; nationally representative data; the ability to identify most family relationships among household members; the reporting of detailed income sources and amounts for all member of the household; and the reporting of program participation among household members. In this analysis, data from the first wave of the 1990 panel are used. Income and program participation assessments result from the reported sources, amounts and participation over the four months covered in the first wave, based on interviews occurring between February and May of 1990 (the reference periods range from October 1989 and April 1990). Variables. In order to conduct the analyses presented here, a number of variables were constructed from the SIPP data on a month-to-month basis. For each month, within each household, a number of household-based variables were constructed, establishing the presence of key household relationships. The economic contributions, by source, of elderly individuals in each household were also determined. The variables used in this paper reflect summarizations of family status, economic well-being, and program participation over the four months of wave 1. Family status and relationships: For each month, variables were constructed which determine the presence of each child's parent(s), the presence of each child's grandparents (insofar as this could be determined), and the number of people aged 65 and over in each child's household. Presence of grandparents could be unambiguously identified in all cases except for a small number of children living with neither parent, where the head of household was not the grandparent (less than 1% of the sample). The presence of a grandparent was established in either one of two cases: (1) when the grandparent was head of the household and the child was reported as grandchild of the head; and (2) when a parent was present, and the parent-child "person number" markers indicated that his/her parent (the child's grandparent) was also present in the household. Economic well-being: A number of variables measuring economic well-being of children are used in this paper. The central focus of the paper is poverty status, and so a considerable emphasis is placed on this indicator of well-being. Poverty status is based on a comparison of household income, summed over the four months covered in the first wave, with the poverty cutoff reported in the SIPP for that household. Many observers have argued that the official poverty line is far below what may reasonably reflect economic hardship, and advocate using a higher cutoff in summarizing economic well-being. Indeed, rates of hardship reflected by incomes under 125% of the poverty cutoff are routinely reported in government statistics (e.g., in Current Population Reports). Schwarz (1990) suggests that public opinion assessments would place economic hardship as occurring in households where incomes fell below 150% of the official poverty cutoff. Edin and Lein (1997) compare resources and expenditures among lowincome single mothers, and find that incomes approximating 150% of the poverty line are required to "make ends meet." In this paper, estimates are generated for the established poverty line, as well as 125% and 150% of the established poverty line. Sources of income. Although one purpose of this study is to assess economic well-being among children, regardless of source, an additional purpose is to consider the differences in sources of income based on household composition. The income sources considered here include the following: earnings income; property income; means-tested transfer income (largely Supplemental Security Income); and other income (including Social Security income, which is also assessed separately). Program participation. The major non-cash means-tested programs from which children may benefit include Food Stamps and housing assistance (including both housing subsidies and public housing). Inasmuch as eligibility requirements are extremely complex, and are variable across place, time, and recipient units (Williams, 1992), no effort is made in this paper to examine recipiency rates based explicitly on eligibility. Instead, patterns of recipiency are examined overall as well as for those children falling below the poverty cutoff, as defined above. The first four tables describe some key features of the population under study. Table 1 provides the unweighted counts and weighted percentages for children based on the presence of parents. In the first wave of the 1990 SIPP panel, 2% of the children are observed living with no parents in the household. An additional 23% live with one parent, and the remaining 75% live with two parents (the child may be the step-child of one of the parents). This is comparable to figures reported in the 1990 Current Population Survey (see column 3). Column 4 shows that coresidence with elderly individuals is not common for any of these children, but does occur among 19% of the children living with neither parent. Table 1: Distribution of Children by Parental Status, 199