In European and Anglophone societies, coresidence of adult children with older parents is not very common, though several countries have documented upward trends in recent years. In the US 18% of all households include at least two adult generations. In EU countries extended family household prevalence ranges from 0.1% of all households in the Netherlands and Denmark to 20% of all households in Bulgaria. Former Eastern European countries and to a lesser extent Southern European countries are characterised by particularly high rates of multigenerational coresidence in Europe (Iacovou & Skew, 2011; Kalmijn & Saraceno, 2008).
East Asian countries are generally characterised by higher levels of intergenerational coresidence compared to European and Anglophone societies. According to a recent UN report, 33% of those aged 60 or older lived independently (alone or with a spouse) in China in 2000; 35% in Hong Kong in 2011; 49% in South Korea in 2005; and 51% in Japan in 2010. The rest resided with their children. By comparison, in 2011 in Finland and France 87% of those aged 60 or over lived alone or with a spouse; in Germany, the figure was 90%, in Italy 70%, in Spain 58%, in Norway and the UK 86%, in the US 71% and Canada 74% (Department of Economic & Social Affairs, 2017). Although the prevalence of coresidence has been falling in Japan over time, it is still a common practice even among the young older adults aged 65–74. Importantly, this decline in coresidence has been accompanied by a steady increase in the prevalence of residential proximity to parents among recently married men and women in Japan (Kato, 2013; Wang & Raymo, 2020). Several generations of one family residing nearby are likely to support each other with unpaid work in ways somewhat similar to residing households (e.g. Chen et al, 2000; Raymo et al., 2010).
How people spend time is influenced by their household compositions. Larger households may have greater potential for economies of scale when it comes to unpaid work. Some members of the household may specialise in unpaid work to reduce the domestic work burden for the rest of the family, especially once new families are formed by adult children. Transitions into marriage (or a stable partnership) and parenthood are typically associated with dramatic changes in domestic workload and a rise in conflict between paid and unpaid work (Borra et al., 2020; Kan, 2009; Zhou & Kan, 2019). Moreover, sometimes several generations may choose to live together because a family member needs care (Takeda et al., 2004). In this scenario, multigenerational living may be associated with an increased need for unpaid work by other household members.
While household composition influences the amount of required unpaid work, gender plays a major role in the way this work is shared. Marriage and children are always associated with the rise in time spent on housework for women (see Sayer, 2010 for a summary of evidence from Western countries; for evidence from East Asia see Sechiyama, 2013). There is a consistent negative association between motherhood and women’s employment (Ahn & Mira, 2002; Miller, 2011), testifying to the rise in intensity in the work-family conflict for women once they have children. For men, marriage and parenthood have historically been associated with decreases in time spent on housework, but this has changed in recent years. In many countries today, marriage is associated with increases in housework time for both women and men, although the change for women is larger while the effects for men tend to be modest. Having children leads to a rise in time spent on care work for both men and women (for a more detailed summary and references see Sayer, 2010). The division of labour in multigenerational households, therefore, is of particular interest in situations where there are children, and we compare families with and without underage children. This paper will focus on families of married adult children only, as these are individuals who have completed the transition to adulthood having formed their own families. We believe the dynamics associated with coresidence for these children are very different from those of unmarried adult children, who need to be analysed separately Fig. 1.
This study contributes a fuller understanding of variation in three major domains of time use (paid work, housework and childcare, and older-adult care) by the types of coresidence arrangements.
Domestic Division of Labour Theories and Intergenerational Support in Multigenerational Households
Our conceptual approach will examine predictions about households where married adult children reside with their older parents from the intergenerational solidarity perspective and the gender perspective.
The intergenerational solidarity theory is specifically designed to make sense of intergenerational relations. In this framework, coresidence with one’s older parents or in-laws results in solidarity between older parents and their adult children(-in-law). This paper focuses on the instrumental aspect of the theory, which refers to non-financial mutual help between generations (e.g., cooking, cleaning, fetching groceries, providing care). According to this theory, we would expect that in households with more than one generation of adults, they will share unpaid work similarly to the ways they share financial and other resources (Shirahase & Raymo, 2014; Takagi & Silverstein, 2011). This approach predicts that the older generation will support their adult children as long as their health allows this. Once the older adults need help themselves, intergenerational solidarity will encourage their working-age children to care for them.
Mutual support between generations of one family is well documented. Parents provide childcare and household chores assistance to adult children who have become parents (Bucx, van Wel, & Knijn 2012; Yamato, 2017). Adult children also support their parents as they age, providing financial support, help with housework, care work and emotional support (Wu, 2021). Scholars have referred to this as “linked lives” to describe the reciprocal exchanges between adult children and their parents (e.g., Greenfield & Marks, 2006).
The intergenerational solidarity approach does not make gender-specific predictions, but empirical research reports gendered patterns in the intergenerational exchange. Analysing longitudinal data from the US, Silverstein et al. (2002) find that mothers on average provide more practical support including domestic help to adult children than fathers. Daughters also provide more support to parents than sons in the US (Silverstein et al., 2002) and the Netherlands (Kalmijin 2007). Similarly, in Canada daughters tend to provide more care than sons and the difference is particularly pronounced in traditionally female (and more time-consuming) tasks such as helping with personal care and domestic assistance (Campbell & Martin-Matthews, 2003). These patterns of intergenerational exchange in support indicate that gender is layered throughout the expressions of intergenerational solidarity. Consequently, in this paper, we propose to combine the intergenerational solidarity approach with the gender-centred perspective.
The gender-centred approach asserts that women perform more unpaid work and less of paid work compared to men because normative femininity and masculinity are associated with domestic work and paid work respectively (Berk, 1985; Brines, 1994; DeVault, 1990; Ferree, 1990; South & Spitze, 1994). By doing housework and caring for family members, women fulfil gendered social expectations and hence “do gender” (Brines, 19C94). Men in turn do not need to engage in unpaid work to establish their masculinity and in cases where their masculinity is damaged through loss of paid work, for example, they can withdraw from unpaid work to prevent further loss in masculinity (Brines, 1994). The gender-centred approach is rooted in marital-dyad assumptions and provides no indication as to how having two generations of adults in the household may influence the working-age couples’ time use (Geist & Ruppanner, 2018). However, if we extrapolate its assumptions from the marital dyad to a larger coresident household it suggests that older parents’ domestic help will not be necessarily a direct substitution of their children’s domestic work time because both older parents and adult children need to “do gender” in domestic work. An older mother or mother-in-law would assume some unpaid work responsibilities within the household as long as her health permits it to address her need to be feminine. For older adults, coresident fathers and fathers-in-law performing masculinity would mean avoiding unpaid work. Similarly, by the logic of doing gender, working-age men are not compelled to do unpaid work. Doing too much unpaid work could even damage their masculinity. Assuming that women are evaluated based on the result of their unpaid labour (e.g., clean home as found by Thébaud et al., 2019) rather than on the length of time they spend to achieve that result, multigenerational households have a clear potential for economies of scale through sharing the unpaid labour. According to the gender-centred approach, domestic work-sharing will largely happen between adult women and consequently the economies of scale will also mostly benefit women. Coresidence with older men is expected to increase adult women’s time spent on cooking, laundry, and cleaning associated with an additional adult in a household and these men are not expected to provide much help around the house themselves.
Before moving to discuss relevant empirical findings, we need to say a few words about the resource approach. Together with the gender-centred approach, this is a key theoretical approach widely used to make sense of paid and unpaid work participation within couples. It is well established that individual choices to allocate time to paid and unpaid work at least partially stem from resource-related factors, in which earning power and lack of time availability allow individuals to bargain away or outsource unpaid work to focus more on paid work (Gupta, 2007; Hook, 2017; Killewald & Gough, 2010). Economic approaches, however, do not appear to be equally useful for understanding dynamics in multigenerational households. While older parents and adult children are known to share resources, the older parents tend to be outside the labour market and hence do not have a similar incentive to focus on paid work and may not suffer from comparable time shortages. Consequently, while access to economic resources influences the ways working-age adults share paid and unpaid labour within couples, resource frameworks are less useful for understanding time allocation dynamics in multigenerational households. In this paper, we will control for resource availability, but we will not base our core hypotheses on the theories focusing on resources.
Empirical Studies on Coresidence with Older Parents
In gender unequal societies, such as China, Japan, and Italy, women living in households with older parents are more likely to stay in paid work (Maurer-Fazio et al., 2011; Sasaki, 2002; Shen et al., 2016; Ta et al., 2018), but see (Yang et al., 2015; Yu & Xie, 2018) for contradicting evidence on China.
In another testimony to reductions in work-family conflict, living with husband’s parents is positively associated with the likelihood of having the first child in Taiwan (Chi & Hsin, 1996; Tsay & Chu, 2005), the second birth in South Korea (Yoon, 2017), and the first and second births in Japan (Fukukawa, 2013), but see some conflicting evidence for Japan and Italy (Raymo et al., 2010).
These studies are largely consistent with the intergenerational solidarity approach in which grandparents share unpaid work with their working-age children and children-in-law, making their domestic load more manageable and compatible with paid work. Existing research largely focuses on women, which is not surprising given that their domestic workload is heavier than men’s across the world (Gershuny & Kan, 2012; Kan & He, 2018; Kan et al., 2021; Sullivan et al., 2018). Women’s increased ability to maintain their attachment to the labour market and couples’ increased ability to have children suggest the flow of support goes from the older adults to the working-age generation. It is however possible that women’s greater ability to participate in the labour market and increased fertility are explained through reasons other than their sharing of domestic burden. If poorer households are more commonly choosing intergenerational living arrangements, then women’s greater likelihood of employment may stem from their greater need to work rather than from grandparental support (Raymo et al., 2010 find some support for this interpretation in Italy and Japan). Couples’ increased fertility in coresident households may be explained by the older adults’ greater power when it comes to family decision-making, rather than help they provide with unpaid work. There is little existing literature verifying whether living with older parents can reduce housework and childcare time and no explicit analysis making predictions about the way gender is (or isn’t) associated with the intergenerational exchange of unpaid labour. Exploring the dynamics in sharing domestic work by type of domestic work will lead to a better understanding of the associations between coresidence and paid work participation.
Finally, previous quantitative studies exclusively focused on women’s employment patterns and how these are associated with coresidence patterns, mentioning unpaid work only as a potential reason behind women’s paid work patterns, rather than analysing it directly. This assumes that the type of coresidence largely matters for women but not men and that intergenerational solidarity, when it happens, is only performed between different generations of women. Such a claim seems unlikely, and the role of both genders needs to be investigated.
In this paper, we will combine the gender-centred approach and the intergenerational solidarity theory and test the hypotheses listed below. The gender approach allows us to nuance the predictions of the intergenerational solidarity theory when it comes to such highly gendered behaviours such as participation in paid and unpaid labour.
Hypothesis 1: Coresidence with older mothers or mothers-in-law is associated with a reduction in unpaid work time and increase of paid work time of working-age women and, to a lesser extent, for men. The differences between working-age men and women will be especially pronounced in couples with children.
Hypothesis 2: Coresidence with older men is associated with longer unpaid work time and shorter paid work time for women.
Hypothesis 3: When older adults require care themselves (whether they are coresident or not), working-age women, but not men, will spend more time on unpaid domestic work and have less time on paid work.
The Japanese Context
Coresidence of adult married children with older relatives has been falling in recent years in Japan, but it is still more common there than in European and Anglophone countries (Department of Economic & Social Affairs, 2017).
Recent research argues that this trend has at least partially been offset with the rise in proximate living (Kato, 2013; Wang & Raymo, 2020).
Japan is a particularly interesting context to investigate issues related to mutual support between generations. It is a highly developed and rapidly ageing country with a cultural environment where the traditional norm of filial piety – that also characterises other Asian countries – coexists with an emerging social norm of independence in old age (Takagi & Saito, 2013). On the policy level, in recent years Japan has shifted from de-familiasation policies that have been associated with a reduction in the family caring responsibilities to measures encouraging intergenerational coresidence (Izuhara, 2020).
At the same time, the rise of women’s educational attainment and employment rate has meant that working-age adults are exposed to greater work-family conflict. In 2018, Japan ranked 110th out of 149 countries in the Global Gender Gap Report, a testament to its low levels of gender equality. Japanese wives continue to be responsible for virtually all housework and care work in married couples (Hertog et al., 2021) so the work-family conflict primarily affects them. Norms about gender roles have shown limited change among Japanese men and women born after the 1950ies (Piotrowski, Yoshida, Johnson, & Wolford, 2019). Behavioural change has also been slow. Average men’s and women’s paid work times and housework times changed only marginally between 1996 and 2016.
In Japan, motherhood is still most compatible with a traditional division of labour within families. Having children is associated with a dramatic rise in women’s unpaid work time, but not men’s (Sechiyama, 2013). Many women quit full-time employment when they have children and return to the labour market only several years later, often into dead-end jobs (Brinton & Oh, 2019).
Given the broad trends described above, it is not clear what coresidence implies for the working-age population in Japan in terms of participation in paid and unpaid labour. In this paper, we focus on the roles older adults play in their married children’s lives, unpacking the flows of intergenerational support from the perspective of adult children. As the arrival of children is associated with a rise in need for both unpaid work time and household income, we analyse families with and without children separately. Time use patterns of working-age adults in families without children offers a closer reflection of behavioural preferences. Paid and unpaid work time of working-age adults in families with children documents the extent to which the grandparents can make a difference in families with particularly high paid and unpaid workloads.
Our analysis is based on data from the 2006 Japanese Survey on Time Use and Leisure Activities, the latest Japanese national time use survey available abroad. The prevalence of coresidence between parents aged 65 and older and their adult children fell from 45% of all households with at least one member aged 65 or older to 39% of such households between 2005 and 2015. This limited change masks a much larger shift in the adult children’s circumstances. In 2005 21.3% of households with at least 1 member aged 65 or older were 3-generation households, i.e. households which contained the older adults, their married adult children and grandchildren. By 2015 this figure fell to 12.3%. This trend was partially offset by the rise in coresidence between the older adults and their adult children (both married and unmarried), who do not yet have children of their own (Cabinet Office, 2017).
As fewer married working-age adults choose multigenerational coresidence in Japan, it is possible, that the ones who do self-select into such an arrangement for a particular reason. Our data only contains information of non-coresident older parents when these parents require care and therefore, we cannot explore such parents’ contribution to their adult children’s families, while they are still healthy. As neighbourhood living is on the rise in Japan, this is an important limitation on the interpretability of our findings.