Matthijs Kalmijn
This chapter is concerned with the relationships between parents and their adult, grown-up children. Traditionally, interest in intergenerational ties came from functionalist perspectives in sociology, which argued that the process of modernization – and in particular the growing rates of social and geographical mobility – had undermined the extended family. Empirically, this hypothesis was quickly rejected, however (Litwak, 1960), and more recent historical research showed that the extended family, at least in terms of residence, was not that prevalent before the modernization process to begin with (Ruggles, 2011). More recently, interest in intergenerational ties has increased again as a result of the aging of Western societies. Life expectancy beyond the retirement age has increased considerably, which has implied increasing shared lifetime of the generations. This not only leads to more demand for personal and physical care in times of illness, it also creates a need for social and emotional support from adult children. While this development may have increased the importance of adult intergenerational relationships, other societal trends may hinder such a (re)strengthening of family ties. Women have been the primary kinkeepers; hence, the increase in women’s employment has reduced the amount of time adult children have available for their parents. The increase in divorce has resulted in a growing number of older parents, especially fathers, who are estranged from their children. Finally, both children and parents have been influenced by individualization. For children, normative obligations to support their parents are replaced by considerations of affection. At the same time, older parents increasingly emphasize the wish to be independent of children for practical support and personal care. These conflicting developments make the study of intergenerational ties both more interesting and more relevant.
Adult intergenerational relations have frequently been studied from the perspective of intergenerational solidarity (Rossi and Rossi, 1990; Roberts, Richards, and Bengtson, 1991). In classic sociological writings, the term solidarity was used to explain the cohesion of social groups or societies more generally. Solidarity referred to a tendency to do or mean something for each other and for the collective, while avoiding the pursuit of pure self-interest that would lead to a disintegration of the group. This more general concept was applied to family relations in two ways (Roberts, Richards, and Bengtson, 1991). From classic sociological theories, the idea was borrowed that cohesion in the family would depend on functional dependencies on the one hand (mechanical solidarity) and normative motivations on the other hand (organic solidarity). From classic microsociological work, especially the work by Homans, the idea was borrowed that family cohesion would further depend on levels of interaction and affection. By using the concept of solidarity, authors also placed intergenerational relations in the debate about aging: how and to what extent is a society able to support the elderly, and what is the ideal division of labor between intergenerational solidarity inside the family and solidarity between (age) generations outside of the family (Kohli, 1999)?
The theoretical developments provided earlier resulted in the idea that intergenerational solidarity can be distinguished into six dimensions (Silverstein and Bengtson, 1997): (i) associational solidarity (contact), (ii) functional solidarity (support giving), (iii) structural solidarity (opportunities for contact), (iv) affective solidarity (positive feelings for each other), (v) consensual solidarity (agreement on values), and (vi) normative solidarity (normative obligations to support each other). Some of these dimensions are behavioral or have underlying behavioral choices (i–iii), whereas others are concerned with mental and cognitive aspects of intergenerational relationships (iv–vi).
Most studies, especially those written from an aging perspective, have focused on functional solidarity. Functional solidarity, or support, occurs in both directions, from adult children to parents (upward stream) and from parents to adult children (downward stream). Support is usually distinguished into instrumental (or practical) support, emotional support, and financial support. Instrumental support covers a variety of behaviors, such as help with household work, taking care of grandchildren, mending the house, shopping, running errands, administrative work, and personal care. These forms of support almost always involve face-to-face contact and can be time consuming, especially when recurrent personal care is involved. Personal care is more directly related to health problems, whereas other practical forms of support are driven by a more heterogeneous set of needs and life events (e.g., moving, unemployment). Daughters provide more support than sons, and consequently, they are often depicted as the main kinkeepers, just as they tend to be the main providers of care to children in the home (Hagestad, 1988). Social stratification plays a role too: in low-status families, there is more informal support exchange between parents and children than in high-status families (Van Groenou et al., 2006).
Even if there is not a high level of support exchanged between the generations at any specific point in time, the family does seem to function as a safety net: in times of crises, or when there are specific needs and problems, the family is mobilized for support (Hogan and Eggebeen, 1995). Because such needs and problems vary over the life course, support exchange is also highly dependent on age and other life course factors (Rossi and Rossi, 1990). For example, parents can support adult children by grandparenting, but this is often limited to the stage in which the grandchildren are preschool and primary school age (Hank and Buber, 2009). The notion of the family as a safety net is an element in the so-called convoy model in gerontology, which argues that people have networks that can be called upon for support if needed, that networks members are hierarchically ordered in terms of support potential, and that parents and children form the inner circle of the potential support network (Antonucci and Akiyama, 1987). There has been debate, however, about the degree to which children respond to parental needs (see also later) and the degree to which structural constraints such as proximity complicate this (Hogan and Eggebeen, 1995). Moreover, for personal and physical care, either the spouse or more formal means of support are still more important than the children (Brandt, Haberkern, and Szydlik, 2009; Pinquart and Sörensen, 2011).
Financial support includes transfers of money, the giving of consumer durables, giving money to pay for college, and helping a child buying a home and setting up a new household (Hochguertel and Ohlsson, 2009). Most financial transfers are downward (Kohli, 1999), but among non-Western migrants, upward financial transfers are common as well (Sana and Massey, 2005). Inheritances are also studied as a form of intergenerational transfers (Szydlik, 2004). Economists use the terms inter vivos transfers and testamentary transfers to distinguish the two forms of giving.
Associational solidarity has also been studied frequently, in part because questions on contact have often been included in large national surveys. Contact is frequent in most Western societies. In a comparative study of Austria, (former) West Germany, Great Britain, and the United States, Kalmijn and De Vries showed that about 40%–60% of the children had at least weekly face-to-face contact with their parents (Kalmijn and De Vries, 2009). When combining face-to-face contact and telephone contact, 80%–90% of older adults in Europe have at least weekly contact with their adult children (Hank, 2007). There is no clear trend in contact frequency with the mother; for fathers, there seems to be a decline, probably as a result of divorce (Kalmijn and De Vries, 2009; Treas and Gubernskaya, 2012). Daughters have more frequent contact than sons, and mothers have more contact than fathers (Rossi and Rossi, 1990; Kalmijn, 2007). Higher educated persons have less frequent contact with their parents than lower educated persons, in part because they tend to live further away from their parents (Kalmijn, 2006).
Research on the subjective dimensions of intergenerational solidarity suggests that both parents and children, although parents more than children, evaluate their relationship generally in a positive way (Mandemakers and Dykstra, 2008). It is not well known if intergenerational relationships have become more or less positive over time. There are good theoretical reasons, however, to expect that the trend has been in a positive direction. First, generational differences in values at the macro level have declined due to the slowing down of cultural trends (Inglehart, 1997). This is likely to have affected generational differences within the family as well, although that remains an empirical question. Other possible contributors to a positive trend are the decline in the number of siblings and the shift from an authoritarian to a liberal child-rearing style (Alwin, 1988). Descriptive research further suggests that normative solidarity is strong (Ganong and Coleman, 2005). Interesting is that older persons emphasize filial norms less strongly than do younger persons, especially when it comes down to personal and physical care (Gans and Silverstein, 2006). One explanation for this is that older persons in many cultures have a desire for autonomy.
Intergenerational solidarity was not only a classificatory scheme for various aspects of intergenerational relations; it was also a theory about how different indicators of solidarity were causally related (Roberts, Richards, and Bengtson, 1991). Consequently, much research has been devoted to assessing the causal relationships between the different dimensions of intergenerational solidarity (Bengtson and Roberts, 1991). In many cases, causal connections can go in both ways. For example, contact is a condition for giving support, but when there is much contact, it is easier to detect a need for support so that support exchange becomes more likely. Similarly, when there is affection, contact is more pleasant, but social-psychological theories also argue that contact itself can lead to more affection. Finally, filial norms can motivate support, but the reports of norms can be affected by support exchange. People tend to avoid cognitive dissonance so that they may weaken their support for the norm of filial obligations when they offer little support to their parents.
Initially, the causal relationships were examined with cross-sectional data, using statistical models that allow for mutual causality (Bengtson and Roberts, 1991; Klein et al., 1999). To study such effects in a more convincing fashion, authors increasingly used longitudinal data. For example, Silverstein, Parrott, and Bengtson (1995) showed that contact and affection at one point in time positively affected support giving to older parents at a later point in time while filial norms did not predict later support (Silverstein, Parrott, and Bengtson, 1995). Other studies showed that norms at one point in time do affect later support but only when the need for support increases (Silverstein, Gans, and Yang, 2006). Because causal arrows can go both ways, and because there are six dimensions of solidarity, the theory became more and more complex. A recent attempt to link seven dimensions of solidarity in a panel design with each other yielded 42 crossdimensional effects ((7 × 7) - 7), of which only 13 were significant (Hogerbrugge and Komter, 2012). This does not provide strong support for the causal implications of the solidarity model, but the length of the panel in this study – and hence the degree of change in solidarity – was rather limited.
The structural dimension of solidarity is quite influential empirically. For example, proximity has a strong positive effect on contact and support. Even for this structural factor, however, the causality can go in both ways. Silverstein, for instance, documented that (the demand for) support exchange can lead to moving closer to one’s parents or not moving further away from parents (Silverstein, 1995). Hence, distance is not a pure restriction and should thus not be included routinely as a control variable. Another structural variable is the number of siblings. The more siblings one has, the less contact there is at the level of the parent–child dyad (Eggebeen, 1992). This effect probably reflects real-time constraints from the parents’ perspective. At the level of the parent, however, the effect is different. Here, there is some advantage for parents who have more children: The more children parents have, the more likely it is that there is contact with or support from at least one child (Grundy and Read, 2012).
To bypass the complexities of the model while retaining the notion of underlying dimensions of solidarity, some authors have resorted to the study of typologies (Silverstein and Bengtson, 1997; Van Gaalen and Dykstra, 2006). Using statistical techniques for clustering cases, groups of children or parents are formed on the basis of a specific combination of scores on multiple dimensions of solidarity. What these analyses reveal is that not all dimensions coincide. For example, there are quite a few parents and children who have frequent contact without feeling very positively about the relationship. The assumption is that for these persons, contact is driven by normative obligations rather than by affection. Moreover, there are also the opposite types of relationships, where parents and children are fond of each other without having much contact or providing much support. This type has been called intimacy at a distance.
A criticism of the solidarity perspective is that it has neglected conflict in families (Luescher, 2002). Whether conflict is incompatible with the solidarity theory remains debated, but the criticism is fair: There are more studies on support, contact, proximity, and affection than there are studies that explicitly examine conflict. Although both everyday observations and popular movies and novels suggest that family conflict is prevalent, descriptive studies do not confirm this: a fair number of people report occasional conflict, but only a small minority of people report frequent conflict (Bengtson et al., 2002; Szydlik, 2008). Several authors have argued that even though overt conflict may be rare, the experience of tensions in the intergenerational relationship is common (Cichy, Fingerman, and Lefkowitz, 2003). It is possible that norms of family solidarity prevent conflicts from escalating. Because avoidance is more difficult in ascribed relationships such as relationships with children or parents, interpersonal tensions may then increase.
There are several sources of conflict (Clarke et al., 1999). Conflicts can first be about differences in taste and values, but this is now less likely than it was in 1960s and 1970s, when the generation gap in norms and values was often severe (Inglehart, 1997). Probably a more important cause of conflict lies in the relationship itself (Clarke et al., 1999). One hypothesis here is the generational stake hypothesis, which argues that children are more important to parents than parents are to children (Rossi and Rossi, 1990). Related to this is the developmental schism hypothesis, which argues that parents and children have different needs because they are in different parts of the life course (Fingerman, 1996). For example, the child may be investing in the start of the career and building a family, with little time left for social contacts, while the parents may just be retiring and in need of social and emotional support. Conflict about the relationship itself is probably especially common for the so-called sandwich generation, parents who are taking care of young children at home and older parents at the same time (Grundy and Henretta, 2006). These types of conflicts have also been an issue in immigrant families where expectations and ideals about intergenerational relationships are often different for migrant parents than for their native born children (Treas and Mazumdar, 2002).
Criticisms of the solidarity perspective have argued that conflict and tensions are not merely the negative poles of solidarity, but to some extent orthogonal dimensions. In other words, high levels of contact and support can go hand in hand with conflict. This argument has been made most strongly in the emerging literature on intergenerational ambivalence (Luescher, 2002; Pillemer and Suitor, 2002; Fingerman et al., 2008). There continues to be debate about what ambivalence means, and various conceptualizations are now being used next to each other. One view is that ambivalence is the combination of positive and negative feelings about one’s parents or children. Such feelings occur when the relationship is intimate but also characterized by conflict. Another view is that ambivalence is a conflict between roles. For example, adult children may desire to be autonomous, while they still receive support from parents, for instance, when setting up their own household. This is a conflict between the role of child and the role of adult, and this can lead to intrapersonal feelings of ambivalence. Sometimes, ambivalence refers to feelings of guilt, for example, when children are not supporting their parents while they feel normatively obliged to do so. This is a conflict between norm and behavior. There can also be conflicts between different norms about family ties, for instance, the norm for parents to treat children equally and the norm to help the neediest child the most.
Papers from several different countries show that some parent–child dyads are indeed characterized by high levels of contact, support, and intimacy on the one hand, and frequent conflict on the other hand (Bengtson et al., 2002; Van Gaalen and Dykstra, 2006; Steinbach, 2008; Szydlik, 2008). Although this is in line with the notion of ambivalence, the prevalence of these types of relationships is low, varying from 5% to 15%. Other authors have measured ambivalence by asking people directly about such feelings (Luescher and Pillemer, 1998) or by looking at sum scores of positive and negative feelings (Fingerman et al., 2008). These studies suggest that adult children often feel ambivalent toward parents, whereas ambivalent feelings among parents are less common (Kiecolt, Blieszner, and Savla, 2011). Moreover, young adults experience more ambivalence than older adults (Fingerman, Hay, and Birditt, 2004). These findings suggest that ambivalence is a developmental phenomenon that has to do with the tensions between autonomy and dependence that adults experience when they move through the life course.
Much of the literature has been devoted to testing theories about why parents and children provide each other with support. Two theoretical perspectives have been used, one based on altruism and one based on exchange. The exchange approach is based on the rational egoistic model of man, as used in economics and parts of sociology. The starting point of this approach is, first, that there are costs and benefits involved in engaging in personal relationships and, second, that the balance of these costs and benefits is the main motivation for what people do in a relationship. The interesting point of this theory is the counterintuitive idea that even in intimate and affective relationships, people are rational and selfish. The application of this general idea to intergenerational relationships comes from exchange theory, which basically argues that the costs of supporting someone else – in terms of time and energy – are compensated by the benefits of receiving support in return. Two forms of exchange exist.
First, there is immediate or direct exchange. In this case, parents may support their children in one way while the children support the parents in another way. For example, parents help children with the painting of their house, and subsequently, children have their parents over for dinner at the end of the day. Instrumental and socioemotional support are exchanged. Second, there is intertemporal exchange, which means that parents give something to their children at a young age in the hope that they will be supported by their children when they are old. In other words, exchange is embedded in the life course. What parents give at an early age can be seen as an investment in children. Investments not only take place when the children are living at home, but also when children are making the transition to adulthood. In this stage, parents can support their children in setting up a new household, for example. Intertemporal exchange is made possible by the norm of reciprocity, which reduces the uncertainty that people would have about whether or not they would receive something back. Because there are norms against selfishness in personal relationships, exchange in parent–child relationships is believed to work in a somewhat subtle, implicit, and possibly hidden way (Batson, 1993).
Exchange theory has often been tested by showing that the support that older parents receive is positively affected by the support that parents give to adult children (Klein Ikkink, van Tilburg, and Knipscheer, 1999; Grundy, 2005). Other studies have pointed to the reverse effect: the support children receive is positively affected by the support children give to elderly parents (Leopold and Raab, 2011). These cross-sectional designs point to instantaneous exchange, but they are not clear about the direction of the exchange. Moreover, upward and downward streams may be correlated due to common causes, such as familialistic norms. Within-family comparisons are better able to rule out such common family causes, and these have provided positive evidence as well: mothers give more support to the child from whom they received more support or from whom they expect more support in the future (Suitor, Pillemer, and Sechrist, 2006; Kalmijn, 2013).
Life course research has tried to test the idea of intertemporal exchange. Authors have pointed to the effects of aging on the balance of support exchange. The theory would suggest that at a certain age, the balance of support would reverse. This is often true in non-Western societies (Cong and Silverstein, 2011), but less so in Western societies where even at quite old ages, parents continue to support their adult children more than children support parents (Kohli, 1999). Longitudinal studies of parent–child dyads have shown that the support children received from parents at an earlier point in the life course has significant effects on the support children later give to parents (Silverstein et al., 2002). This is more direct evidence of intertemporal exchange, but the effects were small in magnitude. Moreover, this finding leaves open the question of whether parents are motivated by exchange; it primarily shows that children reciprocate. In one recent study, it is shown that parents give more money to children when, at an earlier point in time, they received more visits from these children, suggesting that parents also engage in intertemporal exchange (Lennartsson, Silverstein, and Fritzell, 2010).
A different theoretical perspective on intergenerational relationships is altruism. There are many definitions of altruism, but a common one is that people take the well-being of others into account as a benefit in their own utility (Schroeder et al., 1994). Like exchange theory, the perspective of altruism is rational in the sense that people weight the costs and benefits of engaging in intergenerational relationships. A difference between altruism and exchange is that altruism requires empathy: one needs to understand what the costs and benefits are for someone else. Altruism is also used as a framework in economic research (Arrondel and Masson, 2006), but it has been used on a larger scale and for a longer period in (social) psychological research (Schroeder et al., 1994). Social psychologists do not argue against egoism; they argue instead that the frame people use depends on the type of relationship: personal relationships are governed by altruism, while instrumental relationships are governed by exchange (Batson, 1993).
The theory of altruism seems incomplete without an explanation of where altruism comes from. One explanation focuses on the biological nature of altruism. If there are altruistic genes, these are more likely to survive since the children of altruists will have a better chance of reaching the reproductive ages. This explanation of altruism applies to parents and not to children. Another explanation of altruism is that there are side benefits of helping, such as more self-esteem, a sense of meaning, and perhaps a general increase in well-being. These considerations have been summarized with the phrase doing well by doing good (Piliavin and Siegl, 2007). That there are side benefits to helping others is plausible, but because they are basically selfish motivations, the explanation blurs the distinction between altruism and exchange.
The theory has been tested indirectly by focusing on the needs of parents and children. The underlying assumption is that meeting another person’s need makes that other person happier, an assumption that is plausible in most cases. Empirical research has provided considerable evidence in favor of this hypothesis when the focus is on what children give to parents. Virtually all indicators of parental need – for example, living alone, health problems – increase the support that children give (Grundy, 2005). An exception appears to lie in financial transfers. Few children give money to their parents, even when parents are living in poverty (Hogan and Eggebeen, 1995). Parents also respond to the needs of children (Bucx, van Wel, and Knijn, 2012). Especially convincing here is evidence coming from multiactor studies. These within-family comparisons show that children who were single, who were lower educated, and who had health or other personal problems received more support than their married, higher educated siblings who had no problems (Suitor, Pillemer, and Sechrist, 2006; Fingerman et al., 2009; Kalmijn, 2013).
Economists have tested exchange and altruism by studying inter vivos transfers (money donations to children) and inheritances. Research has shown that after controlling for parental income, the income level of a child has a negative effect on the chance to receive a financial transfer from parents (Hochguertel and Ohlsson, 2009). Because children with low incomes are more in need of support, this finding is consistent with altruism and shows that parents compensate for economic inequality within the family. The study of inheritances was motivated by the strategic bequest hypothesis, which argued, on the basis of exchange theory, that parents would strategically use their bequest to secure support from their children (Bernheim, Shleifer, and Summers, 1985). Although originally supported by Bernheim, later studies found no evidence that children provide more support to parents when they expect to inherit a larger amount of money or property (Wilhelm, 1996; Sloan, Picone, and Hoerger, 1997).
The study of inheritances does not seem to support altruism either. Despite income differences among children within families, the large majority of parents give all children the same part of the inheritance (Finch and Mason, 2000; Behrman and Rosenzweig, 2004). This evidence is in contrast to the research on intra vivos transfers, which shows that poorer children receive more financial support from parents. Economists have resolved this discrepancy by arguing that parents can conceal financial gifts but not inheritances (Arrondel and Masson, 2006). Sociologists have argued that parents hardly think about the option of not giving equal shares to the children, in other words, equal division is part of the normative frame that parents use (Finch and Mason, 2000). Qualitative interviews suggest that parents do sometimes differentiate by passing on personal items to specific children to express their preferences (Finch and Mason, 2000). Inheritances thus have an expressive, more symbolic meaning that is suggested by either the altruistic or the exchange perspective.
Marriage and divorce have important effects on adult intergenerational relations, in particular for fathers. All studies find that fathers who divorced have a weaker relationship with their adult children (from the dissolved marriage) than fathers who remained married. These effects occur for a number of dimensions, including contact frequency, support exchange, affection, and conflict (De Graaf and Fokkema, 2007; Albertini and Garriga, 2011). Effects are much stronger when the father divorced while the children were still at home, but there are also small negative effects for later divorces (Aquilino, 1994). There is some tendency that negative effects for one parent do not coincide with negative effects for the other parent; hence, there is some compensation (Kalmijn, 2012). This can be explained by the fact that children are often faced with loyalty conflicts after divorce – they feel caught in the middle – which makes them draw closer to one parent (Amato and Afifi, 2006). The findings for fathers are worrisome, not only for children who miss a father figure in their adult life, but also for fathers. Fathers who do not repartner may miss the support they need when they are old and frail. The divorce revolution has not yet hit the elderly, but this will change as the baby boom enters old age.
A common explanation of these findings lies in the notion of investments and hence, in exchange theory. Fathers rarely get custody, and although many divorced fathers remain involved in the lives of their children, this involvement will typically be less intensive than it was before the divorce. An early decline in involvement can be seen as a decline in investment in children, which, according to principles of reciprocity, will lead to less frequent support from and contact with children when children are older. Fewer investments in children during marriage may also lead to lower-quality ties when the children are older. One piece of evidence for this lies in the effects of the age at divorce. The older the children are when they experience the divorce of their parents, the shorter the period during which fathers are faced with visitation arrangements, and the more they have been able to invest at normal levels in their children. Several studies confirm that there are positive effects of the child’s age at divorce on father–child relationships (Aquilino, 2006; Kalmijn, 2012).
A different type of argument is that fathers maintain ties with adult children through marriage. Mothers are generally the kinkeepers at home; for example, they make arrangements for visits, they call more often, they buy the birthday presents, and so forth. When married, fathers benefit from these investments, but when they divorce, they not only lose a spouse, they also lose a kinkeeper. As a result, outside of marriage, fathers may be less able to maintain ties with their children when the children are adult, and the fathers are living on their own. In other words, marriage and children are often a package deal for men, or to put it more positively, marriage protects men (Kalmijn, 2007; Clark and Kenney, 2010). Direct evidence for this marriage protection effect is scarce, but studies do find that widowed fathers also have less frequent contact with their children than married fathers, suggesting that even without the decline in investment opportunities after divorce, there is a negative effect (Ha et al., 2006; Kalmijn, 2007).
Most theoretical arguments have addressed the role of the father, and there is less attention for how the mother’s ties to the children are affected by a divorce. Adult children with married parents can visit their parents together or can give support to them simultaneously. For children of divorced parents, such economies of scale do not exist. Hence, unless the children of divorce decide to make extra time for their parents, one would expect that they spend less time with their father and their mother, compared to the children of married parents. A divorce may also lead to emotional problems for the parent, which in turn can affect the relationship with the adult child negatively. This applies to mothers and fathers. Studies sometimes find negative effects of divorce on relationships between adult children and mothers, especially for the degree of conflict between mother and child (Kalmijn, 2012).
The role of repartnering for intergenerational relationships has also been studied. One line of inquiry examines how the formation of new unions affects relationships with young children, either children in one’s own home (mostly for mothers who repartner), or for relations to children in the home of the ex-partner (mostly for fathers who repartner). How repartnering affects ties to adult children has been studied less frequently, but there are some studies suggesting that for fathers, repartnering may have cumulative negative effects (Clark and Kenney, 2010). This finding is in line with the notion that marriage and children are a package deal for men. When a new marriage is formed, and possibly new children are born, ties to prior children may become weaker. A more recent strand of research further suggests that relationships to adult stepchildren are somewhat weaker than relationships to adult biological children from the current marriage (van der Pas and van Tilburg, 2010). These effects also depend on the length of time the stepchild was living with the parent. Ties to stepchildren who never shared residence with the father will probably be weaker still. Moreover, it is not yet clear if divorced fathers have more contact with adult stepchildren than with the biological children they had from a prior marriage, although studies on sequential parenthood among fathers do suggest this (Manning and Smock, 2000).
An underlying assumption in the solidarity literature is that adult intergenerational relations, and especially support, are important for individual parents and children. A number of studies have explicitly tested this assumption by relating various solidarity dimensions to measures of well-being. Three theoretical perspectives have been formulated. The first, most basic theory argues in terms of the costs and benefits of support and expects positive effects of receiving support on well-being and negative effects of giving support. The second and slightly different theory argues in terms of equity by stating that the balance of support is crucial (Davey and Eggebeen, 1998). Giving too much support may lead to a decline in well-being because parents or children feel exploited, which may lead to anger and, as a consequence, reduced well-being. Receiving too much support may lead to a decline in well-being as well because it enhances feelings of guilt and helplessness. Especially this last prediction distinguishes the equity perspective from the cost–benefit perspective. The third perspective starts from notions of altruism and argues that helping others may improve well-being. Helping others in need may provide meaning in one’s life, may improve self-esteem, and may lead to normative approval of others. Rather than seeing help to others as a cost, this perspective sees helping others as a benefit.
What has the research shown for various perspectives? Some studies focus only on the main effects of giving and receiving support. These studies generally provide inconsistent results for the hypotheses. Sometimes there is no effect of receiving help on well-being, sometimes there is a positive effect (Ingersoll-Dayton, Morgan, and Antonucci, 1997), and more often, there is a negative effect on well-being (Liang, Krause, and Bennett, 2001). Those who examine equity find that older adults who receive more support than they give have the lowest level of life satisfaction (Davey and Eggebeen, 1998). Overbenefiting thus appears to be negative, seemingly in support of the equity theory.
Selection bias is a serious empirical problem in this area. Older adults who are ill or depressed may have a greater need for support. This will lead to a negative association between support received and well-being, which in turn suppresses the potential positive causal effect of support received. A different reasoning applies to the effect of overbenefiting. Older adults who are very ill will receive much support and will not be able to give support. This leads to a negative association between overbenefiting and equity, and, hence, the observed effect in cross-sections may be spurious. Solving the problem of selection bias in this literature is difficult, in part because adjustments for health problems do not help. After all, health problems are part of the dependent variable and should not be controlled for.
A better solution is to use longitudinal data and to estimate effects of changes in support exchange on changes in well-being. A longitudinal study on children’s well-being finds no effects of support exchange with parents on the well-being of adult children (Merz, Schuengel, and Schulze, 2009). Longitudinal analyses for parents in Southern California show that parents who gave instrumental support to their adult children had fewer depressive symptoms at a later age (Byers et al., 2008). Similar results were found in the European SHARE data: starting to give support to an adult child appeared to improve well-being, whereas starting to give support to an elderly parent did not improve well-being (Opree and Kalmijn, 2011). This asymmetry in well-being effects for parents and children seems in line with the generational stake theory as well as with biological altruism – parents feel good if they help their children. It is less consistent with the notion of self-esteem enhancement since children do not seem to benefit from helping parents, at least not in terms of well-being.
While the evidence for support exchange and well-being is still developing now that more and more longitudinal surveys have become available, there is clearer evidence for the consequences of other aspects of intergenerational relations. Many authors find that intergenerational relationship quality is strongly associated with well-being for both parents and children (Umberson, 1992; Merz et al., 2009). Moreover, intergenerational conflict is associated with a decline in well-being, and ambivalence is associated with poorer well-being of both generations as well (Fingerman et al., 2008). Hence, of all the dimensions of intergenerational solidarity, the affective dimension is probably the most important for well-being.
There are large differences among societies in the degree of intergenerational solidarity. Due to the development of multinational surveys in Europe, country differences have primarily been studied in the European context. The major line of comparison here is North–South, contrasting family-oriented catholic countries in Southern Europe with individualized protestant countries in the North and West. East–West differences are important as well, with earlier marriage and more intergenerational coresidence to the East of the Hajnal line than to the West. Esping-Andersen’s welfare regime typology has also played a role in these comparisons.
The main reason to compare intergenerational relationships among countries lies in institutional effects. One interpretation of institutions focuses on the policies that countries have installed to deal with intergenerational relationships, and especially policies about care for the elderly (Daatland, Herlofson, and Lima, 2011). Another interpretation focuses on the cultural climate that can emphasize or de-emphasize the solidarity of the extended family (Reher, 1998). Important elements of this value climate are norms about intergenerational coresidence, older adult’s preferences for autonomy, and beliefs about the legitimacy of state intervention in personal affairs. There is quite some variation in the strength of intergenerational norms, with stronger norms in Southern and Eastern European countries than in Northern and Western European countries (Kalmijn and Saraceno, 2008; Daatland, Herlofson, and Lima, 2011).
An important theme in this literature has been the crowding-out hypothesis, which argues that generous policies toward elderly care reduce the strength of intergenerational relationships or, in other words, that public transfers crowd out private transfers (Kunemund and Rein, 1999). The crowding-out hypothesis has been examined in a number of studies. First, there are descriptive studies that show that there is more intergenerational contact in Southern European countries than in Northern and Western European countries (Attias-Donfut, Ogg, and Wolff, 2005; Börsch-Supan, et al., 2005; Hank, 2007). For support exchange, descriptive studies have not been entirely supportive. Some comparative analyses suggest that differences in the level of support are small and less systematic (Daatland and Lowenstein, 2005; Ogg and Renaut, 2006) or even absent when controlled for compositional differences in the elderly population (Motel-Klingebiel, Tesch-Römer, and von Kondratowitz, 2005). Results have been less convincing for the 5-nation OASIS study than for the 11-nation SHARE study, suggesting that more variation among countries leads to a better test of the hypothesis.
The descriptive studies do not measure policies and thus leave open the question of the underlying causes. In this sense, they provide a weak test of the crowding-out hypothesis. Cultural values and policies are highly correlated because in democratic countries, policies often reflect underlying values. As a result, the descriptive findings can also be due to cultural differences, that is, different norms and values about family responsibilities. More convincing tests come from studies that operationalize institutional differences with macrolevel variables. In an innovative study, Brandt et al. (2009) use the share of employees working in the service sector as an indicator of policies. They find that in countries where this share is high, children provide more practical help to parents but less personal care (Brandt, Haberkern, and Szydlik, 2009). This suggests that crowding-out refers only to personal care, thereby leaving more room for other types of care, which are not negatively affected. In other words, elderly in more generous welfare states more often receive a mixture of formal and informal care than elderly in less generous welfare states (Motel-Klingebiel, Tesch-Römer, and von Kondratowitz, 2005). Kalmijn and Saraceno (2009) use norms about care to older parents, norms about coresidence, and norms about financial support to measure familialism in a cultural sense. They find that the more familialistic the values in a country, the stronger the effect of the needs of the parents on the degree to which adult children provide support (Kalmijn and Saraceno, 2008). So far, the evidence does support the importance of institutional effects, but the question remains whether these come from policies or from values. Moreover, since it also depends on the type of care we look at, the linear substitution model of crowding-out is not generally confirmed, and a model of specialization appears more fitting.
The literature on adult intergenerational relationships is both theoretically and empirically rich. Due to the aging of modern societies, the relevance of this topic will only be growing. It is difficult to summarize the main conclusions from the research; the research is still developing, and several issues remain undecided. The solidarity perspective has shown that intergenerational relationships are generally strong. There is much contact and support exchange between parents and adult children, but this is highly dependent on the setting (e.g., country differences, ethnic differences), on life course factors (e.g., divorce, remarriage, aging), and on social stratification (e.g., education). Very intensive and time-consuming forms of support of parents by adult children – that is, personal care – are not common, however; for these, the spouse or formal means of support are more dominant (Börsch-Supan et al., 2005). Moreover, parents continue to support children when parents are quite old, in contrast to the traditional exchange model in which there would be an intergenerational reversal in the flow of support exchanged during the life course. Especially for the way parents behave, there is considerable support for altruistic theories. For both parents and children, elements of exchange are also involved. Normative concerns play an important role as well, especially in the degree to which children respond to parental needs.
There is evidence that some intergenerational relationships are characterized by tension and conflict. Because this has been a more recent strand of research than the solidarity research, less is known about this more negative side of intergenerational relationships. Tension and conflict do not seem very common, but social desirability and perhaps selective nonresponse as well may lead to underestimates of such problems. What is clear is that positive and negative aspects of parent–child relationships can go hand in hand. In this sense, the literature on intergenerational ambivalence has added an element to the literature without necessarily contradicting the solidarity perspective. Conclusions about the consequences of intergenerational relationships are least certain, in part because the design to study these issues is demanding. This nonetheless appears an important area for continued study. After all, the claim that “families (still) matter” needs an empirical test, and such a test requires a renewed focus on individual consequences. Consequences can be examined for health and well-being, but outcomes for social stratification seem relevant here as well (e.g., children’s schooling, employment, income, and so forth). Links between the literature on intergenerational relationships and the literature on social stratification and mobility could be strengthened.
Several newer themes have been suggested by the literature as well. Authors have pointed out that solidarity and conflict in (im)migrant families will become increasingly important to study. This is an interesting field because of the many conflicting pressures that are going on in such families, that is, the potentially disruptive force of immigration, the difficulty of maintaining international ties, and the generation gap in traditional family norms and values (Treas and Mazumdar, 2002). There are also large ethnic differences in children’s living arrangements (Landale, Thomas, and Van Hook, 2011), and these may have important repercussions for adult intergenerational relationships. Second, grandparenting, although not a new theme, is becoming more common as more and more mothers are entering the labor force. Grandparenting is not simply another form of instrumental support that parents provide to children; the embeddedness of this support in a three-generational relationship is theoretically challenging as well (Hagestad, 2006). Third, the themes of diversity and complexity will become more important. Much is known about the effects of parental divorce, but this is only one complication in the life course that affects intergenerational ties. Due to remarriage, fertility in remarriage, and second union dissolution, parents can now have many different types of children. For example, a father can have children from a previous union (old children), children from his new and current union (new children), and nonbiological children from his current union (stepchildren). How the ties of these children will develop into adulthood is not well known. Theoretically, this is interesting because it raises important questions about the role of biology, marriage, and shared residence for understanding the development of parent–child ties (King, 2009). Finally, it can be expected that inheritances will become a more important issue in the literature. Pension systems have become more vulnerable due to aging, on the one hand, and risks in the financial sector, on the other hand. At the same time, however, current generations of elderly have accumulated much wealth, in particular in housing. As a result, there is more to pass on than before, but there is also pressure on older adults to consume their wealth when retirement benefits are lower than expected (Szydlik, 2004).