The Implications of Grandparent Coresidence for Economic Hardship among Children in Mother-Only Families

Abstract

Estimates suggest that more than 6 million children live with at least one grandparent. Despite evidence establishing the growing prevalence of this arrangement, limited research has focused on estimating the implications of co-residence for the economic well-being of grandchildren. Using data from the 2001 panel of the Survey of Income and Program Participation, this paper examines levels of financial hardship among a particularly vulnerable group of children – those living in mother-only families. Findings suggest that children living in mother-only families that include a grandparent are substantially less likely to be living below or near the poverty level, compared to children living in mother-only families without a grandparent present. The financial security of children in these three-generation households is enhanced through significant economic contributions of the grandparents, and from household receipt of a wide range of financial resources, including means-tested cash transfers and other income such as Social Security.

Introduction

Multigenerational bonds have been identified as increasingly important in securing well-being, stability, and support for children (Bengtson 2001). Grandparents, in particular, are recognized as being the “first line of defense” when families are troubled, disrupted, or in need, often contributing time, caregiving, and financial resources, as necessary. The bonds between grandparents and grandchildren are noted for their diversity in content, encompassing a broad continuum of intergenerational care and support (Casper & Bianchi 2001; Kemp 2007). Caring “from a distance” is characteristic of many grandparents whose roles may be defined in terms of affection and attachment but little day-to-day interaction. In contrast to this version of the grandparenting role is the grandparent who lives in the same household as a grandchild, contributing extensively to the child's support and supervision, often sharing the parenting role to some extent (Goodman 2007). Those grandparents who share living quarters with a grandchild participate in the pooling of time and financial resources linking family household members; as such, coresiding grandparents may contribute to child well-being in part by ameliorating the economic hardship experienced by children (Deleire & Kalil 2002). These contributions may be especially substantial for children living in mother-only families, who are exposed to particularly high risks of poverty and hardship (Lerman 1996; Lichter & Crowley 2004; Thomas & Sawhill 2002).

The purpose of this study is to examine the implications of grandparent co-residence for the economic well-being of children living in mother-only families. Despite the upward trend in the prevalence of grandchild-grandparent coresidence, virtually no research has documented the economic outcomes associated with this arrangement. We evaluate the levels of poverty and near-poverty experienced by children in mother-only families, and find that living with a grandparent is associated with lower chances of experiencing hardship. As well, we examine the patterns of income packaging evident in children's households by comparing sources of income reported by household members, thereby learning more about the strategies by which financial security is provided. We use data from Wave 1 of the 2001 panel of the Survey of Income and Program Participation (SIPP), focusing on children under 18 who live with their mother but no father.

Background

Recent estimates suggest that about 6.2 million children lived with at least one grandparent in 2001, representing 8.5% of all children under age 18 living in households. Among these children, 2.5 million (or 40%) are part of a single-mother family (Kreider & Fields 2005) that includes the child's mother as well as one or more grandparents. The majority of three-generation households are grandparent-maintained, suggesting that these households may be formed primarily to help the younger generation (Bryson & Casper, 1999). Indeed, a substantial increase in grandparent-headed households has occurred in recent years, especially among children with a single parent (Casper & Bianchi 2001; Casper & Bryson 1998).

Living arrangements and economic hardship among children

The close tie between the living arrangements of children and the level of economic hardship experienced is well established in the literature. Although hardship can be usefully conceptualized in a number of ways—relating to subjective evaluations of difficulty in meeting bills, or noncash resources such as health insurance, for instance—much of the literature examines hardship with reference to income-based measures such as the poverty rate. The research evidence documents especially high poverty rates among children living with a single mother—more than a third of these children were poor in 2001 (37%) compared to about 18% of all children under the age of 18 (Kreider & Fields 2005). Indeed, more than half of all the poor children in the U.S. live with a single mother (Fields 2003). The contribution of coresident fathers to the economic well-being of their children has been extensively explored, and numerous studies have established that child poverty levels are especially high in families with an absent father (Bianchi 1999; Lerman 1996; Lichter & Crowley 2004; Lichter & Eggebeen 1994; Manning & Lichter 1996; Thomas & Sawhill 2002). In contrast, the potential contributions of family members in the household, other than the parents, to the economic well-being of children have been largely ignored in the literature (for an exception, see Manning & Smock 1997).

However, the literature on economic implications of household composition suggests that adding household members or merging households may be a strategy for achieving a more secure standard of living for families. The pooling of resources has long been identified as a benefit of household extension (Angel & Tienda 1982), and the remediation of financial difficulties has been highlighted as an important goal motivating the formation of multigenerational households. The question of “who helps whom” in the multigenerational household has been extensively explored, albeit primarily with respect to elderly parents and their middle-aged children (Casper & Bianchi 2001; Lee 1997; Pebley & Rudkin 1999; Speare & Avery 1993). While some research suggests that older family members often benefit from intergenerational coresidence—for example, needs for assistance on the part of an ill or frail older adult may be met in such a household—financial benefit is often experienced among those in the younger generations. Indeed, intergenerational coresidence commonly occurs due to short- or long-term needs on the part of younger family members associated with family disruption, unemployment, economic distress, or other stresses (Lee 1997; Speare & Avery 1993). Evidence consistent with this generalization is provided by Goodman and Silverstein (2002) in their study of Los Angeles County caregiving grandparents. They note that one of the most common reasons given by grandmothers in three-generation families for forming the intergenerational household is their desire to help the grandchild and his/her parents financially. Additional evidence suggests that grandparents contribute both directly (through the introduction of additional earnings/assets) and indirectly (e.g., through increased educational and employment opportunities of young mothers) to the financial well-being of coresident grandchildren (Gordon, Chase-Lansdale & Brooks-Gunn 2004). Previous research thus prompts our expectation that poverty rates among children living in single-mother families will be notably lower when a grandparent is also present in the household.

Other factors shaping risk of poverty among children

In addition to the influence of family composition, research on child poverty has emphasized the importance of resources possessed by key adults in the child's household. Given our focus on mother-only families, for our study the resources of the child's mother are most significant. Whether the mother has ever been married shapes aspects of the family's social and economic environment that are linked to the experience of poverty (Bianchi 1999) For example, marital history may shape the extent to which the child's father contributes economically to the family. Mother's human capital and other resources are key as well. Children with poorly educated mothers and with mothers who do not engage in paid work are more likely to experience poverty (Eggebeen & Lichter 1991; Lichter & Crowley 2004). Age of mother may also be a risk factor, inasmuch as children with young parents are more likely to be poor (see U.S. Census Bureau 2003).

An additional significant correlate of child poverty is race and ethnic background. Minority children experience far higher chances of experiencing poverty than do those in the White, non-Hispanic majority (Duncan & Rodgers 1988; Eggebeen & Lichter 1991; Lichter & Crowley 2004). Although these surplus risks are reduced when features of mother's resources and demographic characteristics are controlled, they are not eliminated. Without question, the same stratification processes that result in lower wages and other resources among minority parents shape the excess risks of experiencing poverty among their children. The size of this racial gap in poverty may vary depending on the living arrangements of children. For example, some research suggests that members of racial and ethnic minority groups benefit less from extended living arrangements than do non-Hispanic whites (Manning & Smock, 1997). Therefore, an additional goal of the current analysis is to examine the extent to which intergenerational living arrangements, which occur more commonly among non-White and Hispanic families, intersect with the race differences experienced by children living with single mothers.

Data and methods

Data for this paper are drawn from the 2001 panel of the Survey of Income and Program Participation (SIPP). The SIPP is a longitudinal survey of households, obtained from a multistage stratified sample of the noninstitutional resident population of the United States. These data are nationally representative of the household population and include a sample sufficiently large to permit reliable estimates surrounding this somewhat narrowly defined segment of the child population (i.e., children in single-mother families, including those living in three-generation households). Data were collected between February 2001 and May 2001 for Wave 1; responses on income and labor force activity refer to the four months prior to interview (see description of the data in http://www.census.gov/apsd/techdoc/sipp/sipp01w1.pdf). From these data, we structure our work file using children under age 18 as the unit of analysis.¹ We focus on children who live in a mother-only family; that is, families in which no natural, step- or adoptive father is present in the household. The child's mother may be cohabiting, but these children are only included in the study if the cohabiting partner is not identified as the child's father. We note that some research has suggested that poverty rates among children living with a mother and her cohabiting partner are similar to the rates reported for children with non-cohabiting mothers (Lichter & Crowley 2004; Manning & Lichter 1996), suggesting that the inclusion of these children in our analysis likely does not bias our results. Children who have been or are currently married are excluded from the study, as are children who are householders, foster children, or living in the home of a nonrelative.²

We note that coresidence decisions are closely intertwined with resource flows, shaping and perhaps responding to eligibility for financial programs, and resulting from decisions made by each of the several individual adults who may contemplate living together. A single mother who is able to live congenially with her own parents may choose to attend school rather than work full-time, reducing her personal financial resources but potentially providing a higher standard of living for her child within the multigenerational household, as well as investing in a more secure financial future for herself and her child. A similar mother whose parents are not living, who are disagreeable housemates, or who are financially insecure themselves may live in a separate household but experience a poorer standard of living and have fewer opportunities to invest in her future financial well-being. We are unable to determine what a child's resources would be if her mother had chosen a different residential pattern, because we cannot know the range of options available to each mother, nor how the residential choice she has made shapes other, linked choices. Instead, our approach in the current study is to compare the economic well-being of two significant segments of the child population—children in mother-only families who live with a grandparent and those who do not live with a grandparent—controlling for mothers' demographic characteristics. Readers are reminded that this approach cannot tell us what poverty would be among those children who do not live with a grandparent if they were to live in a three-generation household, but does provide useful information on how grandparent coresidence is associated with economic well-being for children living in single-mother families.

Variables

Family structure and grandparent coresidence

Variables are constructed by the authors reflecting the presence of each child's parent(s) and grandparent(s). Presence of parents is established based on person-number “pointers” on each child's record, linking the child to his or her father and mother within the household. Children are included in the study if a mother, but no father, is listed on the child's record. The presence of a grandparent in the household is identified primarily through a person-number “pointer” provided on the mother's SIPP record, identifying the mother's coresident parent(s). Additionally, grandparents are identified through the household roster; i.e. if the child is listed as the grandchild of the householder, the householder and the householder's spouse are identified as grandparents of the child. This second strategy of identification is necessary to locate paternal grandparents within the household; these individuals would not be identified on the mother's record as her parents. Unfortunately, due to the limited household relationship codes provided in the SIPP, paternal grandparents who were neither the householder nor householder's spouse are missed in this analysis.³ Our strategy for identifying children in single-mother families yields an unweighted sample size of 5,866 children, corresponding to a weighted estimate of 16 million children living with a single mother in 2001. This estimate is comparable to that reported in Fields (2003) based on the 2002 Current Population Survey. Through the examination of additional relationship pointers we establish whether a single grandparent is present in the household, or whether two grandparents are present. It is expected that children living with a mother and two grandparents will experience especially reduced likelihoods of poverty. All statistics presented in the tables are weighted, using a centered weight for each child.

Economic hardship

The most widely used indicators of economic hardship are those based on family income relative to household size; specifically, those based on the official government poverty cutoff. This cutoff is based on the expenditures determined to be necessary for a minimally acceptable standard of living (Orshansky 1965). However, the appropriate measurement of hardship is a topic of considerable debate (Edin & Lein 1997; Ruggles 1990), and many alternatives to the conventional measure of poverty have been proposed (Citro & Michael 1995). Our use of the government-defined poverty cutoffs is based on an effort to yield results that can be readily compared to and discussed in the context of existing research, much of which follows this convention. Some observers argue that the official poverty line is far below that which reasonably reflects economic hardship (Ruggles 1990) and advocate using a higher cutoff. Both in recognition of this concern, and in light of the potential for grandparents to assist in the avoidance not only of poverty but also of “near-poverty,” we analyze a three-category measure of hardship: (1) household income falls below the poverty line; (2) household income falls above the poverty line but below 150% of the poverty line; and (3) household income falls above 150% of the poverty line. This classification of “near-poor” has been cited as a common categorization within child and family research (Roosa, Deng, Nair, & Lockhart, 2005). Because this evaluation is based on the combined resources of the household relative to the size of the household, the addition of a household member could either increase or decrease the chances that a given household would be classified as poor. Adding a member with few financial resources would increase the chances of being poor if the increment to household income was low relative to the increased resource needs associated with addition of a member. Conversely, adding a member with more substantial resources could reduce the chances of experiencing poverty.

In the SIPP, economic resources of each person in the household are identified, including sources of income.⁴ An interview occurring in a given month includes reports of income by source for each person, for each of the four months prior to interview. Household income is measured by summing the income reported by all individuals in the household, from all sources. To smooth out month-to-month fluctuations in income, we assess household income for the entire four-month period covered by the interview rather than for a single month. Using this income, and the corresponding poverty cutoff for the household given its size and composition, we determine poverty status for the four-month period for each child.⁵

Other variables

Additional variables in the analysis include child's age (a continuous measure) and child's race and ethnic status (non-Hispanic White, Black, Other, and Hispanic/Latino). Mother's age is included as a categorical variable, because preliminary investigation suggested that mother's age is associated with poverty risk in a non-linear manner; the highest risks of experiencing poverty occur among single mothers who are in their mid-20s through late 30s. We include measures of mother's marital history in the form of dummy variables indicating whether the mother is widowed, divorced, or separated/living apart from her spouse, since mother's marital history may shape the financial resources to which she has access. Mother's paid work status reflects whether she reported working full-time (35 hours or more per week) for each of the four weeks prior to interview, whether she worked part-time (fewer than 35 hours per week, or for only part of the month), or not at all during that interval. Additional variables include mother's disability status (1=self-reported disability limiting the amount of work); and mother's educational status (did not complete high school, completed high school, completed college). The number of children in the household is also included in order to provide an additional control for financial burden in the household. For purposes of our analysis, these variables are considered primarily control variables, but reflect a common set of factors known to shape likelihood of experiencing poverty.

Results

A description of the sample used in the current analysis is presented in Table 1. Our calculations are based on 5,866 children in mother-only families from wave 1 of the 2001 panel of the SIPP who meet the inclusion criteria as described above. Nearly a third of these children are non-Hispanic Black children, and about 19% are Hispanic. Forty-five percent are non-Hispanic White children and just 4% are identified as another race. Children in these mother-only families have a median age of 9 years. More than half of the mothers of these children are between the ages of 25 and 39, and 57% of the children have mothers who are divorced, separated, or living apart from her spouse. Over half have mothers who work full-time; about one-quarter have mothers who are not working for pay during the reference period. The majority of children have a mother who graduated from high school but did not complete a bachelor's degree. The median number of children in the families is two. Just under 14% of the children living with a single mother also live with a grandparent (see Fields 2003 for a similar estimate based on the Current Population Survey). Nine percent live with a single grandparent (typically a grandmother), while almost 5% live with two grandparents.

Table 1. Descriptive statistics, children in mother-only families, 2001.

	Percentage	No Grandparent	One Grandparent	Two Grandparents
Characteristics of the child:
White, non-Hispanic	44.7%	45.6%	35.3%	44.9%
Black, non-Hispanic	32.6%	32.6%	35.9%	25.8%
Latino	18.7%	18.2%	22.0%	22.3%
Other, non-Hispanic	4.0%	3.6%	6.8%	7.0%
Age (median)	9.0	9.0	5.0	3.0
Age (mean)	8.8	9.3	6.3	4.7
Characteristics of the mother:
Under age 20	2.4%	1.0%	9.5%	14.6%
Age 20-24	10.3%	8.2%	19.5%	30.3%
Age 25-39	59.6%	60.7%	54.5%	49.1%
Age 40 and over	27.7%	30.1%	16.5%	6.0%
Divorced	36.1%	38.5%	23.9%	15.0%
Widowed	3.7%	4.0%	2.3%	0.3%
Separated or spouse absent	21.1%	22.3%	10.2%	20.2%
Never married	39.1%	35.2%	63.6%	64.5%
Working full-time	53.5%	55.3%	40.9%	45.8%
Working part-time	18.5%	18.1%	21.1%	20.5%
Not working for pay	28.0%	26.6%	38.0%	33.7%
Work disabled	11.3%	11.8%	9.9%	6.3%
Not work disabled	88.7%	88.2%	90.1%	93.7%
Less than high school	21.8%	21.7%	25.8%	17.1%
High school graduate	67.7%	67.2%	67.9%	75.2%
College graduate	10.5%	11.1%	6.3%	7.7%
Characteristics of the household:
Number of children in the household (median)	2.0	2.0	2.0	2.0
No grandparent of the child is present	86.1%
One grandparent of the child is present	9.0%
Two grandparents of the child are present	4.9%
N of cases (unweighted)	5,866	5,062	525	279

Note: Calculated by the authors from Wave 1 microdata file of the 2001 Survey of Income and Program Participation. The unit of analysis is children under age 18 in mother-only families. Excluded from the sample are children who have ever been married, and children who are non-relatives of householder or who head a household themselves. Except for number of children in the household, all statistics are significantly different across the family categories at p≤.01.

Table 1 also reports descriptive statistics for the sample stratified by the presence of one two, or zero grandparents. Except for the number of children in the household, these descriptive characteristics are all significantly different across the three groups of children, based on global tests of significance. For example, both mothers and children are substantially younger on average when a grandparent is present.

Table 2 includes statistics profiling the economic resources and rates of poverty experienced by children in single-mother families. Differences in hardship levels experienced by children with zero, one, or two grandparents present are also shown. For the sample as a whole, more than one-third of the children live in a poor household; fewer than half live in households with incomes more than 150% of the poverty cutoff. Chances of being poor are substantially lower among children living in a household including a grandparent, especially if two grandparents are present. Whereas 39% of the children living in a household that does not include a grandparent are poor, this rate is reduced by more than half to 14% among children living with two grandparents in the household. The figures on household income included at the bottom of this table highlight the low levels of income available to children living with single mothers—for the sample as a whole, a median household income of just $22,950 is reported. Far higher levels of household income are reported in households that include grandparents, however.

Table 2. Bivariate association between grandparent presence, poverty status and household income, children in mother-only families, 2001.

	No grandparent in the household	One grandparent in the household	Two grandparents in the household	Total sample
Poor (less than 100% of the poverty cutoff)	38.6%	25.8%	14.2%	36.2%
Near poor (100-149% of the poverty cutoff)	18.4%	22.4%	7.6%	18.3%
Not poor (150%+ of the poverty cutoff)	43.0%	51.8%	78.2%	45.5%
	100.0%	100.0%	100.0	100.0%
Household Income (in annual amounts):
Mean	$27,619	$36,697	$63,635	$30,202
Standard Deviation	$26,850	$25,445	$36,134	$28,409
Median	$21,129	$30,714	$59,729	$22,950
N of cases (unweighted)	5,062	525	279	5,866

Note: Calculated by the authors from Wave 1 microdata file of the 2001 Survey of Income and Program Participation. The unit of analysis is children under age 18 in mother-only families. Excluded from the sample are children who have ever been married, and children who are non-relatives of householder or who head a household themselves. Incomes are reported monthly; values in the table are based on the sum of reports across the four months prior to interview, multiplied by three for annual equivalents. All differences in values across columns are statistically significant at p≤.01.

Table 3 includes results from two multinomial logistic regression analyses. Both models reflect a child's likelihood of living in a poor or near-poor household relative to living in a household with income more than 150% of the poverty cutoff. The first model includes all the variables described above that refer to the child and to his or her mother, including the number of children in the household. These results suggest that Black, other non-White, and Hispanic children are substantially more likely than their non-Hispanic White counterparts to live in poor or near-poor households, as are children living in households including more siblings. The experience of poverty is less likely for children with older mothers, with divorced or widowed mothers, children whose mothers work outside the home, and children with more highly educated mothers. In Model 1, the association of mother's age with poverty is curvilinear; both children of younger and older mothers have lower likelihoods of poverty.

Table 3. Multivariate association between grandparent presence and poverty, children in mother-only families, 2001.

	Model I				Model II
	Poor v Not poor (1)		Near poor v Not poor (2)		Poor v Not poor (1)		Near poor v Not poor (2)
	b (s.e.)	Odds ratio	b (s.e.)	Odds ratio	b (s.e.)	Odds ratio	b (s.e.)	Odds ratio
Characteristics of the household:
One grandparent in the household					-1.658** (.142)	0.19	-0.478** (.132)	0.62
Two grandparents in the household (No grandparent)					-2.821** (.214)	0.06	-2.054** (.245)	0.13
Number of children in the household	0.384** (.032)	1.47	0.298** (.034)	1.35	0.427** (.033)	1.53	0.316** (.034)	1.37
Characteristics of the child:
Black, non-Hispanic	.916** (.090)	2.50	0.561** (.096)	1.75	0.841** (.094)	2.32	0.519** (.097)	1.68
Other non-Hispanic	0.351** (.189)	1.42	0.154 (.208)		0.515** (.198)	1.67	0.244 (.211)
Latino/Hispanic	0.416** (.105)	1.52	0.349** (.111)	1.42	0.453** (.109)	1.57	0.382** (.113)	1.47
(White, non-Hispanic)
Age	-0.017 (.009)		-0.019* (.009)	0.98	-0.031** (.009)	0.97	-0.026** (.010)	0.98
Characteristics of the mother:
Under age 20	-1.549** (.229)	0.21	-1.012** (.279)	0.36	-0.525* (.251)	0.59	-0.443 (.292)
Age 20-24	-0.523** (.132)	0.59	-0.314* (.147)	0.73	-0.214 (.140)		-0.120 (.151)
Age 40 and over	-0.238* (.094)	0.79	0.028 (.096)		-0.277** (.096)	0.76	-0.002 (.097)
(Age 25-39)
Divorced	-0.452** (.097)	0.64	-0.088 (.101)		-0.645** (.101)	0.53	-0.198 (.103)
Widowed	-1.204** (.213)	0.30	-0.472* (.210)	0.62	-1.456** (.220)	0.23	-0.597** (.212)	0.55
Separated or spouse absent	0.090 (.100)		-0.071 (.112)		-0.094 (.105)		-0.140 (.115)
(Never married)
Works full-time	-2.363** (.095)	0.09	-0.662** (.111)	0.52	-2.594** (.101)	0.08	-0.806** (.115)	0.45
Works part-time	-0.561** (.106)	0.57	0.202 (.129)		-0.661** (.112)	0.52	0.127 (.132)
(Mom does not work)
Has a work disability	0.495** (.121)	1.64	0.303* (.143)	1.35	0.452** (.125)	1.57	0.261 (.145)
High school graduate	-0.732** (.096)	0.48	-0.517** (.107)	0.60	-0.602** (.100)	0.55	-0.425** (.109)	0.65
College graduate	-2.027** (.170)	0.13	-1.686** (.178)	0.19	-1.996** (.176)	0.14	-1.630** (.180)	0.20
(Less than high school)
Intercept	.888**		-0.590**		1.310**		-0.339
Chi-square/ degrees of freedom	2310/32				2667/36
Pseudo-R2	.372				.418
N of cases	5,866				5,866

Note: Calculated by the authors from Wave 1 microdata file of the 2001 Survey of Income and Program Participation. The unit of analysis is children under age 18 in mother-only families. Excluded from the sample are children who have ever been married, and children who are non-relatives of householder or who head a household themselves.

^*p≤.05

^**p≤.01

Model II in Table 3 includes these attributes and also dummy variables indicating presence of one or two of the child's grandparents in the home. The inclusion of these variables yields a higher pseudo-R² and the associated coefficients suggest that compared to similar children who are not living with a grandparent, children living with one grandparent are far less likely to be poor (O.R.= .19) and also less likely to be near-poor (O.R. = .62). Children living with two grandparents are even less likely to be poor (O.R. = .06) or near-poor (O.R. = .13).⁶ The other coefficients in the analysis are affected only nominally by the inclusion of these variables. The estimated effects of child's race change only slightly, suggesting that estimated differences in poverty risk among children of different races are not diminished by taking into account grandparent presence.

In the literature review, it was noted that some research suggests that the implications of household structure for economic well-being may differ by race of the child (see Manning & Smock 1997). In order to evaluate that possibility for our study, interaction terms between having one or two grandparents in the household and the race of the child were tested for significance (results available from the authors). None of the interaction coefficients reached statistical significance for the contrast between being poor and being not poor. However, two interaction coefficients were statistically significant for near-poverty. These results suggest that the benefit of living with a single grandparent for reducing the likelihood of near-poverty is significantly higher for African American children, whereas the benefit of living with two grandparents is significantly lower for Latino children than it is for non-Hispanic White children.

The global implications of these results for the overall likelihood of experiencing poverty and near-poverty among children living with single mothers are illustrated by calculating predicted probabilities based on the coefficients from the interaction model. We compare predicted probabilities calculated for children living with a one grandparent or two grandparents to those calculated for children who do not have a grandparent in the household. We perform these comparisons for children within each of the four racial and ethnic group categories defined in the study. These predicted probabilities are presented graphically in Figure 1.

Figure 1. Predicted probabilities of being poor or near-poor, by presence of grandparents and child's race.

Note: Predicted probabilities calculated based on coefficients in Table 3, Model II. Calculations assume child aged 9, mother aged 25-39, never married, working full-time, not disabled, with a high school degree, and two children in the household.

The calculations used to generate these predicted probabilities hold most variables in the model at a constant level reflective of the average or “typical” case in the sample.⁷ All calculations assume a constant child's age (9 years), mother's age (age 25-39), mother's marital status (never married), work status (working full-time), disability status (not disabled), educational achievements (high school degree), and number of children in the household (2). The child's race and number of grandparents vary across the calculations. These predicted probabilities suggest that regardless of race, children living with single mothers in households that also include one grandparent experience lower expected levels of poverty, ranging from 3.5% for non-Hispanic White children to 8.7% for non-Hispanic Black children. Even lower chances of experiencing poverty are expected among children living with two grandparents—3% or less for all racial and ethnic groups. Substantially higher poverty levels are expected for same-race children who do not live with a grandparent, ranging from 17.4% for white non-Hispanic children to more than one-quarter of non-Hispanic Black children. Among Black children, more than half are expected to be poor or near-poor in households that do not include their grandparent, compared to less than one-quarter of similar children living with a grandparent. Similar patterns are reflected for children in the other ethnic groups, all suggesting a reduction in risk of economic hardship of more than one-half associated with living with a grandparent.

Income packaging in mother-only families

As noted earlier, the social processes shaping decisions surrounding household composition (with whom shall I live?) are closely entwined with processes shaping income generation (how shall I obtain financial security?). For example, in trying to provide an adequate standard of living for her children, a single mother may perceive a choice between working substantially longer hours, potentially accruing higher child care costs in the process, or merging households with her mother in order to make ends meet. Fully disentangling these processes is a difficult if not impossible task, and beyond the scope of the current paper. However, we present some descriptive information intended to suggest some ways in which the financial resources of single mother families including a grandparent differ from the resources of their counterparts who are absent a grandparent.

Table 4 indicates that the financial contributions of grandparents to the three-generation household are substantial (see Panel A). Most single mothers report some personal income, although those living in three-generation households are less likely to do so than those living without a grandparent in the household (87% compared to 96%). The mean amount reported by mothers is substantially less in three-generation households than in households without a grandparent, by about one-third. In addition to suggesting that mothers living in a three-generation family are less likely to work for pay, this may suggest that that they also work fewer hours, or that they earn a lower wage. Almost all coresident grandparents report personal income, and in three-generation households the income of the grandparent(s) represents the largest single source (45-66% of the total household income). All three types of families frequently report income from someone else living in the household—a sibling, another relative, or someone else—but the values reported do not differ much by type of household. All told, the mean household income for children living with a single mother plus a single grandparent is nearly $37,000 on an annualized basis—one-third higher than the mean reported for children in single mother families without a grandparent (about $28,000). For children living with two grandparents, the difference is even more substantial. These children live in households with mean incomes more than twice as high as those including single mothers only.

Table 4. Income packaging by grandparent presence, children in mother-only families, 2001.

Panel A: by type of recipient
	With no grandparents in the household		With one grandparent in the household		With two grandparents in the household
Recipient of income	Mean	% having income from this individual	Mean	% having income from this individual	Mean	% having income from this individual
Mother	$20,958	96%	$13,674	87%	$14,217	87%
Grandparent	n/a	n/a	$16,566	90%	$42,092	99%
Someone else	$6,660	36%	$6,457	38%	$7,327	43%
Household income	$27,619		$36,697		$63,635
Panel B: by source of income
	With no grandparents in the household		With one grandparent in the household		With two grandparents in the household
Source of income	Mean	% having income from this source	Mean	% having income from this source	Mean	% having income from this source
Earnings	$22,411	84%	$27,344	86%	$53,237	96%
Property	$264	31%	$589	41%	$690	54%
Means-tested cash transfers	$1,164	21%	$2,235	32%	$1,407	20%
Other	$3,780	51%	$6,529	65%	$8,301	67%
Household income	$27,619		$36,697		$63,635

Note: Calculated by the authors from Wave 1 microdata file of the 2001 Survey of Income and Program Participation. The unit of analysis is children under age 18 in mother-only families. Excluded from the sample are children who have ever been married, and children who are non-relatives of householder or who head a household themselves. Incomes are reported monthly; values in the table are based on the sum of reports across the four months prior to interview, multiplied by three for annual equivalents

Another aspect of the household income stream of interest here is the type of income reported within the household. Panel B of Table 4 reports four types of income—earnings, property income (e.g., interest or dividends), means-tested cash transfers (e.g., SSI, TANF) and other types of income (e.g., Social Security). For both types of household, income from earnings is most substantial as well as most common. Households including a grandparent are more likely to have each type of income reported by someone in the household; as well, the mean amount of income reported is higher for each source of income. The only exception to this pattern is means-tested income, which is reported by only 20% of the two-grandparent households. This suggests that children living in households including a grandparent benefit from a wider range of income sources—earned and unearned, and at least for one-grandparent households, means-tested transfers as well as other types of transfers. Especially significant is the substantially higher value of earnings and of “other” income, such as Social Security.

Discussion

Grandparents are frequently a resource for younger family members, often stepping in during family crises. For many children, grandparents form an integral part of their household, contributing time as well as financial resources. This research investigates the role of multigenerational co-residence in enhancing the economic well-being of children in mother-only families. We focus on children living with a single mother because these are the children most at risk of experiencing economic hardship. Based on data from the 2001 Survey of Income and Program Participation, our results show that children living in single-mother households with a grandparent present are significantly and substantially less likely to be living in poverty than are children in single-mother households without a grandparent. We find that in these households, the income of grandparent(s) represents the largest single source of income. Moreover, households including a grandparent are more likely to report receiving most types of income, suggesting that children in such households draw on a wider range of sources of support.

Our multivariate results suggest that children in single mother families who also live with a single grandparent are only 19% as likely to be poor as are their counterparts without a grandparent present, and 62% as likely to be near-poor (that is, to live in households with incomes 100-149% of the poverty cutoff). Children living with their mother plus two grandparents experience an even greater advantage in terms of avoiding poverty. Although these results suggest that, on average, children who live with a grandparent experience substantially less economic hardship than those who do not, we cannot assume that children with single mothers would be better off financially if they also lived with a grandparent. It is possible that the single mothers who live with a grandparent are those who have the most to gain from doing so. For many single mothers, living with their parents is not an option; as well, within many of these families the generations share marginal economic circumstances that would not be improved through coresidence. Our results do suggest that for a segment of the population of children with single mothers, avoidance of poverty and intergenerational coresidence go hand-in-hand, with grandparents making substantial contributions to the economic well-being of the grandchildren with whom they live.

To further expand our understanding of the economic contributions of grandparents to the children with whom they live, future research should focus on the formation of multigenerational households, in order to determine the extent to which these households are part of a strategy to combat poverty and avoid its consequences among single mothers and their children. The research presented here suggests a positive association between grandparent coresidence and reduced chances of experiencing poverty, and past research has suggested that other favorable outcomes for children may also occur, such as improved chances of completing high school (Deleire & Kalil 2002). Yet to be determined is whether living in multigenerational settings represents a long-term solution to economic disadvantage for some families. It may be that the pressures toward residential independence are so great that these arrangements are used as short-term solutions to only the most significant short-falls. Further examination of the specific strategies used by multigenerational families in securing well-being for children would also be helpful. Are these families better able to access services and sources of public support for which they may be eligible, such as food stamps or free lunches at school (Brandon 2005)? And, if so, how might this influence levels of material hardship (e.g. food security) within these households? Learning more about the characteristics of coresident grandparents that are especially conducive to improving economic security for children will also be helpful. Although our study suggests that on average the presence of grandparents is positively associated with well-being among children in mother-only families, it may be that these benefits occur primarily when the grandparent is employed, married, or especially secure financially. Ultimately, the relationship among these factors will be best studied longitudinally, to elaborate on the causal processes that may be at work. It will also be necessary to replicate these analyses with other cohorts, to ensure that these results are not unique to children living with single mothers during the time period examined here. An examination of these issues will go far to elucidate the role of multigenerational co-residence in combating poverty within mother-only households.