Grandparents as family supporters

If the growing number of grandparents in Germany stimulated little sociological research interest until the mid-1990s, the functional dimensions of grandparent-grandchild relationships raised even less. Grandparents were first addressed as support providers in practice-oriented publications in social work and family therapy. As in the United States, the earliest publications emphasised the important role that grandparents play in managing family crises, such as the divorce, separation or unemployment of their children (Fabian Reference Fabian1994; Fthenakis Reference Fthenakis1998). There were similar early studies in educational sociology that focused on child-care provision by grandparents at times of crisis (Sommer-Himmel Reference Sommer-Himmel2001). A study of the informal-support mobilisation strategies of lone mothers, who are arguably more in need of grandparental support than two-parent families, found that on occasion German grandparents played a vital role in providing emergency child-care (Hoff Reference Hoff2006b).

The notion of ‘intergenerational learning’ has motivated other studies of the functional aspects of the grandparent-grandchild relationship (e.g. Krappmann Reference Krappmann, Krappmann and Lepenies1997; Lange and Lauterbach Reference Lange and Lauterbach1998; Lüscher and Liegle Reference Lüscher and Liegle2003). While Krappmann (Reference Krappmann, Krappmann and Lepenies1997) emphasised the role that grandparents played in the socialisation of their grandchildren, Lüscher and Liegle (Reference Lüscher and Liegle2003) stressed the mutuality of intergenerational learning in the grandparent-grandchild relationship. During the 1990s, several researchers with an interest in intergenerational relations began explicitly to examine the relationship between grandparents and grandchildren (e.g. Lange and Lauterbach Reference Lange and Lauterbach1997, Reference Lange and Lauterbach1998; Lauterbach Reference Lauterbach, Sahner and Schwendtner1995b, Reference Lauterbach2002; Marbach Reference Marbach and Bien1994; Wilk Reference Wilk, Lüscher and Schultheis1993).

In contrast to the concerns of German researchers, the functional aspects of the grandparent-grandchild relationship have been a recurring theme in North American family sociology for over 60 years (Szinovacz Reference Szinovacz and Szinovacz1998). One of the first studies of grandparenthood stressed the supporting role of grandmothers as ‘rescuers’ of families (von Hentig Reference Von Hentig1946); this theme re-emerged during the 1970s and 1980s in the wake of the rising numbers of divorces, ‘out-of-wedlock’ births, and lone mothers (e.g. Cherlin and Furstenberg Reference Cherlin and Furstenberg1986; Kivett Reference Kivett1985). More recently, the supportive function of grandparents has remained high on the US research agenda (see Baydar and Brooks-Gunn Reference Baydar and Brooks-Gunn1998; Silverstein, Giarrusso and Bengtson Reference Silverstein, Giarusso, Bengtson and Szinovacz1998). Bass and Caro (Reference Bass and Caro1996) estimated that American grandparents provided child-care services worth $17–29 billion each year.

Recent American and British publications on functional aspects of the grandparent-grandchild relationship have been concerned with the potentially harmful effects of family change. Cooney and Smith (Reference Cooney and Smith1996) examined the impact of parental divorce on functional solidarity between grandparents and their adult grandchildren – but did not find any association. British researchers have studied grandparenting in divorced families (Ferguson Reference Ferguson2004), in support of lone parents (Harper et al. Reference Harper, Smith, Lechtman, Ruicheva and Zeilig2004), and in reconstituted families (Dimmock et al. Reference Dimmock, Bornat, Pearce, Jones and Harper2004). A shared conclusion was that the maternal grandmother line attempts to preserve frequent contacts with grandchildren, whereas the paternal grandparent line maintains only limited contact. A French study produced similar findings and emphasised the implication that there was a higher propensity for support from the maternal than the paternal line (Gauthier Reference Gauthier2002). Custodial grandparenting is another keenly debated issue. According to Harper (Reference Harper and Johnson2005), some 3.7 million grandparents currently help in raising 3.9 million children in the United States. Silverstein (Reference Silverstein, Binstock and George2006) noted that research has for long tended to depict intergenerational transfers by grandparents as an expression of altruism and self-sacrifice, especially given that grandparents who assume parenting roles often do so at great cost to their material, physical and mental wellbeing (Minkler et al. Reference Minkler, Fuller-Thomson, Miller, Driver, Hayslip and Goldberg-Glen2000). On the other hand, Pruchno (Reference Pruchno1999) emphasised that many custodial grandmothers found the role rewarding.

The introduction in the United States during the 1990s of large, population-based surveys with questions about intergenerational support stimulated the research interest in transfers between grandparents and grandchildren (Kronebusch and Schlesinger Reference Kronebusch, Schlesinger, Bengtson and Harootyan1994). Not only did this provide considerable improvements in the data resources and methodologies for studying supportive relations between grandparents and grandchildren, but several theoretical concepts that were originally developed to explain supportive behaviour between contiguous generations were applied to the grandparent-grandchild relationship. They have provided the theoretical framework for the present analysis.

Hypothesising intergenerational support

Intergenerational support across one and two generations

The concern that population ageing would overburden familial support networks is prominent in the German public debate about the implications of population ageing, which frequently pictures older people as needy and as passive recipients of help. In contradiction to this anachronistic but persistent stereotype, the majority of older people support the younger generations in their families. Several publications based on the first wave of the German Ageing Survey showed that the older generations give far more support than they receive (e.g. Kohli et al. Reference Kohli, Künemund, Motel, Szydlik, Kohli and Künemund2000; Künemund, Motel-Klingebiel and Kohli Reference Künemund, Motel-Klingebiel and Kohli2005; Motel-Klingebiel Reference Motel-Klingebiel2000; Szydlik Reference Szydlik2000). But whereas inter-personal transfers between parents and children have been relatively well charted, those between grandparents and grandchildren have received little attention.

To measure the provision and receipt of informal financial transfers, the interviewees were asked the following questions (in my translation from the German): ‘Many people make gifts of money or gifts in kind, or support others financially, for example, parents, children, grandchildren, other relatives, friends or acquaintances. How does this apply to you? During the previous 12 months, did you give money or large presents to anyone, or did you support anyone financially on a regular basis? [And the other way round]: Did you receive gifts of money or large gifts in kind or did anyone support you financially regularly during the previous 12 months?’ The present analysis is limited to 62–85-year-olds, i.e. the birth cohorts in which the majority had already become grandparents.

Although the questions allowed the provision or receipt of regular financial support to be reported, such exchanges were rare and instead one-off gifts were the rule. Only six per cent of the grandparents aged 62–85 years in 2002 who supported others financially at all gave money regularly to their grandchildren, but 17.5 per cent did so to their children. In contrast, one-off financial gifts to grandchildren were common and by far the most frequent type of private financial transfer. Of the grandparents in this age group who had provided financial support, 90 per cent had given money to their grandchildren, compared with 83 per cent who had made such gifts to their children. Nearly one-quarter made large gifts to their grandchildren – one-fifth did the same for their children. These differentials applied to men and women and to both East Germans and West Germans. Some provided several types of financial transfers, while others reported just one transaction. A limitation of the data is, however, that there is no information on the frequency or amounts of the transfers during the previous year.

Instrumental support was measured similarly by asking for details of the help provided by individuals living outside the interviewee's household. The question was: ‘Apart from caring activities already mentioned and other activities that you carry out as part of your job, during the previous 12 months did you help anyone who does not live in your household with the chores, e.g. with cleaning, small repairs, or shopping? And please tell me about the converse: during the previous 12 months, did anyone who does not live in your household help you with chores, e.g. cleaning, small repairs, or shopping?’ Unlike the questions on financial transfers, the survey did not ask about different types of instrumental support.

Figure 2 summarises the patterns of mutual financial and instrumental support between grandparents and both their children and grandchildren (but not between the two younger generations) in 1996 and 2002. The strength of the arrows on Figure 2 indicates the proportion of the group that provided support during the two years.Footnote ³ The upper part of the Figure depicts a pattern of support between parents and their adult children that was well established during the mid-1990s. Older Germans supported their children mainly by means of financial transfers, and their children reciprocated through instrumental assistance. They also made financial contributions to their grandchildren, but children were far more likely to be recipients of financial support than grandchildren. More than one-quarter (26%) provided financial support to their children at least once during the previous year, compared with 12 per cent who did the same for their grandchildren.

Figure 2. Intergenerational support patterns in Germany, 1996 and 2002.

Figure 2 also demonstrates that the mutual support was asymmetrical. Although the younger generations returned financial transfers with instrumental assistance, they were far less likely to do so than to receive support: 19 per cent of children and only four per cent of grandchildren had helped the grandparents at least once during the previous year, consistent with the intergenerational stake hypothesis. The mutuality of transfers in the parent-child relationship supported the functional solidarity hypothesis, but too few grandparents reported instrumental assistance from their grandchildren to confirm the applicability of the functional solidarity hypothesis. The counter-flows (financial transfers from the younger generations to the grandparents and instrumental help from the grandparents to the younger generations) were insignificant (the thinnest lines on the figure indicate a prevalence of less than one per cent).

Broadly, the patterns described for 1996 persisted in 2002. The lower panel of Figure 2 shows that older Germans were still more likely to support their children financially (23%) than to receive instrumental support (17%). Moreover, a greater proportion transferred financial assistance to their children than to their grandchildren. There had, however, been a significant shift in the proportion that provided financial transfers over the previous year – away from children (−3%) and towards grandchildren (+5%). Support given to grandchildren of course also helps their parents, but nevertheless this was a remarkable change in just a few years.

The impact of grandparents' age on intergenerational support

A limitation of the German Ageing Survey is that it did not collect grandchildren's ages. Table 1 distinguishes four age groups of grandparents (62–67, 68–73, 74–79 and 80–85 years), born respectively in 1929–34, 1923–28, 1917–22 and 1911–16 (1996 sample) and in 1935–40, 1929–34, 1923–28 and 1917–22 (2002 sample). In 1996, financial transfers to grandchildren were most common by the oldest age group (17 per cent reported such gifts), while among those aged 62–67 years only one-half of that percentage made similar transfers. In comparison, there was little variation across the four age groups in the proportions that made financial transfers to children, although the percentage of grandparents that gave to children was significantly higher than the percentage that gave to grandchildren. The age distribution of those who gave financial support to their children was strikingly different from the age distribution of those who gave to grandchildren. Financial transfers to children were most likely by those in the two youngest age groups, while such support to grandchildren was clearly more common from the older age cohorts. It is unclear to what extent this is an age or a cohort effect (or both), but age-differentiation is clearly confirmed.

Table 1. Intergenerational support patterns by age group of grandparents, 1996 and 2002, Germany

Source: German Ageing Survey, 1996: all respondents aged 62–85 years who were grandparents in baseline sample (N=1,586), weighted data. 2002: all respondents aged 62–85 years who were grandparents in replication sample (N=1,211); weighted data.

The receipt of instrumental assistance from children was by contrast largely age-dependent and needs-based, confirming the fourth hypothesis. Instrumental support was most likely to be given to 80–85 year-olds. The likelihood of instrumental help from children exceeded that of financial transfers in the other direction. Evidence for the applicability of both the ‘functional solidarity’ and the ‘needs and capacity’ hypotheses to the parent-child relationship was found, but the ‘intergenerational stake’ hypothesis was not supported for the oldest age-group's relationships with their children, who were more likely to provide instrumental assistance than to receive financial help. The grandparent-grandchild relationship was imbalanced at the grandparents' expense. Grandparents were far more likely to give financial support to their grandchildren than to receive instrumental assistance from them. In fact, they received hardly any instrumental assistance from their grandchildren. This finding supports the validity and dominant influence of the ‘intergenerational stake’ hypothesis in grandparent-grandchild relationships.

It has been noted that a greater percentage of financial transfers went to grandchildren in 2002 than in 1996 (Figure 2). An analysis by age groups revealed that it was the three younger age groups that were responsible for the higher likelihood, while the gifts made by the oldest age group hardly changed. On the other hand, 80–85-year-olds were mainly responsible for the eight per cent decrease since 1996 in financial transfers to children. The remarkable consequence was that the same percentage of children and grandchildren (15%) of the oldest cohort received financial support in 2002. How are these changes to be interpreted? They may signal a period effect, by which changes in the social environment manifested in changed attitudes to support. One may speculate that the increasingly intense public debate on pension reform in Germany since the late 1990s has resulted in greater financial caution among the oldest age groups, who arguably are the most vulnerable to changes in pension, investment and social security income, but the possibility of a cohort effect must be considered. Three of the four age groups examined in 1996 were still in the 2002 sample, specifically those born in 1917–22, 1923–28 and 1929–34. If a particular support pattern was rooted in the life experience of a specific birth cohort – let's say, those born 1929–34 – it should be recognisable among those aged 62–67 years in 1996 and those aged 68–73 years in 2002. This was not, however, the case. Whereas only eight per cent of this cohort reported financial transfers to grandchildren in 1996, more than one-fifth made such gifts in 2002. Similar differences can be found when following the other two 1996 cohorts in 2002. Hence, cohort effects do not appear to be the main explanatory force. Age-related effects, possibly confounded by period effects, were more likely to account for the changed pattern of transfers.

The occurrence of support exchange in intergenerational relations

The presented findings give the impression that intergenerational support followed an exchange pattern, if slightly imbalanced at the expense of the older generation, but was this really the case? The results actually show that the older generation provided mainly financial support to both their children and grandchildren, and that their children assisted the parents with instrumental help. A cross-cohort analysis in 1996 and 2002 enabled a more authoritative test of whether intergenerational support between grandparents and grandchildren followed an exchange pattern. For this analysis, each logical combination of financial (A, yes or no) or instrumental (B) support provision by grandparents (C) or grandchildren (D) was assigned a category value, thereby creating 2⁴ or 16 combinations. They ranged from a comprehensive pattern of transfers, with both financial and instrumental support being provided by both grandparents and grandchildren, to no support transfers at all.

The vast majority of the grandparents aged 62–85 years in the 2002 replication sample indicated that they had provided neither financial nor instrumental support to grandchildren, and that they had not received any such support from them. Most of those who did provide support to their grandchildren only made financial transfers – without receiving anything in return. The mutual exchange of financial and instrumental support between grandparents and grandchildren was practically non-existent – there were only 10 cases of exchanged support. Support exchange was slightly more common between parents and their adult children. Nearly one-fifth of the grandparents aged 62–85 years exchanged financial and instrumental transfers with their children, but the provision of only financial support by the older generation, without receiving anything in return, was again predominant. On the other hand, about one-quarter received instrumental assistance from their children without giving any support in return. Genuine dyadic exchange was far less likely than the summary statistics suggest.

Financial transfers to children and grandchildren

A significant shift in the proportion of grandparents making financial transfers toward grandchildren, seemingly at the expense of children, occurred between 1996 and 2002. To gain a better understanding of the change, we now compare financial transfers towards grandchildren and children more closely, first, by examining the estimated values of the financial transfers. The respondents were asked to indicate the amount of money that they had given to others over the previous 12 months. They were presented with categories of value, ranging from ‘less than €250’ to ‘more than €10,000’. For the analysis, the categories were represented by the mid-class values, e.g. 250–500 was represented by 375. The principal finding is that older people provided far more financial support to their children than to their grandchildren. In 2002, each grandparent aged 62–85 years who provided any financial support gave on average €1,250 to her or his grandchildren, but nearly three times more (€3,500) to his or her children. It is however notable that overall financial support to the younger generations had decreased since 1996. For each of the four age-groups, grandparents in 1996 on average provided a larger sum to both their children and their grandchildren than their counterparts in 2002. Generally, grandfathers transferred a larger amount than grandmothers, which reflects their higher pensions and savings.

Longitudinal perspectives on financial transfers

The main finding of the analyses reported thus far is that German grandparents aged 62–85 years were more likely to support their grandchildren in 2002 than six years earlier. The increasing propensity to support grandchildren was particularly pronounced in the younger age groups, whereas a reduction in the propensity to support children was most evident in the oldest age group. It is important to note that all the findings so far discussed make statements about groups of older people and do not allow conclusions to be drawn about individual behaviour. It may well be that those who gave support to grandchildren in 2002 were not the same as those who did not support their children. To examine the combinations of support further, we first explore if grandparents gave money exclusively to their children, exclusively to their grandchildren, or to both generations.

It is not possible to observe changes of individual behaviour over time from two independent samples, such as the 1996 baseline and 2002 replication samples. The support propensities were therefore re-estimated from the longitudinal sample of the German Ageing Survey. These estimates corroborated the finding that the proportion of grandparents that provided financial support exclusively to their grandchildren rose from one-fifth in 1996 (when aged 62–85 years) to one-third in 2002 (when aged 68–91 years). Over the same period, the share of those who gave money only to their children fell from more than 60 per cent to just over 40 per cent. While the comparison of the 2002 replication and the 1996 baseline samples suggested cross-cohort changes, any cohort effects would have been confounded by age effects in the longitudinal analysis. The sample respondents were aged 62–85 years in 1996; six years later, when they were 68–91 years old, they had more grandchildren, which would at least partly explain the higher percentage making financial transfers to grandchildren. Furthermore, the grandchildren were now older and their need for financial assistance would have been greater. Table 2 presents a more detailed picture of the changes between 1996 and 2002. The combination of four variants of support (financial support to neither children nor grandchildren, to children only, to grandchildren only, to both children and grandchildren) in the first and second wave implied 4² or 16 possible combinations.

Table 2. Changing patterns of financial support of grandchildren and children, Germany 1996 and 2002

Source: German Ageing Survey, panel sample 1996 and 2002 with all respondents aged 62–85 years in 1996 who were grandparents (N=371) and who gave money to children/grandchildren in 1996 or 2002 (N=210).

While the results for 1996 and 2002 suggested that German grandparents shifted a considerable proportion of their financial support away from their children and towards their grandchildren, the longitudinal analysis did not confirm this change. Only 10 grandparents in the selected age group made financial transfers exclusively to their children in 1996 and then in 2002 exclusively to their grandchildren. Similarly, only a low percentage (6.5) of those who gave money only to their children in 1996 gave to both children and grandchildren in 2002. On the other hand, a larger percentage (12.6) of those who provided no financial support in 1996 six years later gave money to their grandchildren but not to their children. Most importantly, however, more than one-fifth of those who had given money to their children in 1996 did not make any financial transfers to children or grandchildren in 2002. All in all, the reduction in financial support of children was more pronounced than that of grandchildren. An age effect seems to be the most likely explanation, given that in 2002 the youngest respondents were 68 years old, and the oldest were well into the ‘fourth age’.

Conclusions: a changing balance of intergenerational support?

The objective of this paper was to look at change and continuity in the patterns of supportive transfers between grandparents and grandchildren in Germany between 1996 and 2002. Whereas most research on the functional aspect of the grandparent-grandchild relationship has focused on child-care issues, this paper has concentrated on two other support types that are at the heart of support exchanges between parents and children: financial transfers and instrumental assistance. Two theoretical constructs led to the hypotheses that have guided this research: the intergenerational solidarity hypothesis of Bengtson and Roberts (Reference Bengtson and Roberts1991), and the intergenerational stake hypothesis of Giarrusso, Stallings and Bengtson (Reference Giarrusso, Stallings, Bengtson, Bengtson, Schaie and Burton1995). Two additional hypotheses that take into account the effect of age on supportive behaviour were proposed: (a) that the extent of support varies with age, and (b) that the intergenerational stake hypothesis does not apply to supportive relations between the younger generations and very old grandparents, because of their high need for instrumental assistance.

To summarise the findings, financial and instrumental support are first considered separately. In 2002, nearly one-fifth of 62–85 year-old grandparents provided financial support to their grandchildren, and one-quarter to their children. A remarkable shift was found in the likelihood of financial transfers between 1996 and 2002, away from adult children and towards grandchildren. From a comparison of the 1996 baseline sample with the 2002 replication sample, it has been shown that the proportion of grandparents that reported financial transfers over the previous year to grandchildren increased by five per cent, but that to children decreased by three per cent. Analysis of the reports from the longitudinal panel in 1996 and 2002 found that an increasing share of 62–85 year-old grandparents gave financial support solely to grandchildren, and a decreasing share gave exclusively to children. Nevertheless, adult children were still more likely to be recipients of financial transfers than grandchildren, and the value of the financial gifts to children was almost three times higher than to grandchildren. When grandparents financially supported their grandchildren, one-off transfers were the rule, and very few provided regular financial support.

The age differentiation hypothesis is that the provision and receipt of financial and instrumental transfers alter with the age of the grandparents, so the analysis was repeated controlling for the grandparents' ages. The results showed that the above-mentioned shift in intergenerational support patterns occurred for all age groups of grandparents. While a higher likelihood of financial transfers to grandchildren was found in all but the oldest cohort, only the oldest cohort reported a considerable decrease in their propensity to provide their children with financial assistance. The fourth hypothesis – the changing needs or capacity hypothesis – was particularly expected to apply to instrumental assistance for grandparents aged 80 or more years. The hypothesis was largely based on the idea that the oldest people need more support for reasons of frailty and multi-morbidity, but also that adult grandchildren have a greater support capacity than young grandchildren, not least to provide instrumental help. In contrast to adult children, very few grandchildren reciprocated financial transfers by means of instrumental assistance in either 1996 or 2002. Thus, the ‘greater needs or capacity hypothesis’ is rejected for the grandparent-grandchild relationship – but there is clear evidence of its validity for the parent-child relationship.

A final comment about the mutuality of support provision is appropriate. This article has shown, in line with previous findings, that the functional dimensions of intergenerational relationships are imbalanced at the older generation's expense. In other words, the support relationship of grandparents with their grandchildren is even more imbalanced than that with their children. The 62–85 year-old German grandparents were generally more likely to support their grandchildren and children than vice versa. This outcome confirms the overall validity of the ‘intergenerational stake hypothesis’, but there was one exception: the alternative ‘greater needs or capacity hypothesis’ applied to very old grandparents. Only for these dyads was the propensity of the younger generation to provide support (in this case instrumental assistance) greater than that of the older generation to provide financial support. The grandparent-grandchild relationship is not therefore characterised by mutual exchange. The substantial ‘downwards’ financial transfers were not matched by reciprocal ‘upwards’ instrumental support – not even toward the ‘oldest old’. Overall therefore, whereas supportive behaviour in the parent-child relationship is governed by the principle of intergenerational (functional) solidarity, the grandparent-grandchild relationship is guided by an ‘intergenerational stake’ and not by intergenerational (functional) solidarity.

What broader conclusions can be drawn from these findings, particularly in the light of anticipated demographic and family changes? The spread of the ‘beanpole family’ (Bengtson, Rosenthal and Burton Reference Bengtson, Rosenthal, Burton, Binstock and George1990) will result in fewer grandchildren, but on the other hand, rising average life expectancies will increase the average number of years that each person spends as both a grandchild and a grandparent. Thus, the prospect of four healthy grandparents – in step-families possibly even more – competing for the attention of just one grandchild is no longer fanciful. A shift in the destination of financial transfers towards grandchildren may signal a changing attitude towards grandchildren, and perhaps they will be valued more highly. Future research on change and continuity in intergenerational norms, values and attitudes should monitor and test this claim.

The analysis reported in this paper has several limitations, not least that it has been confined to financial and instrumental support. The incorporation of other support dimensions, such as emotional support, cognitive support and child-care, would give a more complete understanding of the grandparent-grandchild relationship. A major shortcoming arose from the nature of the data. All national ageing surveys face the dilemma of wanting to collect as much information as possible but being constrained in doing so. There is an obvious trade-off between maximising the data collected and restraining the time spent on the task. Single-purpose surveys of the grandparent-grandchild relationship would provide fuller and better information for the analysis of mutual support relations. Ideally, such surveys should incorporate all aspects of the grandparent-grandchild dyad and the grandparent-child-grandchild triad.