Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Mar 1.
Published in final edited form as: Popul Dev Rev. 2015 Mar 1;41(1):127–146. doi: 10.1111/j.1728-4457.2015.00029.x

Competing Demands from Aging Parents and Adult Children in Two Cohorts of American Women

Emily E Wiemers 1, Suzanne M Bianchi 2
PMCID: PMC4649941  NIHMSID: NIHMS725296  PMID: 26594071

Abstract

In late middle age, individuals may face competing demands on their time and financial resources from elderly parents and young adult children. This study uses the Panel Study of Income Dynamics to examine changes over time in the probability of having children and living parents for women age 45 to 64. We compare two cohorts: those born in the 1920s and 1930s and those born in the 1940s and 1950s. We find that there has been a dramatic increase in the probability of having children and living parents and that this increase has been driven mainly by changes in life expectancy of the parent generation. We further examine transfers of money and co-residence for women in the later cohort. We find that while women may not give to parents and children concurrently, approximately thirty percent of them have provided support to both parents and children at some point in the past.

Introduction

In the decades leading up to retirement, individuals may have obligations to multiple family members including adult children and grandchildren as well as aging parents and parents-in-law. In the aging literature, the majority of research has focused on care for parents (Coward and Dwyer 1990, Dwyer and Coward 1991, McGarry 1998, 2006, Wolf, Soldo, and Freedman 1996, Wolf, Soldo, and Freedman 1997) but care and support for adult children is actually more prevalent in late middle age (Kahn et al. 2011). Demands for care and support are unlikely to come from only one family member and individuals in late middle age may often be sandwiched between the needs of their children and grandchildren and the needs of their parents. Fertility, mortality and marriage trends have reshaped these intergenerational ties over time. Trends toward later births, fewer births, and a delayed transition to adulthood alter the time required to fully launch one’s offspring. Longer life expectancy increases the chances that one will have surviving parents later in the life course. Marriage, which brings ties to parents-in-law, is increasingly delayed or foregone altogether by some segments of the population.

This study uses the Panel Study of Income Dynamics (PSID) to analyze the relationship between demographic trends in mortality, fertility, and marriage to changes in the size and attributes of the so-called “sandwich generation” between 1988 and 2007. There has been some assessment of how these demographic trends may have affected the “sandwich generation” at a single point in time but there has been no assessment of change over time. To assess change over time, we compare two cohorts of women: those born between 1924 and 1943 and the Baby Boom cohort born between 1943 and 1962. We begin by examining the differences in the potential for supporting to multiple generations between cohorts by assessing how many women have living parents and children.1 We then use the more recent cohort to examine current and past support that individuals have provided to parents and children.

Our analysis examines both the potential for obligations to multiple generations and the actual transfers women give to parents and children in terms of transfers of money, time, and co-residence. We do so because intergenerational ties, demands, obligations, needs, and even transfers are complex and this complexity makes it hard to define who is a member of the “sandwich generation.” Actual transfers may only represent a small fraction of what women anticipate giving to parents and children and anticipated transfers may affect behavior and cause emotional stress and worry even if actual transfers are never observed. Similarly, actual transfers may not always be given in response to a demand; parents may transfer money to young adult children because they think it is helpful, not because their children demand or even need money. Conversely, women with few financial resources may not be able to make any transfers of money even though they have needy children and parents. While the term “sandwich generation” evokes the costs of providing for multiple generations, ties with multiple generations can also be beneficial. Older parents may help by caring for younger children or grandchildren (Bryson and Casper 1999, Pebley and Rudkin 1999, Wang and Marcotte 2007), by financing college attendance or other costs associated with the transition to adulthood (Arrondal and Mason 2001, Kahn et al. 2011), and by providing emotional support, acting as role models, or increasing family solidarity for younger generations (Bengtson 2001). Having both older parents and young adult children may also benefit the middle generation explicitly by allowing them to “demonstrate” how to care for an elderly parent to their young adult children (Cox and Stark 2005). Like much of the literature on the “sandwich generation” we frame our analysis in terms of caring for multiple generations but also discuss the problems with measuring demands and needs using actual transfers, the importance of anticipated transfers, and the benefits of having parents and children in late middle age.

The Changing Demography of Kin Availability and Need

Mortality

Between 1970 and 2007, life expectancy at birth increased from 71 to 78 years (Miniño et al. 2011). By 2008, women lived an average of 80.6 years and men 75.6 years (Miniño et al. 2011). Uhlenberg (1996) estimates that about 37 percent of 60 year olds had at least one living parent in 1980 compared with 44 percent of 60 year olds in 2000. Watkins et al. (1987) show that the number of years women spend with living parents has also increased over time. Other things equal, as parents live longer, individuals in late middle age have an increased likelihood of having at least one parent who survives into old age when frailty and disability increase. This prediction is complicated, however, by the fact that increases in life expectancy have been coupled with declines in disability later in life (Christensen et al. 2009, Crimmins 2004, Cutler 2001, Martin et al. 2010). The trend in declines in functional limitations among the elderly in the 1990s may have stalled more recently but does not seem to have reversed (Freedman et al. 2013). This increase in “healthy life expectancy” may lessen care demands from elderly parents even as parents live longer.

Fertility

The effect of fertility trends on later life caregiving is even more difficult to predict than the effect of mortality trends because some changes increase the likelihood of having “needy” children later in life whereas other changes lessen this probability. Between 1970 and 2009, the TFR declined from 2.5 to about 2.0 children per woman (Martin et al. 2012: Table 4). The decline in TFR is due to both the reduction in the number of children for women who have children and the increase in childlessness. The percent of women 40–44 without children has increased from 11 percent in 1984 to 18 percent in 2008 (Pew Center Report 2010). Other things equal, the decline in the number of children should have decreased the demands on parents for care and support of children. However, the timing of childbearing has also undergone change, with the median age at first birth rising from 22.7 to 25.2 between 1980 and 2009 (Martin et al. 2012: Table I-1). This suggests that women today may reach late middle age with younger children than in the past, increasing the likelihood that they still have dependent children – either children under age 18 in the home or older children who are not yet fully financially independent. Change due to the later timing of fertility is muted, however, by the fact that earlier cohorts had more children than later cohorts. Thus, the change in age at last birth may not be as dramatic as the change in age at first birth across cohorts. It is likely that the timing of last birth, more so than timing of the first birth, more often determines when the nest is truly empty.

Table 4.

Mean Characteristics of Women with Children and Living Parents, Age 45–64

PSID
1988 2007
Married 78.9 79.7
Age 52.5 52.7
Number of Own Parents 0.97 1.1
Number of Own Parents and In-Laws 1.6 1.9
Mean Age Own Parent 78.5 79.7
Has a Sibling 0.95 0.97
Has a Sister 0.76 0.76
Number of Children 3.3 2.4
Mean Age of Youngest Child 23.5 23.9
Mean Age of Oldest Child 30.7 28.7
N 642 858
Open in a new tab

Notes: Weighted using individual weights. N for age of parents is lower (N=503 in 1988 and N=678 in 2007) because of missing data on parental age. Note that the probability of having a sibling for the overall sample of women 45–64 is lower in 1988 than in 2007 (93% vs 96%) consistent with the increase in the probability of having children during the Baby Boom. However, the difference in the probability of having a sibling conditional on having a living parent and a living child is smaller between the two cohorts.

Young adults today remain in their parents’ households longer before striking out on their own than in the recent past (Furstenberg et al. 2005). Higher rates of college attendance have extended the period of financial dependence on parents in more affluent families (Schoeni and Ross 2005) and young adults who do not go to college have great difficulty finding good jobs. The extent to which young adults take longer to settle into stable careers and family lives – and parents help finance their slow transition – have likely increased the need for support from parents later in the life course (Kahn et al. 2013).

Marriage

Marriage – as a formal link that ties families together - creates obligations to parents-in-law and increases the sets of older kin who may need assistance. Pierret (2006) shows that married women aged 43 to 54 in the late 1990s, had more living parents (or parents-in-law) than unmarried women. Whereas only 11 percent of married women had no parents, almost one-third of unmarried women had no living parents. Differences in caregiving between married and unmarried women may be dampened because unmarried daughters more often provide care than married daughters (McGarry 1998). Married daughters greater number of kin who might need care may be counterbalanced by unmarried daughter’s greater propensity to provide care when a need arises.

Studies of Sandwich Care

Studies of sandwich caregiving are limited but one of the best descriptions is Pierret’s (2006) analysis of the National Longitudinal Study – Young Women cohort. He estimates both the percentage of women who are at risk of being sandwiched between the needs of children and parents and, using various definitions, the percentage giving help to two generations simultaneously. Using a definition of “sandwich caregiving” in which assistance to parents includes either co-residence, having a parent in a support facility, giving aid of $200 or more in the previous year, or providing 100 hours or more of assistance, and, assistance to children includes co-residence, support for college, aid of $200 or more, or assistance of 100 or more hours, 33 percent of women in their mid-forties to mid-fifties could be considered sandwiched caregivers in the late 1990s in the U.S.. If higher levels of assistance are used to define sandwich caregiving, e.g., parental co-residence or providing aid of $1,000 or more, or help amounting to 500 hours or more combined with support for children that includes either co-residence or support for college, or aid of $1,000 or more, or assistance of 500 hours or more, than a smaller nine percent of women are classified as sandwiched caregivers.

A recent Pew Research Center Report on the sandwich generation (Pew Research Center, 2013a) takes a somewhat different approach. Instead of defining the population at risk for facing competing demands and then examining actual transfers, they define the “sandwich generation” as adults who have a parent age 65 or older who are also either raising a young child or financially supporting an adult child. Under this definition, 47 percent of adults in their 40s and 50s are members of the sandwich generation. Consistent with other research on transfers, they find that financial transfers are more likely to flow to children (McGarry and Schoeni, 1995). They also find that the sandwich generation is more likely to provide emotional support to a child than to a parent.

Henretta, Grundy and Harris (2001) use the 1994 Health and Retirement Study (HRS) to estimate the percentage at risk of sandwich care for those aged 50 and over. Between 32 and 37 percent have both living children and at least one living parent, with higher estimates for more highly educated women than for women with less education. Grundy and Henretta (2006) combine financial and time assistance to estimate the percentage of women age 55 to 69 who provide care to both generations. About 36 percent of married women and 27 percent of unmarried women are helping both parents and children simultaneously.

A handful of other studies focus on whether there are negative outcomes, in terms of health, for those they define as “sandwiched caregivers.” These studies are not always careful about determining who is at risk of being sandwiched between multiple generations and vary in their comparison groups. Chassin et al. (2010) find that sandwiched caregivers engage in less healthy behaviors than others (e.g., more smoking, less seat belt usage, less exercise, less health conscious food shopping), suggesting that caregiving may reduce time for activities that enhance well-being or that the stress of caregiving puts one at risk of engaging in unhealthy behaviors (e.g., smoking). Other studies do not find detrimental effects of sandwich caregiving on well-being (Künemund (2006) for a German sample; Loomis and Booth (1995) for a U.S. sample, with Williams (2004) reporting mixed results for a Canadian sample).

To summarize, the existing literature on sandwich caregiving is not extensive and is largely cross-sectional. Definitions of what constitutes sandwich caregiving vary across studies, making it difficult to compare findings. Some of the best studies have used nationally representative data to define sandwich care and estimate its prevalence among women later in life, but only for one point in time or for one cohort. We build in particular on the work of Pierret (2005) and Henretta, Grundy, and Harris (2001) by first examining changes over time in the population of women with children and living parents and then exploring actual transfers that women make to children and parents. This paper extends these two studies by looking at the population at risk across cohorts to examine changes in the potential for demands from both children and elderly parents over time. We also contribute to the literature by showing that the time period over which we explore transfers has a large impact on the probability of women giving transfers to multiple generations in late middle age.

We ask three main questions: 1) Who is potentially at risk of being sandwiched between needs of children and parents and has this changed over time? 2) What is the contribution of changing fertility, mortality and marriage patterns to the change in the population with children and living parents? 3) Focusing on those with both living children and parents, who gives to both generations simultaneously? We address these questions using the Panel Study of Income Dynamics (PSID), the longitudinal data set that allows estimation of change in likelihood of being sandwiched over time and that provides indicators of support and care for those at risk.

Data and Measures

The PSID is the premier dataset in the U.S. for studying intergenerational ties because of its genealogical design, its long life histories of linked family members, and its high wave-to-wave response rates. Begun in 1968, the study follows individuals whether or not they are living in the same dwelling as the original sample household or with the same people. All individuals in households recruited into the PSID in 1968 are said to have the PSID “gene.” Individuals who are born to or adopted by someone with the PSID gene acquire the gene themselves and are followed and become members of the PSID sample for the rest of their lives. This design feature implies that the study provides, at each wave, data on a sample of extended families. Interviews were conducted annually until 1997 when PSID moved to an every other year schedule.

This paper uses data on two cohorts of women. The first cohort is age 45–64 in 1988 and was born between 1924 and 1943. The second cohort is age 45–64 in 2007 and was born between 1943 and 1962. Analysis samples include a total of 1,207 women in 1988 and 1,369 women in 2007.2 We focus on women between the ages of 45 to 64 because these are the ages in which women are most likely to have young adult children and parents who may require care and it is when retirement savings typically reach their peak (Attanasio 1998). To facilitate comparisons across cohorts we use only individuals with the PSID “gene” in both samples, we exclude individuals in the immigrant sample3, and we weight using the longitudinal weights that are comparable across years. In both 1988 and in 2007, information was collected on living parents and parents-in-law, making it possible to identify women in late middle age who have children and living parents.

Defining the Potentially Sandwiched

The first thing we do in this paper is to determine the percentage of women who are potentially sandwiched between children and parents in order to assess whether this has changed over time. Women have the potential to be sandwiched if they have at least one child (of any age) and at least one living parent. Our analysis requires information on the number and age of children and whether parents and parents-in-law are living. The children of the women in our sample are PSID sample members and are followed over time. Using both information derived from PSID interviews and information from birth histories, we have consistent information about the number and year of birth of children for both cohorts in our sample.4

We measure having a living parent in two ways—having a living parent and having a living parent or parent-in-law. We include parents-in-law because obligations to older generations may come in the form of caring for an in-law. Consistent information on the parents and parents-in-law of the women in our sample is more problematic than consistent information on children. For the first cohort, most parents are not PSID sample members and are not interviewed. Many individuals in the second cohort have at least one PSID sample parent. However, two special data collection efforts allow us to have consistent information about parents and parents-in-law. In 1988 a special supplement on time and money transfers was added to the data collection. This supplement obtained several characteristics of parents and parents-in-law including whether they were currently living in 1988, their age if living and year of birth and death if deceased, and their marital status. In 2007, household heads and spouses were asked whether their parents were living along with set of transfer questions. These special supplements allow us to have consistent information about the number of parents and parents-in-law who are alive at the time of the interview for the women in both cohorts of our sample.

While knowing whether an individual has a child and a living parent may seem trivial, this information is included in very few nationally representative surveys.5 The National Survey of Families and Households (NSFH) and the Health and Retirement Study (HRS) are two other surveys in which individuals in this age group are asked about both parents and children. The NSFH would allow for a similar comparison over time but the latest wave is 2003. The HRS would allow for examining changes in potentially sandwiched women over time but the cohort structure of the data does not allow for a consistent examination of individuals under 50. Including these younger women seems particularly important for capturing individuals with lower socio-economic status, who, because of differential mortality and fertility patterns, are more likely to experience competing demands from parents and children at younger ages. In fact, the Pew Center (2013a) finds that 71 percent of individuals with a parent 65 or older and either a minor child or an adult child to whom they are providing support are between 40–59 with a further 19 percent younger than age 40. Only 10 percent of these individuals are 60 and over.

Defining Who Gives Transfers in 2007

After describing change over time in the potential for being sandwiched and examining changes in the characteristics of these individuals over time, we focus on those with children and living parents in 2007 and examine the likelihood that they make transfers up and down the generational ladder.6 In particular, we are interested in how different definitions of the timing of transfers affects the estimates of care to multiple generations. Our contribution is to consider not just simultaneous caregiving to multiple generations but to consider providing support to multiple generations over a longer time period.

Co-residence

One of the ways in which individuals provide support for others is to co-reside and hence we assess whether those with living parents and children have children, parents or both generations co-residing in the household. Current co-residence with parents and children is assessed using household rosters. Past co-residence for parents is measured using a question in the 2007 supplement in which PSID respondents were asked “In the past, did your [or your wife’s] parents live with you [and your wife] one year or longer in your home?” Past co-residence with children is measured using household rosters for the past ten years (1997–2007). For each wave, we measure whether any children, any children 18 and over, and any children 25 and older are co-residing with the respondent. We use only the last ten years because at some point, virtually all of our sample co-resided with a child. We break up the child co-residence by age of the child to examine differences between providing support to a minor child and providing support to an adult child.

Financial Assistance

In each survey year, the PSID includes a question on whether financial support has been given to anyone outside of the household during the past year. Respondents are asked “In 2006, did you or anyone else in your family living there give any money toward the support of anyone who was not living there at the time, including child support, alimony, money given to parents, and things like that? Don’t include loans or charitable contributions to organizations; we'll ask about them later.” If respondents answer yes, the recipient’s relationship to the household head is identified.7 We use this question to identify whether the respondent has given money to children or to parents in the past year. We also use information from this annual question on financial giving from past waves of the PSID to identify whether respondent had given money to children in the past ten years. We note that the measure of transfers of money was not asked separately for parents and children. Analysis of the 1988 PSID Time and Money Transfer Data suggests that this leads to lower reports of transfers (Altonji et al. 2000, McGarry and Schoeni 1995). Because we believe that financial transfers may be underreported and they exclude money given to individuals living in the household, our estimates on financial assistance should be considered a lower bound estimate.8 We supplement information on past financial transfers to children with information on past financial transfers to parents. Instead of using the annual reports, we use responses to the question, asked in 2007, “In the past, did you [or your wife] give significant financial support to your [or your wife’s] parents?” We use this information to expand the examination of financial transfers up the generational ladder.

Care

We do not have a measure of time transfers to parents or children in the 2007 PSID but there was a question about whether parents had received care in the past. Respondents were asked “In the past, did you [or your wife] spend a lot of time caring for your [or your wife’s] parents or parents?” We show estimates of this type of transfer and also provide overall estimates of support to multiple generations that include time transfers.

Results: “Potentially” Sandwiched

We begin our analysis by examining the potential for being sandwiched between a younger and an older generation at two points in time. Table 1 shows the percent of women with living parents, children, and both. The percent of women who have both living parents and children—the potentially sandwiched—increased by 11 percentage points from 43 to 54 percent between 1988 and 2007. If we include parents-in-law, 55 percent of women have a living parent and child in 1988, increasing to 64 percent in 2007.9 These increases are despite the fact that the percent of women age 45–64 with children decreased slightly from 92 percent in 1988 to 88 percent 2007.

Table 1.

Percentage of Women, Age 45–64, with Surviving Parents and Children

PSID
1988 2007
PARENTS
At Least One Surviving Parent 47.4 61.8
At Least One Surviving Parent or Parent-in-Law 59.5 72.9
CHILDREN*
At Least One Child 91.4 87.3
“SANDWICHED”
At Least One Surviving Parent Plus Child 42.6 54.0
At Least One Surviving Parent or Parent-in-Law Plus Child 54.3 64.1
Sample Size 1,207 1,369
Open in a new tab

Notes:

*

At least one live birth . Weighted using individual weights. The percent of individuals with living parents and children in the 1988 PSID matches that in the 1994 HRS (Henretta, Grundy, and Harris 2001) and the 1988 NSFH very closely (Reading down the column the NSFH numbers for 1998 are 70.1, 43.4, 55.8, 90.2, 39.8, 51.4). The percent of individuals with living parents in the 2007 PSID matches that in the 2011 SCA very closely.

Table 1 shows that the potential for women being sandwiched between the needs of aging parents and adult children in the decade or two before retirement has increased quite dramatically—by about 20 percent—between the two cohorts in our study. By 2007, almost two thirds of women in later middle age are “at risk” of simultaneously facing needs of their children and of their elderly parents.

The changes in the probability of having living parents and children over time are necessarily driven by three demographic factors: the number of parents and in-laws a woman has, the percent of a women’s parents and in-laws who still are living, and whether a woman has children. Having a living parent or in-law is in turn influenced by life expectancy of parents, the age difference between middle-aged adults and their parents, and the age difference between spouses. Because childbearing is largely complete by age 45, the number of children and fertility timing are not related to whether a woman is potentially sandwiched between the ages of 45 and 64—other than through childlessness—though family size is related to the scope of demands that are likely from the younger generation.

Table 2 presents information on the components that contribute to the likelihood of being potentially sandwiched; the relative importance of marriage, mortality, fertility, and age differences between parents and children and husbands and wives as factors underlying the change in the size of the population at risk of facing competing demands over time. Table 2 focuses on all women in a cohort and shows the percent of women 45–64 that are married, the average total number of own parents, the average total number of parents and in-laws, the average age of mothers when the 45–64 year old women were born, the average age difference between husbands and wives, and the percent of women that are childless.10 Increases in the average number of own parents over time could be the result of increases in parental life expectancy or of declines in the age of parents when the 45–64 year old women were born. Similarly, increases in the average total number of parents plus in-laws over time could be the result of increases in parental life expectancy, declines in the age of parents when the 45–64 year old women were born, and increases in marriage or in the age homogamy of marriages. Increasing childlessness decreases the percent of women with living parents and children and declines in marriage decrease the percent of women with both parents and in-laws.

Table 2.

Mean Characteristics of Women, Age 45–64

PSID
1988 2007
Married 68.7 62.7
Number of Own Parents 0.58 0.84
Number of Own Parents and In-Laws 0.95 1.36
Age of Mother at Birth 27.7 27.8
Difference in Spouse and Own Age 2.6 1.9
Childless 8.59 12.45
N 1,207 1,369
Open in a new tab

Notes: Weighted using individual weights. Marriage rates for 1988 match the marriage rates (measured as the percent of non-Hispanic white and black women 45–64) for the 1990 Census closely (68.1%) and the 1988 CPS (70.6%). Marriage rates for 2007 match the 2007 ACS (61.4%) and 2007 CPS (65.5%) closely. Although marriages rates are slightly higher in the CPS than in the PSID or the ACS, all show a decline of between 5 and 6 percentage points. The differences in age of mother at birth are consistent with Kirmeyer and Hamilton’s (2011) analysis of fertility in the 1910 and 1935 birth cohorts. Kirmeyer and Hamilton’s (2011) analysis is based partially on Heuser’s (1976) fertility data. Comparable figures on childlessness are difficult to obtain. In the 1990 Census 11.7% of non-Hispanic white and black women 45–64 were childless. Age homogamy of first marriages have increased over time (U.S. Census Bureau, 2011) but increases in the PSID data are larger. Because some of the women in our sample are not in their first marriage, exactly comparable statistics are difficult to obtain. Childlessness is slightly lower in the PSID in 1988 at 8.6%. The CPS fertility supplements suggest that childlessness has increased over time by about 7 percentage points for women age 40–44 (Pew Research Center, 2010). Childlessness among women age 40–44 matches the Pew Report on Childlessness (2010) closely. For this older age group, childlessness has also increased in the PSID but by about 3.5 percentage points. Age of mother at birth has a slightly smaller N (1120 for 1988 and 1225 for 2007) because of missing data on mother’s year of birth.

Table 2 shows that the average total number of own parents and the average total number of parents plus in-laws is increasing over time. The average number of own parents alive increases by 45 percent between 1988 and 2007 and the average number of parents-in-law increases by 43 percent. Table 2 shows that there are not significant changes in the age difference between mothers and daughters over time. Data on age differences between fathers and daughters is reported less consistently in the PSID, however, age differences between husbands and wives in the parent generation has decreased over time so smaller age differences between fathers and daughters in the later cohort could also contribute the increase in the number of own parents (U.S. Census Bureau 2011). Combined, these trends suggest that most of the increase in the average total number of own parents is the result of increases in parental life expectancy with some contribution from increases in the age homogamy of marriage in the parent generation. Table 2 also shows an increase in the age homogamy of marriage for the middle generation over time and a six percentage point decline in the percent of women who are married. The decline in marriage implies that the total number of women with both parents and in-laws is falling over time though increases in the age homogamy of marriages for the middle generation increase the probability of having a living in-law conditional on marriage. The trends in marriage and age homogamy of spouses in the middle generation—which offset one another—and increases in parental life expectancy and parental age homogamy have combined to increase the average total number of parents plus in-laws over time. The similar increase in the number of own parents and the number of parents-in-law implies that the increase in longevity and increased age homogamy of marriages in both the parent and middle generation more than compensates for the decline in marriage. Table 2 also shows a four percentage point increase in the percent of women who are childless, a factor that on its own would imply a decline in the percent of women with living parents and children. Table 2 shows that the increase in the number of living parents is the main factor that drives the increase in the population of women with children and living parents over time.

Table 3 examines the full distribution of the number of parents by marital status. For married women, there is an over 40 percent decline in the percent of women without any living parents. There are also nearly triple the percent of married women with three or more parents in 2007 over 1988. We see similar trends for single women with a declining percent of women without any living parents and increasing percent of women with one and two parents. Overall in the sample, the percent of women without any parents declined 13 percentage points to 27 percent. The percent of women with one parent remained constant while the percent with two or more living parents increased by 14 percentage points with most of the increase, 10 percentage points, occurring among women with three or four living parents.

Table 3.

Distribution of Parents, Women Age 45–64

PSID
1988 2007
Married
 0 Parents/In-laws 33.5 19.5
 1 Parent/In-law 32.9 28.6
 2 Parents/In-laws 22.9 24.9
 3 Parents/In-laws 8.0 19.1
 4 Parents/In-laws 2.7 7.9
Single
 0 Parents 55.7 41.8
 1 Parent 34.9 41.5
 2 Parents 9.4 16.7
All
 0 Parents/In-laws 40.6 27.1
 1 Parent/In-law 33.5 33.0
 2 Parents/In-laws 18.6 22.2
 3 Parents/In-laws 5.5 12.5
 4 Parents/In-laws 1.8 5.2
N 1,207 1,369
Open in a new tab

Notes: Weighted using individual weights.

The number of women in late middle age who have the potential to experience competing demands up and down the generations has increased by approximately 10 percentage points or twenty percent since 1988. This increase has been driven by an increase in the number of living parents with trends in childlessness and marriage exerting countervailing pressure. While changes in the pattern of fertility—namely the increase in the age of first birth and the reduction in the number of births conditional on having children—cannot affect the number of women who are “potentially sandwiched”, these patterns can affect the demands that these women may face. Similarly, increases in life expectancy may increase both the number of women with living parents, and the number and age of parents conditional on having living parents. To examine the change over time in potential demands both up and down the generations we examine several characteristics of the sample of women with living parents and children in 1988 and 2007.11

Table 4 shows the marital status and age of women with children and living parents in both 1988 and 2007 along with the number and average age of their parents and children and the percent of women any siblings and with sisters. Even though marriage rates have declined among women 45–64 overall, marriage rates among women with children and living parents have remained relatively constant around 79 percent. The average age of women with children and living parents has also remained similar at around 53 years of age. Table 4 also shows the average size of the older and younger generation. There has been a 13 percent increase in the mean number of own parents and a 19 percent increase in the mean number of parents and in-laws among women with children and living parents. Parents are older by about one year on average. Even if healthy life expectancy has increased, women have a larger number of parents for whom they may be expected to provide care.

While the potential demands from parents may be increasing, the burden of caring for parents is often shared among siblings with sisters typically providing more care than brothers (McGarry 1998, Pillemer and Suitor 2013). Table 4 shows that nearly all of the women in both cohorts have at least one sibling though the probability of having a sibling has increased over time. The probability of having a sister has remained constant over time and is substantial with over 75% of women having a sister.

In contrast to the increase in the number of parents, women with children and living parents in 2007 have fewer children overall than their counterparts in 1988—nearly one fewer child on average. However, their children are younger on average—oldest children are nearly two years younger while youngest children are the same age.12 Younger children may require more financial resources and be more likely to co-reside (Furstenberg 2005, Schoeni and Ross 2005).

Table 5 shows the distribution of the number of children and number of parents for women with children and living parents in 1988 and 2007. The full distribution allows us to examine more closely the change over time in the relative numbers of parents and children. Table 5 shows that the biggest change for the child generation has been a decline in the percent of women with more than three children and an increase in the percent of women with one or two children. For the parent generation, there has been a parallel increase in the percent of women with three and four parents and a decline in the percent of women with only one living parent. In 2007, women 45–64 with children and living parents are likely to have fewer children and more parents while women in 1988 had more children and fewer parents.

Table 5.

Distribution of Parents and Children of Women with Children and Living Parents, Age 45–64

PSID
1988 2007
Number of Children
 1 10.2 16.6
 2 27.7 45.0
 3 23.7 27.2
 4 18.2 8.2
 5+ 20.2 3.0
Number of Parents or In-Laws
 1 55.9 44.0
 2 31.4 30.8
 3 9.9 17.6
 4 2.8 7.6
N 642 858
Open in a new tab

Notes: Weighted using individual weights.

Having living parents in late middle age need not imply that women are giving or anticipate giving simultaneously to multiple generations, or that there are unmet needs in multiple generations. Indeed there may be potential benefits as well as potential costs of having intergenerational ties. However, understanding how the population of women with living parents and children has changed over time is a necessary first step in understanding the costs and benefits that these women may experience. Changes in parental life expectancy and increases in the age homogamy of marriages have lead to a ten percentage point increase in the percent of women with living parents and children between 1988 and 2007. This increase in the percent of women with children and living parents is in spite of declines in marriage rates and increases in childlessness over time. Among women with children and living parents, there has been an increase in the number of living parents and a decrease in the total number of children over time. The increase in the number of parents has been coupled with an increase in the average age of parents and the decline in the number of children has been accompanied by a decline in their average age. In 1988, the median women with children and living parents had three children and one parent. In 2007, the balance has shifted to two parents and two children.

Results: Actual Transfers for Baby Boom Cohort

In describing the changes over time in the percent of women with children and living parents and the context in which women with children and living parents find themselves in terms of their age, and the age of their potential dependents, several general themes emerge. First, the potential for being sandwiched has increased quite dramatically over time. Second, this trend has been driven mainly by changes in life expectancy of the parent generation. Finally, among those with living parents and children in late middle age, the balance has shifted towards having more aging parents and fewer young adult children. These general themes are consistent with the larger demographic trends in life expectancy, birth timing, fertility, and marriage. To further understand the actual transfers that women make in late middle age, in this section we focus on current and past transfers of money or co-residence with parents and children for the Baby Boom cohort using data from the 1997–2007 PSID.13

Most of the literature defines women to be sandwiched caregivers if they provide support to both parents and children in terms of time or money at a single point in time (Grundy and Henretta 2006, Pew Research Center 2013a, Pierret 2006). However, as we note in the introduction, the concept of being sandwiched is complicated because transfers are not always given in response to needs, women may anticipate transfers even when none are given, and women may be reaping rewards of having multigenerational ties even if these ties sometimes come with obligations. Our data does not allow us to capture all of these complexities. However, we can take a more broad interpretation by examining transfers over a longer period of time. As one example, children may not be getting money transfers in every year, rather, they may only receive money transfers when they are unemployed. Similarly, parents may not require care each year but instead only if they are sick. If in one year, a woman gives money to an unemployed child and then next gives care to an aging parent, she would not be sandwiched between children and parents under the strict definition of the term. However, the cyclical nature of the demands suggest she is still facing competing demands from both generations—her behavior in the labor market, for example, may reflect that she anticipates actual transfers to both parents and children simultaneously. We look at measures of support over time to explore one of the complexities in defining a woman as sandwiched between parents and children.

We begin similar to the existing literature by analyzing current transfers of money to parents and children and co-residence with parents and children for women with children and living parents in 2007. Table 6 shows current transfers of money and/or co-residence, to children, to parents and to both children and parents. Consistent with the Pew Center Report (2013a), transfers of both money and co-residence are more common with children than with parents. Co-residence is also more common with younger children although approximately 10 percent of women with parents and children live with a child over 25 (with or without parents present). If we consider money transfers and co-residence with any child together, about 3 percent of women with living parents and children are making a financial transfer or co-residing with more than one generation. We do not have any measure of current transfers of care so our figures of giving transfers to multiple generations, in its strictest form, are a lower bound estimate.14

Table 6.

Current Giving to Children or Parents by Women with Children and Living Parents, Age 45–64 in 2007

Transfers
To Children To Parents To Children and Parents
Money Support 8.9 2.7 1.0
Co-Residence 45.3 2.2 0.5
Co-Residence 18+ 27.8 2.2 0.5
Co-Residence 25+ 9.8 2.2 0.4
Money or Co-Residence (All) 49.4 3.4 2.8
Money or Co-Residence (18+) 34.5 4.2 1.9
Money or Co-Residence (25+) 18.1 4.4 1.8
Open in a new tab

Notes: Weighted using individual weights.

The estimates of giving transfers to multiple generations in 2007 in the PSID are low—lower than estimates from other data—and our purpose is not to argue that these estimates reflect the actual percent of women giving transfers to parents and children at any one time. Time transfers are often larger than financial transfers by about 50 percent and excluding time transfers is likely to exclude many women who are providing care to multiple generations. However, even these lower bound estimates allow for an exploration of the change in the incidence of transfers if we consider periods of time longer than one year. If we are concerned about savings, labor supply decisions, and even physical or emotional strain on middle-aged adults, it is not clear that transfers over a single year are the correct unit of measurement. Giving money to a child and parent concurrently is more likely to strain the budget of a family in a given year, but, if borrowing is possible, giving concurrently would have the same overall effect on savings as giving money to parents and children sequentially. Similarly, sequential demands from multiple generations may in fact have larger implications for labor supply than acute demands in a single year—for example policies like FMLA may make it easier to deal with acute demands while remaining in the labor force than dealing with more chronic demands from family members. For this reason, we use the PSID panel to examine transfers over a longer period of time.

In Table 7, we explore the incidence of giving transfers multiple generations under a more broad definition than in Table 6. We combine information on current transfers to children and parents with the incidence of past transfers to parents and children that we outlined in our discussion of the data. Table 7 shows that while in any given year only nine percent of parents report having given a money transfer to a child, over 30 percent of parents have given a money transfer to a child in the last ten years. Similarly, less than three percent of women report giving money to a parent in a single year but over twelve percent of women report having given substantial money transfers to parents in the past. For co-residence, extending the period about which transfers are calculated also dramatically increases the incidence of co-residence with both parents and children. In 2007, 45 percent of women live with a child and 2.2 percent of women live with a parent. However, 89 percent of women have lived with a child in the past ten years and nearly 10 percent of women have lived with a parent for one year or more in the past. Even excluding co-residence with young children, 19 percent of women have lived with a child over 25 in the past 10 years. Extending the period over which transfers are considered also dramatically increases the percent of women who have given transfers of money or co-residence to both children and parents from less than three percent (at the most) in a given year to approximately 15 percent.

Table 7.

Current and Past Giving to Children or Parents by Women with Children and Living Parents, Age 45–64 in 2007

Transfers
To Children To Parents To Children and Parents
Money Support 31.3 12.6 5.5
Co-Residence 89.4 9.8 8.0
Co-Residence 18+ 70.5 9.8 7.2
Co-Residence 25+ 18.9 9.8 2.2
Money or Co-Residence (All) 89.4 20.7 17.9
Money or Co-Residence (18+) 78.1 20.7 16.1
Money or Co-Residence (25+) 44.3 20.7 10.6
Money, Co-Residence, or Parent Care 89.4 45.1 39.1
Money, Co-Residence (18+), or Parent Care 78.1 45.1 35.4
Money, Co-Residence (25+), or Parent Care 44.3 45.1 21.7
Open in a new tab

Notes: Weighted using individual weights.

Finally, while we cannot examine current care to parents or any care for children, we can examine past care for parents. Nearly 65 percent of women with children and living parents report having given significant care to parents in the past. We combine this with co-residence and money transfers to examine a more complete set of transfers. We find that including care for parents increases the percent of women who giving to multiple generations to about 30 percent depending on how we count co-residence with children. This is still a lower bound estimate as it excludes transfers of time with parents that are either not for care or not “significant” and it excludes all transfers of time with children. These estimates are much closer to the 30 percent estimated by Grundy and Henretta (2006) and Pierret (2006).

Even after recognizing that transfers to multiple generations need not take place at the same time for women to be sandwiched caregivers, using transfers to measure needs and demands remains problematic. Women may be sandwiched between parents and children because both generations have unmet needs but may be unable to provide for these needs because of financial or time constraints. Similarly, women may be giving transfers to multiple generations even though neither generation is particularly needy. The potential to face demands from both generations may also be important and even when transfers are not given, anticipating transfers may be emotionally costly and may also affect behavior such as savings or labor supply. Because of these complexities, we do not use transfers to define sandwiched caregivers but instead focus on showing that the population at risk for facing competing demands is increasing over time and that many women give transfers to multiple generations over the course of late middle age.

Conclusions

As women in the Baby Boom generation begin to retire, they face a larger potential for intergenerational demands than previous cohorts. We show that over 60 percent of Baby Boom women have a living parent or in-law and at least one child, up ten percentage points from their parents’ generation. The increase over time is driven by increases in the probability of having living parents—mainly a product of changes in life expectancy. Changes in parental mortality are large enough to increase the likelihood of having children and living parents over time despite declines in fertility and marriage and increases in the number of women without any children. We also show that women in the Baby Boom generation who have children and living parents are more likely to have more parents and fewer children than an earlier cohort. On average, the parents of the Baby Boom cohort are older and children are younger than earlier cohorts. While healthy life expectancy has increased, so too has support to young adult children. Overall, the trends outlined here point to an increase over time in the potential for transfers to multiple generations.

In the context of an increase in the number of women who are at risk to face demands from multiple generations, we explore actual transfers to children and parents for the Baby Boom cohort. Using data from the 2007 PSID we find that only 3 percent of women are currently making transfers of money or co-residence with multiple generations. However, our estimates of proving concurrent support to multiple generations are low because of the limitations of information on transfers in the 2007 PSID. In particular, this estimate understates the transfers in two major ways: it does not capture transfers of time and is only focused on the present in a dynamic context that unfolds over relatively long periods. Despite the limitations of the transfer data, we show that the estimates of providing support to multiple generations increase dramatically when we consider transfers over a longer time period. When we examine past transfers to parents and children and include care or time, transfers to parents and children increase substantially and care is the dominant component of support that women report having given to parents in the past. We show that considering transfers to multiple generations that are consecutive and not just concurrent increases the estimates of transfers to multiple generations ten fold. Thus, the range of our estimates of is quite large (as is true of other estimates, such as those of Pierret (2006) with NLS-YW data). Transfers to multiple generations rises to 15 percent when we add in past money transfers and past co-residence with parents and children and to approximately 30 percent when we add in past care for parents. Analysis of current and past transfers reveal that a substantial fraction of women are likely to give to parents and children at some point. These transfers may not always occur simultaneously but the potential for transfers exists over a long period of time. Intergenerational demands may affect well-being in retirement if women decrease labor supply to care for parents (Van Houtven et al. 2013) or if money transfers to children crowd-out retirement savings.

For the Baby Boom cohort, supporting multiple generations is particularly salient, as retirement age and increasing intergenerational ties have coincided with the Great Recession. The Great Recession reduced wealth and increased unemployment among older workers while at the same time increasing the demand for intergenerational support to young adults from their Baby Boom generation parents. While stock markets and housing markets are recovering, the Great Recession has had a large, lasting effect on the labor market. Declining asset values and increases in unemployment that pushed some Baby Boomers who became unemployed to “retire” early and collect Social Security benefits and have made financial transfers to young adults children more difficult (Coile and Levine 2011). At the same time, high youth unemployment has increased the demand from children for financial transfers and co-residence (Pew Research Center 2013b).

In this paper we analyze transfers to parents and children between 1997 and 2007—before the Great Recession—and transfers have likely increased since 2007. Even in 2007, we show that the population at risk of supporting multiple generations has increased over time and that transfers are substantial for the Baby Boom cohort. The 2007 data do have limitations. These data do not allow for a detailed analysis of time transfers to both generations nor does the timing allow for an analysis of transfers during the Great Recession. Perhaps even more importantly, these data do not contain detailed information on the needs of the parent and child generation in terms of health, assets, employment, or number of children and so we are limited to exploring actual transfers instead of needs or demands. New data on time and money transfers was collected in the PSID in 2013. These data include more detailed information on the financial, demographic, and employment characteristics of children and parents and allow for an exploration of needs as well as transfers. In the future, we plan to use these rich data on transfers to and from parents and children to further analyze transfers between Baby Boom generation retirees and their parents and children in the wake of the Great Recession.

Acknowledgments

The authors gratefully acknowledge support of this research from the Alfred P. Sloan Foundation (grant no. 2011-6-24). We also thank Sung S. Park for exceptional research assistance and appreciate helpful comments from Megan Sweeney, Judith Seltzer, and Joan Kahn.

Footnotes

1

We do not attempt to separate period and cohort effects. We do sometimes refer to time trends but time and cohort are not separable in this analysis.

2

Note that our cohorts overlap by one year. This is because of biennial interviewing in the PSID and the inclusion of questions about parents in the 2007 PSID interview. There are 93 women in 2007 who are also in our 1988 sample. Results excluding these women are unchanged.

3

We include only whites and blacks in the analysis and exclude Asians and Latinos. The PSID added a Latino subsample in 1990 that was dropped in 1995 and added an post-1968 immigrant subsample in 1997 and 1999. Both are excluded from the analysis. The sample in 1988 did not include a population representative sample of Asian and Latino women and so including Asians and Latinos would make cross-cohort comparisons impossible.

4

There are some women in the first cohort who have children who had already left the parental home by the time of the original 1968 PSID interview. These children are not sample members. However, because information on children in this paper is derived from birth histories, having children who are not PSID sample members does not affect the analysis.

5

Knowing about living parents is quite rare but even information on children is difficult to find. For example, the CPS fertility supplement is not given to women over 44 and the ACS only asks about births in the last year.

6

Ideally we would also look at actual caregiving and support over time but measures collected in 1988 are not strictly comparable to those collected in 2007. We therefore opt to provide estimates from the most recent time period.

7

We include people who have provided child support as providing financial support to children because it is not possible to separate child support from other support for individuals giving both child support and other support. However, we have calculated financial support to children excluding all support for respondents who gave any child support to children. Excluding these individuals reduces the percent giving to children by approximately 5 percentage points in each case.

8

New data collection in the 2013 PSID will ask information on financial transfers for parents and children separately.

9

All analyses include biological or adopted parents and current step-parents. We do not have information on more distant kin ties such as former step-parents. Some individuals may have a parent about whom they know very little—including not knowing if they are alive. We do not know whether respondents have such a parent but these parents are not counted as living. Therefore any increase in the number of absent parents over time would bias our estimates of the change in the “potentially sandwiched” downward.

10

See notes to Table 2 on comparisons of marriage rates, age of birth, age of spouses, and childlessness with other data sources.

11

In the text below when we refer to women who have children and living parents we are including those who have living in-laws but no living parents. We do this for expositional ease.

12

As we outline in the introduction, these figures imply that the age at first birth is rising but due to the decline in the total number of children, the age at last birth has changed very little.

13

We do not use the 1988 data in this section because the transfers are not comparable to those collected in 2007. Data in the 2013 PSID that will be more comparable to those collected in 1988. When final results of the 2013 PSID become available, we will be able to assess inter-cohort change in actual, not just potential, sandwich caregiving. For this paper, we choose to focus on the 2007 time point because these data provide the most recent data with which to assess “actual” support for two generations.

14

Indeed these estimates are much lower than previous estimates that ranged about 30 percent (Grundy and Henretta, 2006; Pierret, 2006). We believe this is because we are unable to include any transfers of time—other than the implied transfer through co-residence, and because of the nature of questions on financial transfers in the PSID outlined in the data section.

Contributor Information

Emily E. Wiemers, University of Massachusetts

Suzanne M. Bianchi, University of California, Los Angeles

References

  1. Altonji Joseph G, Hayashi Fumio, Kotlikoff Laurence. The Effects of Income and Wealth on Time and Money Transfers between Parents and Children. In: Mason Andrew, Tapinos George., editors. Sharing the Wealth. Oxford: Oxford University Press; 2000. pp. 306–357. [Google Scholar]
  2. Arrondal Luc, Mason Andre. Family Transfers Involving Three Generations. Scandinavian Journal of Economics. 2001;103(3):415–443. [Google Scholar]
  3. Attanasio Orazio. Cohort Analysis of Saving Behavior by U.S. Households. Journal of Human Resources. 1998;33(3):575–609. [Google Scholar]
  4. Bengtson Vern L. Beyond the Nuclear Family: The Increasing Importance of Multigenerational Bonds. Journal of Marriage and Family. 2001;63(1):1–16. [Google Scholar]
  5. Brown Dustin C, et al. The significance of education for mortality compression in the United States. Demography. 2012;49:819–40. doi: 10.1007/s13524-012-0104-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bryson Kenneth R, Casper Lynne M. Current Population Reports. Washington, DC: U.S. Bureau of the Census; 1999. Co-resident Grandparents and Their Grandchildren; pp. P23–198. [Google Scholar]
  7. Chassin Laurie, Macy Jon T, Seo Dong-Chul, Presson Clark C, Sherman Steven J. The Association between Membership in the Sandwich Generation and Health Behaviors: A Longitudinal Study. Journal of Applied Developmental Psychology. 2010;31(1):38–46. doi: 10.1016/j.appdev.2009.06.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cherlin Andrew J. The Marriage Go-Round: The State of Marriage and the Family in America Today. New York: Knopf; 2009. [Google Scholar]
  9. Christensen Kaare, Doblhammer Gabriele, Rau Roland, Vaupel James W. Ageing Populations: The Challenges Ahead. Lancet. 2009;374(9696):1196–208. doi: 10.1016/S0140-6736(09)61460-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Coile Courtney, Levine Phillip. The Market Crash and Mass Layoffs: How the Current Economic Crisis May Affect Retirement. The BE Journal of Economic Analysis & Policy (Contributions) 2011;11(1):Article 2. [Google Scholar]
  11. Coward Raymond T, Dwyer Jeffrey W. The Association of Gender, Sibling Network Composition, and Patterns of Parent Care by Adult Children. Research on Aging. 1990;12(2):158–181. doi: 10.1177/0164027590122002. [DOI] [PubMed] [Google Scholar]
  12. Cox David, Stark Oded. On the Demand for Grandchildren: Tied Transfers and the Demonstration Effect. Journal of Public Economics. 2005;89:1665–1697. [Google Scholar]
  13. Cutler David. Declining Disability among the Elderly. Health Affairs. 2001;20(6):11–27. doi: 10.1377/hlthaff.20.6.11. [DOI] [PubMed] [Google Scholar]
  14. Crimmins Eileen M. Trends in the Health of the Elderly. Annual Review of Public Health. 2004;25(1):79–98. doi: 10.1146/annurev.publhealth.25.102802.124401. [DOI] [PubMed] [Google Scholar]
  15. Dwyer Jeffrey W, Coward Raymond T. A Multivariate Comparison of the Involvement of Adult Sons versus Daughters in the Care of Impaired Parents. Journal of Gerontology: Social Sciences. 1991;46:S259–S269. doi: 10.1093/geronj/46.5.s259. [DOI] [PubMed] [Google Scholar]
  16. Eggebeen David J. Cohabitation and Exchanges of Support. Social Forces. 2005;83:1097–110. [Google Scholar]
  17. Fingerman Karen L, et al. Who Gets What and Why? Help Middle-Aged Adults Provide to Parents and Grown Children. Journals of Gerontology: Social Sciences. 2010;66B(1, 2010):87–98. doi: 10.1093/geronb/gbq009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Vicki Freedman, et al. Trends in Late-Life Activity Limitations in the United States: An Update from Five National Surveys. Demography. 2013;50(2):661–671. doi: 10.1007/s13524-012-0167-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Furstenberg Frank F, Jr, Rumbaut Ruben G, Settersten Richard A., Jr . On the Frontier of Adulthood: Emerging Themes and New Directions. In: Furstenberg Frank F Jr, Rumbaut Ruben G, Settersten Richard A Jr, editors. On the Frontier of Adulthood: Theory, Research, and Public Policy. Chicago: Univ. of Chicago Press; 2005. pp. 3–25. [Google Scholar]
  20. Grundy Emily, Henretta John C. Between Elderly Parents and Adult Children: A New Look at the Intergenerational Care Provided by the 'Sandwich Generation,'. Ageing & Society. 2006;26(5):707–722. [Google Scholar]
  21. Henretta John C, Grundy Emily, Harris Susan. Socioeconomic Differences in Having Living Parents and Children: A U.S. – British Comparison of Middle-Aged women. Journal of Marriage and Family. 2001;63:852–867. [Google Scholar]
  22. Heuser Robert L. DHEW Publication No. 76–1152. Rockville, MD: National Center for Health Statistics; 1976. Fertility Tables for Birth Cohorts by Color: United States, 1917–1973. [Google Scholar]
  23. Hogerbrugge Martijn JA, Dykstra Pearl A. The Family Ties of Unmarried Cohabiting and Married Persons in the Netherlands. Journal of Marriage and Family. 2009;71:135–145. [Google Scholar]
  24. Kahn Joan R, McGill Brittany S, Bianchi Suzanne M. Help to Family and Friends: Are there Gender Differences at Older Ages? Journal of Marriage and Family. 2011;73:77–92. doi: 10.1111/j.1741-3737.2010.00790.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kahn Joan R, Goldscheider Frances, Garcia-Manglano Javier. Growing Parental Economic Power in Parent-Adult Child Households: Coresidence and Financial Dependency in the United States, 1960–2010. Demography. 2013;50(4):1449–1475. doi: 10.1007/s13524-013-0196-2. [DOI] [PubMed] [Google Scholar]
  26. Kennedy Sheela, Bumpass Larry. Cohabitation and trends in structure and stability of children’s family lives. Paper presented at the Annual Meeting of the Population Association of America; Washington, DC. 31 March – 2 April.2011. [Google Scholar]
  27. Kirmeyer Sharon E, Hamilton Brady E. NCHS Data Brief 68. Hyattsville, MD: National Center for Health Statistics; 2011. Childbearing Differences Among Three Generations of U.S. Women. [PubMed] [Google Scholar]
  28. Künemund Harald. Changing Welfare State and the “Sandwich Generation”: Increasing Burden for the Next Generation? International Journal of Ageing and Later Life. 2006;1(2):11–29. [Google Scholar]
  29. Loomis Laura S, Booth Alan. Multigenerational Caregiving and Well-being: The Myth of the Beleaguered Sandwich Generation. Journal of Family Issues. 1995;16(2):131–148. [Google Scholar]
  30. Martin Joyce A, et al. National Vital Statistics Reports. 1. Vol. 60. Hyattsville, MD: National Center for Health Statistics; 2012. Births: final data for 2009. [PubMed] [Google Scholar]
  31. Martin Linda G, Schoeni Robert F, Andreski Patricia. Trends in the Health of Older Adults in the United States: Past, Present, and Future. Demography. 2010;47(supplement):S17–S40. doi: 10.1353/dem.2010.0003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. McGarry Kathleen. Caring for the Elderly: The Role of Adult Children. In: Wise David A., editor. Inquiries in the Economics of Aging. Chicago: University of Chicago Press; 1998. pp. 133–163. [Google Scholar]
  33. McGarry Kathleen. Does Caregiving Affect Work: Evidence Based on Prior Labor Force Experience. In: Wise David A, Yashiro Naohiro., editors. Health Care Issues in the United States and Japan. Chicago: University of Chicago Press; 2006. pp. 209–228. [Google Scholar]
  34. McGarry Kathleen, Schoeni Robert F. Transfer Behavior in the Health and Retirement Study. Journal of Human Resources. 1995;30(supplement):S184–S226. [Google Scholar]
  35. Miniño Arialdi M, Murphy Sherry L, Xu Jaiquan, Kochanek Kenneth D. National Vital Statistics Reports. 10. Vol. 52. Hyattsville, MD: National Center for Health Statistics; 2011. Deaths: Final Data for 2008. [PubMed] [Google Scholar]
  36. Pebley A, Rukdin LL. Grandparents Caring for Grandchildren: What Do We Know? Journal of Family Issues. 1999;20:218–42. [Google Scholar]
  37. Pew Research Center. Childlessness up Among All Women; Down among Women with Advanced Degrees. 2010 http://www.pewsocialtrends.org/2010/06/25/childlessness-up-among-all-women-down-among-women-with-advanced-degrees/June.
  38. Pew Research Center. The Sandwich Generation: Rising Financial Burdens for Middle-Aged Americans. 2013a http://www.pewsocialtrends.org/2013/01/30/the-sandwich-generation/January.
  39. Pew Research Center. A Rising Share of Young Adults Live in their Parents’ Home. 2013b http://www.pewsocialtrends.org/2013/08/01/a-rising-share-of-young-adults-live-in-their-parents-home/August.
  40. Pierret Charles R. The “Sandwich Generation”: Women Caring for Parents and Children. Monthly Labor Review. 2006 Sep;:3–9. [Google Scholar]
  41. Pillemer Karl, Jill Suitor J. Who Provides Care? A Prospective Study of Caregiving Among Adult Siblings. The Gerontologist. 2013 Jul 9; doi: 10.1093/geront/gnt066. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Raley R Kelly, Bumpass Larry. The Topography of the Plateau in Divorce: Levels and Trends in Union Stability after 1980. Demographic Research. 2003;8:246–58. [Google Scholar]
  43. Schoeni Robert F, Ross Karen E. Material Assistance Received from Families during the Transition to Adulthood. In: Furstenberg Frank F Jr, Rumbaut Ruben G, Settersten Richard A Jr, editors. On the Frontier of Adulthood: Theory, Research, and Public Policy. Chicago: Univ. of Chicago Press; 2005. [Google Scholar]
  44. Singh Gopal K, Siahpush Mohammad. Widening Socioeconomic Inequalities in US Life Expectancy, 1980–2000. International Journal of Epidemiology. 2006;35:969–979. doi: 10.1093/ije/dyl083. [DOI] [PubMed] [Google Scholar]
  45. Uhlenberg Peter. Mortality Decline in the Twentieth Century and Supply of Kin Over the Life Course. The Gerontologist. 1996;36(5):681–685. doi: 10.1093/geront/36.5.681. [DOI] [PubMed] [Google Scholar]
  46. U.S. Census Bureau. Estimated Median Age at First Marriage, by Sex: 1890 to the Present. 2011. Nov, [Google Scholar]
  47. Van Houtven Courtney H, Coe Norma B, Skira Meghan M. The Effect of Informal Care on Work and Wages. Journal of Health Economics. 2013;32:240–252. doi: 10.1016/j.jhealeco.2012.10.006. [DOI] [PubMed] [Google Scholar]
  48. Wang Ying, Marcotte Dave E. Golden Years? The Labor Market Effects of Caring for Grandchildren. Journal of Marriage and Family. 2007;69(December):1283–1296. [Google Scholar]
  49. Watkins Susan Cotts, Menken Jane A, Bongaarts John. Demographic Foundations of Family Change. American Sociological Review. 1987;52(3):346–358. [Google Scholar]
  50. Williams Cara. The Sandwich Generation. Perspectives on Labour and Income. 2004;16(4):7–14. [Google Scholar]
  51. Wolf Douglas A, Soldo Beth J, Freedman Vicki. The Demography of Family Care for the Elderly. In: Hareven Tamara., editor. Aging and Generational Relations Over the Life Course : A Historical and Cross-Cultural Perspective. Hawthorne, NY: Walter de Gruyter; Los Angeles, CA: Rand Corporation; 1996. pp. 115–138. [Google Scholar]
  52. Wolf Douglas A, Soldo Beth J, Freedman Vicki. The Division of Family Labor: Care for Elderly Parents. Journal of Gerontology: Social Sciences. 1997;52B:102–109. doi: 10.1093/geronb/52b.special_issue.102. [DOI] [PubMed] [Google Scholar]
  53. Wu Lawrence L, Bumpass Larry, Musick Kelly A. Historical and Life Course Trajectories of Nonmarital Childbearing. In: Wu Lawrence L, Wolfe Barbara., editors. Out of wedlock: Causes and Consequences of Nonmarital Fertility. New York: Russell Sage Foundation; 2001. pp. 3–48. [Google Scholar]

RESOURCES