Kōtuitui

New Zealand Journal of Social Sciences Online

---

Maintaining solidarity across generations in New Zealand: support from mid-life adult to ageing parent

Sarah Hillcoat-Nallétamby

Department of Applied Social Sciences
School of Human Sciences
University of Swansea
Singleton Park , Swansea SA2 8PP
Wales, UK

Arunachalam Dharmalingam

Department of Sociology
School of Political and Social Inquiry
Monash University
Clayton, Vic 3800
Australia

Abstract This paper establishes the key factors influencing functional solidarity—the material and emotional support mid-life New Zealanders provide to an ageing parent or in-law. Using the theoretical framework of intergenerational family solidarity, empirical analysis draws on the 1997 Transactions in the Mid-Life Family survey of individuals aged 40–54. Results indicate that mature, adult children with no partner and living far from their parent are the least likely to provide emotional and in-kind help and will be more hesitant to offer emotional support to in-laws compared to their own parents. Adult children of Christian background, and possibly those who are homemakers and not employed, are more likely to provide in-kind help, but those in rural areas are less likely. Surprisingly, ill health, a large household, being employed, and continuing to provide help to a young adult child who has left home do not necessarily hinder the respondent in helping an ageing parent. The argument that intergenerational relations are increasingly subject to compromise as mid-lifers face unprecedented demands for support from other generations needs critical consideration. Further analysis should consider the cultural significance of intergenerational transactions, and how they may be influenced by the cognitive-affective dimensions of intergenerational family solidarity.

Keywords intergenerational family solidarity; ageing parents; mid-life; adult children

INTRODUCTION

Taking a micro-level perspective, the broad aim of this paper is to explore the contention advanced by several researchers that those in mid-life are increasingly likely to be at the nexus of intergenerational tensions (Brody 1990; Hamill & Goldberg 1997; Kinsella & Velkoff 2001). Their argument is premised on the idea that demographic change—particularly delayed childbearing, coupled with increased life expectancies—means that individuals in mid-life will be confronted by an unprecedented demand for support from both younger and older generations of kin. As a result, any foundations of solidarity previously underpinning these bonds will be subject to compromise or even conflict. Other scholars have, however, challenged these interpretations on the basis that they underestimate the strength of intergenerational bonds or paint an “apocalyptic” picture of the consequences of demographic change (Evandrou 2002; Agree et al. 2003; Kohli et al. 2003).

In New Zealand , the thesis of generational conflict associated with population ageing has focused on tensions that may arise if public-level transfers between generations cannot be sustained (Thomson 1989, 1993), but minimal attention has been paid to the role that private transactions between kin may play in maintaining intergenerational family solidarity. The paper redresses this imbalance by focusing on transactions of material, financial, and affective support provided by mid-life individuals to an ageing parent or parent-in-law. The analysis undertaken complements previous New Zealand research which has examined the factors influencing the likelihood of such transactions occurring between parents in mid-life and their young, adult child.

To address this aim, the paper provides an empirical analysis of the types of support that a sample of New Zealanders in the early phase of mid-life, aged between 40 and 54, have provided to one of their ageing parents or in-laws, establishing the most influential factors in this process of intergenerational transaction. Analysis also takes into account the influence that the presence of a third generation, one of the mid-life respondent’s own children, may have in structuring such transactions.

The paper outlines the theoretical background underpinning the analysis, the method, and data sources used and an overview of findings from previous research on intergenerational transactions which has drawn on the same dataset. The study population and dependent variables are described, followed by results from multivariate analysis. Overall, findings suggest that receiving in-kind or emotional help in later life from a mature adult child will be less likely if the latter has no partner and lives far away. If child and ageing parent are linked through marriage as in-laws as opposed to biological descent, emotional support is less likely. In-kind help is more likely when children are of Christian background and probably also easier for homemakers as opposed to those in employment, but appears more problematic for those in rural locations. Surprisingly, ill health, a large household, being employed, and continuing to provide help to a young adult child who has left home do not necessarily hinder the respondent in helping an ageing parent. The paper ends with a discussion of the broader theoretical significance of these findings in relation to New Zealand and international literature on -intergenerational transactions.

THEORETICAL BACKGROUND

Empirical analysis draws on the theoretical framework of micro-level intergenerational family solidarity developed and applied by Bengtson and others (Bengtson et al. 1976; Attias-Donfut 1995b; Dykstra et al. 1999; Katz et al. 2003). The framework is composed of six elements representing the types of bonds existing between parents and their adult children, and can be divided into two dimensions. The first includes the structural-behavioural elements, which are conceptualised and measured as functional solidarity (the degree of helping with daily tasks and the exchange of resources), structural solidarity (the opportunity structures for intergenerational relationships reflected in co-residence and geographic proximity of other generations), and associational solidarity (the frequency and patterns of contact and interaction). The second dimension groups the cognitive-affective elements including affectual solidarity (feelings of emotional closeness or affinity and degree of reciprocity of such sentiments between generations), consensual solidarity (degree of shared opinions, attitudes, and beliefs between generations), and normative solidarity (degree of commitment to assuming responsibility and obligations for other family members, degree of familism). Testing of the model has concluded that intergenerational family solidarity is not a uni-dimensional construct, and that each element may be tested separately, or used to elaborate typologies of solidarity (Mangen et al. 1988; Silverstein & Bengtson 1997). More recently, the concept of ambivalence has been developed in order to address criticisms that the solidarity framework posits a dualistic approach to micro-level relations characterised by either conflict or solidarity (Lüscher 2004).

The focus of this paper is on the structural-behavioural dimension of solidarity, with functional solidarity as the dependent variable, that is, the likelihood of an ageing parent receiving material (or in-kind), emotional, or financial support from their mature, mid-life child.

DATA AND METHOD

Data are from the 1997 New Zealand sample survey Transactions in the Mid-Life Family (Koopman-Boyden et al. 2000), which provides information for 750 men and women aged between 40 and 54, including some characteristics of all their surviving children and ageing parents and parents-in-law. The sample was identified by area stratification according to population size, yielding a final response rate of 54%. Although the survey represents a unique, national-level source of unit-record data on family transactions, it does have its limitations. For example, no data are available for the ageing parent’s marital status, age, or ethnicity nor for the child’s family or employment situations, and only three dimensions of the solidarity framework can be measured: association, structure, and function. In addition, the mid-life individual has acted as a proxy respondent for both the ageing parent and child populations, which suggests that data may reflect reporting discrepancies (Shapiro 2004). Finally, no over-sampling by ethnicity was carried out, and this may explain the lack of statistical significance of some of the results presented for this variable.¹

Multivariate logistic regression techniques are used to estimate the relationship between transactions of material, emotional, and financial support from mid-life child to ageing parent, and selected covariates for all three generations. The models obtained show how the probability of being in a particular category or group, versus the likelihood of being in another, changes when independent variables are introduced (Tabachnick & Fidell 1996).

In a second stage of analysis, an additional variable was introduced in the models as an index in order to try and capture the effect of the normative dimension of the solidarity framework on the results initially obtained. The index, referred to in the discussion section of the paper as the “family values” variable, was constructed based on four variables originating from a smaller subset of the original Transactions survey which measured the degree of respondents’ agreement on the importance of family life. Questions were phrased as: “Many of the important things that happen to me involve family”, “A lot of my interests are centred around my family”, “I am very much involved personally in my family”, and “To me, family is a large part of who I am”, with responses graded on 1 = strongly agree to 7 = strongly disagree. Taken as a proxy indication of the importance respondents attach to family life, these data provide the only variable which goes some way towards capturing the cognitive-affective dimensions of the solidarity model.

PREVIOUS FINDINGS ON INTERGENERATIONAL TRANSACTIONS

Previous New Zealand research using the Transactions dataset has examined the factors most likely to influence the structural, associational, and functional bonds of solidarity between mid-life parent and one of their young, adult children. Focusing first on structural solidarity (the likelihood of co-residence between parent and child) and then on associational solidarity (the frequency of contact between parent and child when they no longer co-reside), findings suggest that parent and child gender and ethnicity, as well as a parent’s religious background, significantly influence both dimensions of solidarity (Hillcoat-Nallétamby et al. 1998). Although daughters are less likely than sons to be living at home, mothers and daughters keep in touch once living apart more than their male counterparts. Pacific Island and Asian parents and those of Hindu or Muslim religious backgrounds are more likely than other ethnic and religious groups to have young adult children living with them, but once they leave home, children of Muslim or Hindu parents seem much less likely than those from other religious backgrounds to keep in touch. Surprisingly, results suggest that young Maori adults are less likely to keep in regular contact with their parents when living independently, compared to their European/Pakeha counterparts. Further analysis showed that neither dimension of solidarity was weakened as might be anticipated when the kin network included at least one of the respondent’s ageing parents or in-laws, or when household income and the mid-life respondents’ employment status were considered, three factors which could potentially represent financial and time constraints to intergenerational co-residence or contact. These findings suggest that both independent living and communication for the young generation reflect the effects of a natural transition to adulthood, to some extent unconstrained by parental resources (Hillcoat-Nallétamby et al. 2006).

This analysis was further extended to include functional solidarity as a dependent variable—that is, the giving of material, financial, and emotional support by a mid-life parent to a non co-resident young, adult child. Findings show that financial and emotional transactions decline with increasing child age, but that emotional support is more likely when the mid-lifer is not the child’s biological parent. Mothers are significantly more likely than fathers to provide emotional and in-kind help, and children whose parents are separated benefit the least from all types of help. Infrequent contact (association) reduces emotional and financial transfers, and increased geographic distance reduces the likelihood of material and emotional transfers. The presence of a third generation (ageing parent/grandparent) appears to reduce the likelihood of a young, adult child receiving in-kind help, but they will benefit more from emotional help if at least one member of the grandparent generation is alive (Hillcoat-Nallétamby & Dharmalingam 2003).

Further descriptive analysis comparing the types of help the mid-lifer provides to adult child and ageing parent, respectively, indicates an asymmetry in terms of the nature of support provided. Compared to their grandparents, the youngest generation receives more financial and material support but both generations benefit from emotional support (Hillcoat-Nallétamby & Dharmalingam 2002).

Finally, when considering how the mid-life parent may alter the support provided to their adult child when certain characteristics of their own ageing parent are taken into account, it was found that the child’s chances of benefiting from all types of help are increased when their grandparent receives emotional support, and that these chances remain unaffected regardless of the total number of surviving grandparents they have in their kin network (Hillcoat-Nallétamby & Dharmalingam 2006).

At a broader level, initial interpretations of these findings suggest: (1) preliminary evidence of the asymmetrical nature of intergenerational transactions which reflect a more pronounced response from parents to the life-course needs of youth than to those of ageing parents; (2) that such an asymmetry can potentially offset the conflicting constraints that mid-life individuals are anticipated to experience in a three-generation context; and (3) that transactions are patterned by gender, suggesting that women in New Zealand continue to play a key role in the maintenance of micro-level bonds of solidarity between generations.

STUDY POPULATION AND DEPENDENT VARIABLE

The study population included all mid-life respondents who had at least one parent or parent-in-law alive but with whom they did not cohabit. If more than one parent or in-law was alive, a focal individual was selected based on the number of types of assistance received or the frequency of contact in the absence of such transactions (Lee et al. 1994). This gives a total study population of 584 individuals. Analysis also includes data for the respondent’s children aged 15 or more, who were selected using the same criteria. Of all respondents, 287 had a child matching these criteria and the remaining missing values in the dataset were recoded to correspond to a category of “no focal child”.²   The purpose of this was to assess whether the presence of a third, younger generation influences the mid-life adult’s transactions with their ageing parent.

Analysis for the study population included several explanatory variables: age, gender, ethnicity, residence, religion, marital status, educational achievement, household size, employment status, age at first birth, and long-term health problems (see Table 1 for univariate description of sample characteristics). About one-quarter of the respondents are in their early 50s, and the majority in a union, and for those who are not, most have experienced separation or divorce. Almost 30% experience a long-term health problem or illness. Predominantly of Pakeha ethnicity (Maori and other ethnic groups representing less than 15%), of Christian religions (about 70%), and city or town dwellers, respondents are on the whole involved in some form of paid employment, earning less than $50,000 per year, and just under one-fifth have never completed any educational qualification. About 7% have no children and a similar proportion only one, but over 1 in 5 have four offspring or more, and about one-half live in households of four or more people.

Table 1 Univariate distribution of explanatory variables (n = 584).

		%	n
Respondent
Gender	Male	49.8	248
	Female	50.2	336
Age	40–44	41.7	238
	45–49	34.6	213
	50–54	23.7	133
Marital status	Single/widowed	3.0	51
	Divorced/separated	14.3	94
	Married/de facto	82.7	439
Employment	Full-time	47.4	260
	Part-time	11.1	103
	Self-employed	25.5	121	Includes family business not paid
	Homemaker	12.6	54	Includes students, retired, voluntary
	Unemployed	3.4	46
Residence	City	65.4	387
	Town	13.9	93
	Rural	20.7	104
Ethnicity	Maori	7.5	26
	Pakeha	86.1	517
	Other	6.4	41	Includes Asian and Pacific Island groups
Religion	None	25.8	148
	Christian	58.0	328
	Other Christian	11.9	74
	Non-christian	4.2	34	Includes don’t know
Educational qualification	None	18.6	101
	Secondary	42.9	255
	Tertiary	24.0	140
	Bachelor/PostDoc	14.6	88
Total children	0	7.2	69	Includes born, adopted, step, foster
	1	7.1	44
	2	37.1	205
	3	27.2	153
	4+	21.4	113
Age 1st birthday	<21 years	17.2	100
	21+	74.8	410
	No live birth	8.0	74
Focal child	Yes	51.6	287	Child aged 15+ not living at home
Has health problem	Yes	28.7	179	Long-term health problem or illness lasting 6 or more months.
Number in household	1	4.1	43
	2	26.4	155
	3	20.1	119
	4	26.1	145
	5+	23.2	122
Personal income	<$30,000	39.9	263
	30,001–50,000	27.5	148
	50,001+	19.5	103
	Don’t know	13.1	71
Child exchanges	No focal child	51.6	287
In-kind (y/n?)	Receives	14.2	83
	Does not receive	34.2	204
Emotional (y/n?)	No focal child	51.6	287
	Receives	26.7	156
	Does not receive	21.7	131
Ageing parent
Gender	Male	27.4	155
	Female	72.5	429
Health problem?	Yes	47.5	282
Number in ascending generation	1	28.7	183	Total number of parents or in-law alive—based on respondent’s current union only
	2	34.1	205
	3	20.8	114
	4	16.4	82
Kin relationship	Mother	60.5	366
	Father	23.3	130
	Mother-in-law	12.2	63
	Father-in-law	4.1	25
Distance (km)	≤3	12.7	85
	3–20	29.4	170
	21–100	12.2	78
	100+
Child (focal child not living with respondent and aged 15+; n = 287)
Gender	Male	46.2	130
	Female	53.8	157
Age	<25	64.2	182
	≥25	35.8	105

 

 

The dataset provides four variables for the ageing parent group: geographic distance from children, kinship relationship to child (biological or in-law), health status, and the number of parents and in-laws alive (see Table 1). Most ageing parents are female,³  nearly one-half have a health problem, and a similar proportion live at least 100 km from their adult child. About 1 in 3 represent the respondent’s only surviving parent or parent-in-law but 6 out of 10 are related through a maternal kinship bond.

Of the child population, slightly more than one-half are female but only about one-third have reached their mid 20s. For the purposes of multivariate analysis, only one variable was used as an explanatory factor, a functional solidarity variable indicating whether they have benefited from material help from their parent over the last 12 months. Two reasons underpin this decision. First, to include receipt of emotional support as a child characteristic would suggest a hypothetical link between this factor and the likelihood of an ageing parent receiving in-kind help, an association that it is difficult to substantiate from a theoretical perspective. Second, it is assumed that the functional solidarity variable indicating whether the young adult has received in-kind help from their parent over the last 12 months will capture the influence of other child characteristics of age and gender.

DEPENDENT VARIABLE

The types of help the mid-life respondent has reported providing to their ageing parent at least once within the last 12 months provide the basis for the development of the dependent variable. Responses were first regrouped into three categories to represent an indicator of functional solidarity: emotional support provided; financial support provided; in-kind or material⁴  support provided. Each category is treated as dichotomous (1 = respondent has provided emotional support to their ageing parent, 0 = has not provided this support,⁵ etc.). Individuals can receive more than one type of assistance, so analysis is limited to whether giving at least one type of assistance was reported. As less than 1 out of 10 adult children actually reported providing their ageing parent with direct financial help, these cases (n = 40) were merged with the material support responses, to provide two dependent variables—in-kind and emotional support.

Initial descriptive results indicate that about one-third of respondents (35%) say that they have not provided any help to their parent over the past year,⁶ and of those who have, similar proportions have provided in-kind or emotional support (37 and 39%, respectively). The most pressing types of in-kind assistance which appear to be needed are those involving an activity outside the home—transport, house maintenance, gardening, and shopping—but few require help in the home, either with housework or personal health.

MULTIVARIATE RESULTS

Multivariate results are presented in Table 2. Model 1 assesses the factors influencing the probability of the respondent providing in-kind support to their ageing parent as opposed to not providing it. Model 2 assesses the influence of these factors on the likelihood of providing emotional support. When interpreting these results, the reader should bear in mind that due to small cell sizes, the category “other” of the original ethnicity variable has been collapsed with “Maori” (hence the “non-Pakeha” category in Table 2).⁷

IN-KIND SUPPORT TO AGEING PARENTS

The likelihood of an ageing parent or parent-in-law receiving in-kind help is influenced by how close they live to their adult child (or daughter/son-in-law), whether they have a health problem, and how many individuals of the ascending generation are present (co-surviving). In addition, this type of exchange varies depending upon their adult child’s residential location, religious background, marital and employment status, the age at which they had their first child, whether they have a health problem, and the number of people with whom they co-reside. It also varies quite significantly depending upon whether they have a grandchild who continues to receive in-kind help from their parents once they have left home or not.

As might be expected, whether an ageing parent receives help will reflect how far away they live from their offspring; those living nearby, within 20 km of their adult child, for example, will be much more likely (over five times in fact) to receive in-kind help than if they lived over 100 km from them (Table 2, Model 1, section “Ageing Parent”, odds ratio: 5.16). Unsurprisingly, having a long-term health problem means that, in later life, a parent can expect to receive more material help from their child than would otherwise be the case if they were in good health; the odds are increased by about 50% because of poor health status (Table 2, Model 1, odds ratio: 1.46). The fewer members of the ascending generation there are (the parents or parents-in-laws generation), the more likely the mid-life respondent will be to provide them with in-kind help. Interestingly, and although results are not statistically significant, the adult child does not appear to make a distinction between mother or mother-in-law when providing in-kind help, but may be less inclined to offer it to their father or father-in-law, a finding which suggests the predominance of gender over kinship bond.

Table 2 Logistic regression models—covariates expressed as odds ratios (1 = yes, 0 = no). Effects of selected characteristics on the likelihood of a mid-life respondent providing in-kind or emotional support to their ageing parent (n = 584).

		Model 1	Model 2
		In-kind	Emotional
Ageing parent
Distance (km)	100+	1.00	1.00
	≤3–20	5.16***	2.16***
	21–100	7.65***	1.60
Kinship	Mother	1.00	1.00
	Father	0.67	0.80
	Mother-in-law	1.16	0.41**
	Father-in-law	0.53	0.14***
Health problem	No	1.00	1.00
	Yes	1.46*	0.97
Number in ascending generation	1	1.00	1.00
	2	1.01	0.70
	3–4	0.54**	0.66
Mid-life respondent
Gender	Male	1.00	1.00
	Female	1.25	1.32
Age at survey	40–44	1.00	1.00
	45–49	1.10	0.83
	50–54	0.72	0.75
Ethnicity	Non-Pakeha	1.00	1.00
	Pakeha	0.91	1.21
Residence	City	1.00	1.00
	Town	0.77	0.99
	Rural	0.56*	1.07
Religion	None	1.00	1.00
	Christian	1.65*	1.10
	Other Christian	1.93*	1.18
	Non-Christian	1.32	1.03
Marital status	Married/de facto	1.00	1.00
	Single/widowed	0.44*	0.23***
	Divorced/separated	0.50*	0.43**
Educational achievement	None	1.00	1.00
	Secondary	0.76	2.86***
	Tertiary	0.89	3.16***
	Bachelors	1.66	4.44***
Number in household	1–2	1.00	1.00
	3	0.55*	1.03
	4	0.58	0.63
	5+	0.38**	1.04
Employment status	Full-time	1.00	1.00
	Part-time	0.70	0.94
	Self-employed	0.99	1.20
	Homemaker	1.58	1.00
	Unemployed	0.37***	1.77
Age at first birth	< 21	1.00	1.00
	21+	2.01**	0.90
	No child	2.91*	0.86
Health problem	No	1.00	1.00
	Yes	2.01***	1.56*
Respondent’s child
Receives in-kind help	Does not receive	1.00	1.00
	Yes, receives	2.90***	5.98***
	No focal child	2.15**	4.52***
Log likelihood		–334	–319
Wald χ2		78	107
P-value		.0001	.0001

*** P < 0.01; ** P < 0.05; * P < 0.10.  Reference category = 1.00.

 

When an ageing parent has a child living in a rural area, or who is not in a partnership because of divorce, separation, or widowhood and who shares their household with several people (e.g., with five or more), their chances of benefiting from material support will probably decrease (Table 2, Model 1, section “Respondent”). Adult children who are of Christian background will also be more likely to offer help compared to those with no religion (65 and 93% more likely). In terms of employment status, it may be that only adult children who are homemakers will be more likely to give in-kind support, presumably because of more free time (but the relationship is not statistically significant), whereas the unemployed are clearly much less likely than those employed full time to have provided it. The age at which the adult child has had a first birth influences this form of functional solidarity quite significantly, particularly amongst those who had a child after the age of 20. Surprisingly, parents or in-laws are more likely to have received in-kind support from an adult child who is experiencing long-term health problems.

The last explanatory factor included in Model 1 has been designed to capture the effect that the presence of a third generation, a grandchild (a focal child), might have on the likelihood of the respondent’s ageing parent receiving in-kind help (Table 2, Model 1, section “Respondent’s Child”). A focal child, as noted earlier, is a child aged 15 or more who does not live with the respondent, that is, their mid-life parent.⁸ For respondents who do have a focal child, the variable reflects whether this child received in-kind support from them. Interestingly, an ageing parent is over two times more likely to receive in-kind support when the respondent has no focal child, or if they do, when this child also receives in-kind support (odds ratios 2.15 and 2.90). The significance of this result is the object of further commentary in the Discussion section of the paper.

EMOTIONAL SUPPORT TO PARENTS

Logistic regression results show that a small number of explanatory variables had a significant effect on the likelihood of an ageing parent receiving emotional support. For the ageing parent, these include residential proximity and the type of kinship bond they share with the mid-life respondent (Table 2, Model 2), and for the mid-life respondent, the key variables were their marital and health status and educational achievement. As was the case for in-kind help, receipt of emotional help is also associated with the presence of a focal child.

The residential proximity of ageing parent to adult child does influence this type of transaction, but only when distances separating them are small: the closer they live to each other, the more likely the ageing parent will be to benefit from emotional support (when living within 20 km of each other, parents are more than twice as likely as those living over a 100 km from their child to receive it). The kinship bond between ageing parent and adult child proves more important than an ageing parent’s gender: both father- and mother-in-law are much less likely to receive emotional support than the respondent’s own parents, but for in-laws, this difference is more pronounced for males. Parental health status and the total number of parents alive do not make any difference to the likelihood of receiving emotional support.

Of all the characteristics of the adult child, only three were of significance: marital status, educational achievement, and health status. An ageing parent whose adult child had no partner was much less likely to receive emotional support compared to those who did. This support was more forthcoming from adult children the higher their level of educational achievement, and proves to be a factor of significant import: when respondents had achieved at least bachelor level qualifications, for example, they were about four and a half times more likely to provide emotional support compared to those who had no educational qualification. As was the case for in-kind support, adult children with a health problem are more likely (more than 50%) to give emotional support to their parents.

Finally, the relationship between the focal child and the ageing parent’s receipt of emotional support is examined. Respondents with no focal child, and those with one who receives in-kind help, are four and a half times to six times more likely to provide emotional support to an ageing parent than those who have a focal child to whom they give no in-kind help. As an initial interpretation of these findings which are elaborated further in the Discussion section, what appears to matter here is not the presence or absence of a focal child per se, but rather whether this child actually benefits from support from their own parent.

DISCUSSION

Informed by research developed in the field of intergenerational family solidarity, the purpose of this paper was to establish the most influential factors affecting functional solidarity—the likelihood of an adult child in mid-life providing material/in-kind, emotional or financial support to an ageing parent or in-law. This investigation was grounded on the postulate that these transactions would be structured in particular by gender, and that any asymmetry in terms of the types of help provided to ageing parent and young, adult child could potentially offset the conflicting constraints that mid-life individuals are anticipated to experience in a three-generation context. Underpinning this work has been an investigation of the broader debate surrounding the mid-life period as one in which intergenerational bonds of solidarity may be increasingly compromised as those in this period of the life course confront the potentially competing support or care needs of older and younger generations.

The remainder of the discussion will focus on the key factors influencing functional -solidarity, in-kind and emotional support, highlighting the broader theoretical significance of these findings when appropriate.

Factors influencing functional solidarity

We have found that it is the ageing parent’s gender which will exert some influence on functional solidarity and may increase their chances of receiving in-kind help in their roles as mother or mother-in-law. The mid-life respondent’s ethnicity and age do not affect the likelihood that they will continue to keep links of functional solidarity with their ageing parent, but religious background will be a factor to consider when it comes to the continuity of in-kind help.

The most striking result is that the more successful an adult child has been in pursuing educational qualifications, the more an ageing parent can expect to receive emotional support from them. Interestingly, these results mirror those found for the mid-life parent who provides emotional support to their young, adult child who has left home (Hillcoat-Nallétamby & Dharmalingam 2003). For the results of the present analysis then, could it be that if higher socio-economic status contributes to increased geographic mobility, then adult children may be more likely to substitute in-kind help which requires their physical presence, with emotional help which is not necessarily mediated by proximity? Some support for this interpretation comes from findings on the influence of social class (and income) and its association with geographic mobility. Rossi & Rossi (1990: 434) have found that upward mobile adult children (low-income parents and high income children) engage less in exchanges of help than downwardly mobile children (high-income parents, low-income children). In particular, low-income children provide more in-kind help (chores) than high-income children. Silverstein & Bengtson (1997: 450) found that higher income is associated with adult children being more likely to have “intimate but distant” relations with fathers (strong emotional closeness but not geographic proximity, contact, providing help or receiving help), and this is consistent with the greater geographic dispersion and a lower affiliation to family found amongst higher social classes.

Some of the factors examined influence both dimensions of functional solidarity. When an adult child has no partner, for example, their ability or willingness to engage in assisting their parent with either emotional or in-kind help is certainly reduced, suggesting that their inaction reflects some constraint, such as time. This said, neither geographic distance (unless they live a considerable distance from their parent—more than 100 km) nor a personal, long-term health condition appear to be associated with the provision of either type of support. It also seems that an adult child will be far more likely to maintain these bonds of solidarity with their ageing parent provided all their children who are aged 15 or more are still living at home (Table 2, category “no focal child”⁹ of respondent’s child variable). In addition, when respondents do have one of their children in this age range who has left home (categories “does not receive” and “yes, receives”), grandparents are much more likely (nearly six times in fact) to benefit from emotional support when the grandchild continues to receive in-kind support from home compared to one who does not.

The first two results concur with international findings on the effect of adult children’s marital status in influencing contact or provision of support to ageing parents (Cicirelli 1983, 1984; Attias-Donfut 1995a), and on the determining effect of residential proximity in regulating the exchange of resources (Hoyert 1991; Arrondel & Masson 2001), but the latter two findings require further interpretation. The more intriguing of findings is that older parents’ receipt of emotional and in-kind help is somehow tied to the likelihood of a grandchild also receiving the latter type of support or of there being no grandchild (i.e., no grandchild who has already reached aged 15 and left home, the focal child in this study). Given that these results are net of the influence of all the other individual and group level factors, further explanation probably comes from the factors which were not included in the models.

The first of these is the lack of data on the ageing parent’s age. Rossi & Rossi (1990) have identified the effect of life course changes in influencing the flow of support provided across the parent-adult child dyad. They find, for example, that an increasing number of adult children provide money to their mothers as the latter age, or that daughters provide increasingly more help to mothers with domestic chores as their parents age.

Another explanation could be that a key interpretative element is missing because there are no data on the normative, affective or consensual aspects of intergenerational relations. Others have found, for example, that with higher affective solidarity, comes greater associative solidarity (Roberts & Bengtson 1990) and, in turn, a greater likelihood of the evidence of functional solidarity, particularly financial exchanges (Rossi & Rossi 1990). Shuey & Hardy (2003), in their study of family allocation decisions regarding intergenerational transfers of assistance between couples and their ageing parent, find that those providing financial help to their children were also more inclined to provide help to their ageing parent or in-law. They point to the importance of considering unobserved but underlying dispositions that may influence intergenerational transfers because they represent a trait of generosity underpinning supportive behaviour in certain families.

More recent work focusing on classificatory approaches identifying types of family intergenerational relationships provides a further avenue of explanation for differences in family relationships, in terms of transfers of support between adult children and their parents (Berkman et al. 1991; Silverstein & Litwak 1993; Pyke & Bengtson 1996; Silverstein & Bengtson 1997; Burholt & Wenger 1998). Silverstein & Bengtson’s (1997) typology of family types, for example, includes “detached” relations which are qualified by a lack of engagement between adult child and parent on any of the solidarity elements of association, affect or geographic proximity.

Could the respondents in the New Zealand sample who give less readily to their ageing parent when also not providing in-kind support to their own child (the focal child who receives no in-kind help) be displaying a certain type of family relationship or “culture” with regard to exchanges, governed perhaps by norms, values, or attitudes towards their relations with family members? Hung and colleagues’ research in New Zealand, for example, on notions of obligation felt by younger European and Chinese family members to family elders, were able to identify family types through the combination of responses regarding feelings of filial obligation (financial support, respect, obedience, maintaining contact, etc.). They found that divergence with regard to filial obligations is characteristic of parent-child bonds for European families but that convergence is the norm for the Chinese dyads observed (Ng et al. 2000).

To explore this idea further, a “family values” index, previously described in the Data and Method section of the paper, was constructed. When included in the regression model for in-kind help, the coefficient for the variable “child receives in-kind help” became statistically insignificant but the new variable had an estimated odds ratio of 1.47 with a P-value of 0.12.

The interpretation offered of these results is t