Abstract
Older people play a key role in intergenerational households across the developing world, including in caring for the children of migrants. However, research on ‘translocal households’ in Africa and Asia focuses almost exclusively on households headed by working-age husbands and wives and fails to substantially incorporate older people, whether as household heads or household members. This article presents findings from Kiboga District, Uganda on intrahousehold dynamics and wellbeing within translocal households containing older people and younger migrants. A relational approach is adopted to interrogate gendered notions of dependence, independence and interdependence within these intergenerational relationships. ‘Translocal interdependencies’ are found to underpin these households and are critical to the wellbeing of older people, migrants and their children. However, households are also diverse in character, occupying a ‘spectrum of translocality’ ranging from coherent, supportive translocal households, to those with less unified and less reliable—but often highly persistent—linkages between migrants and older people. Further, older women may overstate, and older men understate, the translocal support to which they have access, with important consequences for development programming targeted at older people. Debates about ageing and development need to reflect the key role played by older people within these gendered webs of translocal interdependence, while acknowledging and seeking to address the many challenges they face.
I. Introduction
The ‘household’ is a central concept within development research, policy and programming and has well-established practical and theoretical advantages as a generalizable unit of analysis (Andersson Djurfeldt, 2021; De Haas, 2010; Muniina et al., 2014; Rabe, 2008). However, in contexts where migration is fundamental in many people’s livelihoods and relationships, the consequences of migration at the household level are far reaching, spanning productive and reproductive domains and bringing binary distinctions between ‘monolocal’ households at ‘origin’ and ‘destination’ into question. This applies to both international and internal migration, and thus to ‘householding’ at both transnational and translocal scales (Andersson Djurfeldt, 2021; Douglass, 2006; Yeoh et al., 2005).
Efforts to conceptualize spatially distributed or ‘translocal’ households in Africa have a particularly long track record, drawing on concerns about the social and economic impacts of rural–urban labour migration in Africa from at least the 1950s onwards (Englund, 2002; Potts, 2010), and more recent work demonstrating the resilience of rural–urban links and ongoing ‘fluid’ household membership across the continent (Ramisch, 2016; Rao et al., 2020). In Asia, a growing number of studies on translocal householding reflect similar concerns but often incorporate a more explicit focus on gendered outcomes within the household (Andersson Djurfeldt, 2021; Fan, 2021).
Studies of translocal households share a common interest in how household and family relationships are (re)configured through spatial separation and most profess a concern with how these processes are shaped by intergenerational relations. In practice, however, the focus is almost exclusively on households headed by ‘working-age’—often male—household heads, where husbands and wives are spatially separated. There are exceptions, such as Korzenevica and Agergaard’s (2019) focus on siblings in Nepal, but there are no studies which focus on—or substantially incorporate—older people, whether as household heads or as members of intergenerational households. This is despite the critical productive and reproductive roles older people continue to play across the developing world (Bastia et al., 2022).
Further, while some recent work on translocal households in Africa does incorporate a concern with social and emotional bonds within the household (Steinbrink and Niedenführ, 2019), earlier approaches frequently emphasized productive livelihoods and remittances to the exclusion of gendered concerns with social reproduction. This masked the care work and emotional labour involved in the maintenance of households, reinforcing distinctions between ‘dependence’ and ‘independence’ in intrahousehold and intergenerational dynamics that have been subject to substantial critique across diverse fields including feminist care ethics, ageing and disability theory (Barnes, 2012; Christensen, 2009; Ward, 2014). This reflects a wider aversion to dependence in contemporary political and policy discourse across both the Global North and South, rooted ultimately in Enlightenment ideas of the ‘autonomous individual’ (Barnes, 2012). Thus, as Ferguson notes, despite a long history of relational thinking in Africa, social policies generally focus on maximizing independence and denigrate ‘dependence’, precluding positive opportunities to construct desirable forms of dependence that respect ‘vernacular aspirations to social relationality’ (2015: 163–64). In contrast, a relational approach, in which social relations are understood as foundational to wellbeing (Gough et al., 2007; White, 2016, 2017), is better able to capture the centrality of social relationships and ‘translocal interdependencies’ to older people’s and migrants’ lives. It also resonates well with African notions of ‘ubuntu’ that emphasize the fundamentally relational nature of being and with efforts to incorporate these ideas into development thinking (Hoffmann and Metz, 2017).
A focus on these interdependencies brings to the fore reciprocities over both the short- and long-term, as well as the role of older people as both givers and receivers of care (Aboderin, 2006). At the same time, as feminist critiques of the household from the 1980s onwards showed (Harris, 1984; Wolf, 1990), it is important to avoid reifying the household by obscuring intrahousehold inequalities based on age or gender and downplaying the importance of individual agency (De Haas, 2010). Spatial mobility reconfigures caring practices for the young and the elderly in complex ways and involves substantial, gendered reproductive and emotional labour (Baldassar, 2007; Parreñas, 2009). In translocal contexts, these intrahousehold interdependencies may thus involve a greater burden of responsibility for women overall, even if, as will be shown below, they also pose particular challenges for men in the context of ageing and societal norms of masculinity (Mudege and Ezeh, 2009).
Development efforts to support older people in the African context often pay limited attention to these translocal dynamics. For example, social pensions are paid to individual older people, but are widely understood by policymakers and programme officials as de facto household transfers (Gelders and Athias, 2019; Barrientos, 2009). However, in efforts to understand how intrahousehold dynamics are affected by pension receipt, the focus is almost exclusively on the immediate, ‘monolocal’ household, despite evidence suggesting social pensions may increase the out-migration of working-age individuals living with pensioners (Hagen-Zanker and Himmelstine, 2013). At the same time, the contributions of older people are often neglected in broader development debates, even though most are still working and are active providers of care to family members (Barrientos et al., 2003; Yarris, 2017). Less focus still is placed on the importance of older people to migrants themselves, despite their key role in providing care and ‘informal social protection’ for migrants who fall sick or whose livelihoods or personal relationships break down (Locke et al., 2013; Schatz and Seeley, 2015).
In this article, I explore intrahousehold dynamics and wellbeing within translocal households in Uganda containing older people and migrants. Reflecting their complexity and contextual specificity, conceptualizations of these households are grounded in the understandings of participants rather than preconceived or bureaucratic categories. Adopting a relational approach to interrogate gendered notions of dependence and independence, I show that ‘translocal interdependencies’ underpin households and are critical to the wellbeing of both older people and migrants. However, household structures are also diverse in character and, as I discuss below, are not static but respond to changing circumstances over time. Rather than fixed and binary distinctions between ‘translocal’ and ‘monolocal’ households, it is more useful to consider an evolving ‘spectrum of translocality’ in which some households are translocal in all senses, some only partially or temporarily, and a minority are genuinely isolated from wider household and family support systems. Further, where households are situated on this spectrum is mediated by the gender, relationship status, ethnicity and political connections of older people. By exploring how these dynamics of support play out in intergenerational, translocal households containing older people, migrants and their children, this article argues that debates about ageing and development need to better reflect the key role played by older people within these webs of translocal interdependence, while also acknowledging—and seeking ways to address—the many challenges they face.
The article begins with an overview of the research context, study site and research methods. This is followed by the main findings presented in two sections: the key characteristics of intergenerational, translocal households headed by older people, and the gendered, translocal interdependencies in which older people and migrants in these households are embedded. The article concludes with reflections on the relevance of these findings to research and policy on ageing and development.
II. Study Context
Internal migration is a significant phenomenon in Uganda and linkages between migrants and family members in multiple locations are often substantial and persistent. The 2016/2017 National Household Survey in Uganda found that 16% of the population had migrated in the last five years (Uganda Bureau of Statistics [UBOS], 2018). Migrant-sending households nationally are larger and more likely to be headed by a woman, while migrants themselves are mainly younger adults, with men and women just as likely to migrate (World Bank, 2016). Peaking among individuals in their twenties, migration declines significantly with age and relatively few individuals over 50 years migrate. Migration appears to be beneficial in terms of welfare gains for migrants measured through consumption, with bigger improvements for those moving to urban areas (World Bank, 2016). Links between rural and urban areas are also substantial and persistent over time (Barratt et al., 2012). One study found that over a quarter of households in Uganda receive remittances from internal migrants, with both rural and urban migrants mainly remitting to rural family members (McKay and Deshingkar, 2014). At the same time, cultivation on rural family farms ‘back home’ continues to play a critical role in food security for many urban households in Kampala (Pottier, 2015)
Urban livelihoods are, however, highly precarious and skewed towards Kampala, where inadequate infrastructure and limited employment opportunities leave urban migrants highly vulnerable (Bidandi and Williams, 2017; Mukwaya et al., 2010). Migration thus forms an important livelihood option for younger generations but does not imply a severance of ties with rural sites of origin which remain an important form of ‘informal social protection’ to cope with livelihood failure, illness or breakdown in relationships (Getu and Devereux, 2013). Just as significant is the longstanding tradition of leaving children in the care of grandparents for a variety of economic, emotional and socio-cultural reasons, with children also being reciprocal providers of care (Schatz et al., 2018). These practices have, however, been profoundly affected by the HIV epidemic which placed substantial strain on grandparents so that by 2012/13, 45% of children who had lost one or both parents were living in the care of older people (Ministry of Gender Labour and Social Development [MGLSD], 2015).
People over 60 represent 4.5% of the population nationally, or 1.7 million individuals, of whom 19% are living in poverty (UBOS, 2018). Most (83%) live in rural areas and are still working (67%), almost all in agricultural livelihoods. Over 70% of households headed by older people have caring responsibilities for children (MGLSD, 2015) and one study found a 20-percentage point increase in poverty in these households compared to those without children, attributing this to the greater resources needed for their care (Kakwani and Subbarao, 2007). Older women face particular challenges related to inheritance practices which prioritize the division of land among male children upon the death of husbands, and households headed by female widows are significantly more likely to be poor than those headed by male widowers (World Bank, 2016: 20).
The study site, Kiboga District, is approximately 120 km north-west of Kampala in the Central Region. It was selected for the study because it has a history of in-migration and, more recently, substantial out-migration. The availability of cheap land was previously a driver of in-migration from other regions of Uganda, and the population grew by 5.8% annually between 1969 and 1980 (National Environment Management Authority [NEMA], 1998). It was badly affected by the Civil War in the 1980s but another period of population growth followed the end of the conflict. By the end of the 1990s, despite being within the Buganda Kingdom, only half (53%) of the population was Baganda (NEMA, 1998). However, the availability of land has declined over time and land ownership is now a source of tension within the District and a reason for younger generations to migrate elsewhere (Green, 2006; Kjaer, 2017).
Kibiga Sub-county was selected as the study site within Kiboga as it has a relatively high population density and a mixed agricultural economy, in contrast with the much lower-density, pastoralist ‘cattle corridor’ running through the north and east of the District. Kibiga contained 25,685 people in the 2014 census, of whom 1,352 (5.3%) were 60 years old and above (UBOS, 2017).
III. Methods
A combination of quantitative and qualitative data was collected during two main phases of fieldwork in 2017 and 2018. The findings discussed here were part of a broader study on the intrahousehold effects of the Senior Citizen Grant (SCG)—a non-means-tested social pension for older people in Uganda—so local officials involved in the pension helped select participants, all of whom were over 60 (this age threshold was the SCG’s eligibility criteria in the District). The research methods were a household questionnaire, in-depth interviews with older people and migrants and focus group discussions (FGDs) with older people. Interviews were carried out in local languages with the support of a research assistant. Seventy households were purposively selected for the questionnaire, with local guides asked to prioritize those containing individuals that had recently out-migrated. Efforts were also made to include different types of households by gender of the head of household, household size and relative wealth.
The primary purpose of the questionnaire was to collect data to assist in identifying household types and to identify households for the in-depth interviews. Questions were included on socio-economic characteristics (housing conditions, assets, source of income, food security and perception of relative wealth) to identify wealthier and poorer households. To allow for critical engagement with its findings, the majority of questions—including those on housing, income, assets and migration—were based on the household questionnaire for the impact evaluation carried out during the pilot phase of the SCG. 1 Two FGDs were also held at this stage to explore older people’s understandings of wellbeing and how this was affected by the migration of family members. Importantly, given the findings below, basic questions on ethnicity were included within the questionnaire even though this data had not been collected in the impact evaluation of the SCG. This was because key informants had highlighted the ethnic diversity of the field site during initial discussions and stressed the importance of hiring a research assistant who could speak multiple local languages. However, households were not purposively selected on this basis since it was not anticipated in advance to be a major factor in differentiating between them.
Twenty-five households were selected for in-depth interviews. A balance of female- and male-headed households were included and a range of other characteristics were also taken into account. These included the number of children in the household, self-reported and asset-based wealth and the destination of migrants. Based on preliminary analysis of the data in the field, which suggested there were some differences between Baganda and non-Baganda households, ethnicity was also taken into account in selecting participants at this stage. Interviews were semi-structured and designed to investigate household living arrangements, livelihoods, relationships and support networks of older people and migrants in translocal households as well as how these impacted their overall wellbeing. Twenty interviews with migrants from older people’s households were also conducted to explore migrants’ perspectives on these issues.
In a second round of fieldwork one year later, follow-up interviews were conducted with 16 older people. This round primarily focused on relational dimensions of wellbeing beyond material needs and understanding how households had changed over the course of the year. Two further FGDs were also carried out with groups of older men and women to generate insights into gendered findings from the first round of interviews as well as differences between households headed by Baganda individuals and older people from other ethnicities.
All interviews were transcribed in English by the research assistant with coding and analysis conducted by the author using NVivo 11. The coding was carried out using a combination of themes identified prior to fieldwork and those that arose during data collection and analysis. Responses to the questionnaire and answers from older people and migrants during in-depth interviews were also compared and contrasted to build up a more detailed picture of households and to identify any differences in the responses between the different sources.
IV. Findings
Intergenerational, Translocal Households
The household questionnaire identified a number of common features across households in the sample. These included the diverse ethnicity of older people, their almost exclusively agricultural livelihoods and the high prevalence of grandchildren in households containing older women. Further, almost all households were characterized by multiple migrations across different generations, with older people having migrated into the area, while younger generations were now moving out.
Households headed by older people were large, containing 5.5 people, compared with 4.35 for Kiboga as a whole (UBOS, 2017). Each contained nearly three children aged 14 or under on average, with female-headed households having one child more than male-headed households. However, the presence of an older woman was crucial in both cases as households headed by single men without female partners had only 0.5 children on average. Excluding those over 60, 49% of the remaining household members were aged 15–18, a further 23% were 19–30 and only 28% were aged 31–59. So while households were ‘intergenerational’ in nature, most co-resident members were either children, young adults or older people.
Using a self-reported measure of poverty with values ranging from ‘very poor’ to ‘wealthy’, female-headed households described themselves as slightly poorer than male-headed households. No one self-identified as ‘wealthy’ and only nine individuals (13%) considered themselves ‘above average’. Thirty-one individuals (44%) felt themselves to be ‘average’ in terms of wealth, 21 (30%) considered themselves ‘poor’ and 9 (13%) said that they were ‘very poor’. All households were engaged in agriculture, with 82% of older people still working in the fields. Overall, households were thus dominated by agricultural livelihoods with older people shouldering a significant burden of work alongside a relatively small proportion of co-resident ‘working age’ individuals—mostly young adults engaged in agriculture, often as wage labourers—and school-age children, 95% of whom were reported to be also attending school.
Only 40% of heads of households said that they were Baganda, with most identifying as members of tribes from West or South-West Uganda or Rwanda. During the qualitative interviews, two pensioners who had described themselves as Baganda turned out to be Banyoro and Banyarwanda, respectively, suggesting that even these figures may overestimate Baganda households. Further, almost all older people, regardless of ethnicity, had moved to their current location from elsewhere, with only two interviewees living where they were born. Most, however, were long-term residents, with almost all arriving in the area by the 1980s. Nonetheless, it was clear that in-migrants from other ethnicities were less well connected politically than their Baganda peers. Four out of 10 Baganda interviewees had occupied formal political roles such as the Parish Chief or other local leadership positions, but no-one from other ethnicities had done so.
Fifty-three percent of households reported a ‘recent out-migrant’ within the last 12 months, while 90% reported migrants who had left previously but retained substantial links with the household. These were identified through the following question: ‘Are there any other people who left more than one year ago who you consider to be ongoing members of the household?’
A total of 126 individuals were identified in this way, or 1.8 per household. In addition, a third set of individuals was captured at the end of the questionnaire by asking about the parents of children living with the pensioner who were still alive but had not been mentioned previously. This identified a further 41 individuals in 37% of households, of whom 71% were men (primarily adult sons). Having a child within the household was thus not always sufficient for pensioners to identify individuals as an ‘ongoing member’. As will become clear below, the nature of the relationship—including the level of communication and flow of support—was generally the more important factor.
The questionnaire findings supported the notion that older people had migrant family members who retained substantial and persistent linkages to the co-resident household. Building on this, the in-depth interviews with a sub-sample of 25 households clarified the character of these ongoing relationships. Narratives on initial migration choices often focused on individual decision-making by migrants, rather than a planned or ‘strategic’ approach at the household level. Nonetheless, linkages with migrants were not only persistent, they were found to be critical to the functioning of most households. In other words, without taking the contribution of migrants into account, it was not possible to adequately understand the nature of the older person’s ‘household’.
These linkages between older people and migrants were not uniform in nature even within individual households, and there was considerable diversity in perceptions of the nature and strength of these connections. However, most of the 25 households who took part in the in-depth interviews were translocal in at least some of the following senses (adapted from De Haan et al., 2000; Potts, 2010; Spiegel et al., 1996):
Persistent linkages involving the flow of resources and support in multiple directions. The presence of children of migrants at sites of origin. Fluid membership of the household across different locations. The regular return of migrants in cases of sickness or other crises such as relationship breakdown or livelihood failure. Shared common purpose among individuals in different locations. Co-insurance and risk pooling.
In Table 1, these dimensions of translocality are mapped against households grouped into six categories based on the gender, ethnicity and relationship status of household heads. As mentioned above, these categories were developed during analysis of the first round of data and were subsequently discussed during interviews and FGDs in the second round of data collection. Although there is considerable variation, households in the first and second categories tend to correspond to the fewest dimensions; those in the third category are quite mixed, reflecting how older men often portrayed relationships with migrants—including resource flows, risk pooling and planning—in quite a negative light, while differences are less pronounced between the remaining categories where households frequently meet many of the criteria, or in one case, all of them. This shows how households in the research sample occupied a ‘spectrum of translocality’ ranging from coherent, supportive translocal households, to those with less unified and less reliable—but often highly persistent—linkages between migrants and older people.
Dimensions of Translocality by Household Category.
Support in these households could potentially flow in both directions and the option of return for themselves or their children represents an important safety net for migrants. Households were thus not static. Of 16 households revisited in the second phase of fieldwork only one had not altered in some way, most commonly by the departure or arrival of grandchildren. Further, since grandchildren were only living in significant numbers alongside grandmothers, even where older men were present it was older women who ‘anchored’ the translocal household in its rural setting.
Importantly, the gender, relationship or marital status, ethnicity and political connections of older people were found by the researcher to be key factors in explaining the position of households along this spectrum. These categories, in turn, corresponded to different degrees of translocal intrahousehold support (based on the list above) that both reflected—and were shaped by—these demographic and socio-economic factors. Table 2 summarizes the features of these households and how they related to poverty, land ownership and the average age of household heads. The concept of the translocal household thus emphasizes the persistent linkages and translocal interdependencies between older people and migrants, but individual households must be understood in relation to this wider set of characteristics that shaped and underpinned these interdependencies.
Key Features of Households.
Gendered, Translocal Interdependencies
The discussion of households presented above is primarily structured along the lines of gender, relationship status, ethnicity, relative wealth and political connections. As noted, there were important and unanticipated differences between Baganda households and other ethnic groups within the quantitative and qualitative sample in terms of the strength of their informal support systems and social networks. Baganda households were wealthier with more translocal linkages involving greater flows of support. Households headed by older people from other ethnic groups had weaker social support networks, both locally and translocally, and more limited political connections and traditional ties with clans. I will return to this issue in more detail towards the end of the section. Arguably, however, it was the gender of the older people in the household that was dominant in structuring the translocal household and interdependencies with migrants.
Relationships with migrants were presented as contingent by almost all older people and most felt that only some individuals could be relied upon. Particularly supportive children were often singled out, as were the least helpful individuals, and this was true even in households that were strongly translocal in character. Reflections on the provision of support permeated these discussions, whether relationships were broadly positive or negative. Older men, however, more frequently spoke in critical terms. For example, highlighting a tendency to support mothers over fathers, or that migrants sometimes wanted to help themselves.
According to me I think we no longer have good relationships with our children. And the children mostly consider supporting their mothers compared to their fathers so for us we have to work hard to make sure that we always have something to survive on. (Older man, FGD4)
No way, because they also don’t have anything to help me with and they are not even bothered, they are also looking for help from me. (Older man, HH021)
Older women, in contrast, generally offered a more balanced view, weighing upon positives and negatives.
Ah, the positive factors, yes there are bad things, and also good things. Like for me I have some older children that stay away from home who send me some money once in a while. I also have some young grandchildren whose parents ran away and abandoned them, especially our daughters.… So there are some good sides … [and] some dark moments. (Older woman, FGD1)
They were also more likely to describe supportive relationships in moral and emotional as well as practical terms.
What makes me happy is that we have now grown old, God gave us children, and in the process our children have also produced and given us grandchildren. They give us happiness, they protect us and we also protect them. (Older women, FGD2)
A balanced perspective was also characteristic of how older women spoke of young children in the household. Grandchildren were frequently cited as a positive feature of their lives and—as in the quote above—a rewarding aspect of ageing.
The joy of having the grandchildren here, knowing that all my children went, but at least they are left and so I look at them and I don’t worry so much. (Older woman, HH004)
However, older women nonetheless emphasized the challenges they faced in providing care.
We see it as a very big burden because of the responsibility attached, but on the other side we see it as a blessing to have a grandchild.… They are a blessing, but we carry a lot of responsibilities trying to take care of them. (Older woman, FGD2)
These caring roles required significant resources, including cash from the social pension to cover school fees, necessitating flows of support in both directions between older people and migrants, both directly and indirectly, where migrants’ children were involved. Indeed, across the 25 interviewees, only 3 older people were not paying any school fees or costs.
Sometimes I pay half the fees for those whose parents can contribute but there are others whose parents can’t manage and I pay. Especially those who are staying in Kampala … their mother pays half and I top up. (Older woman, HH059)
Support to migrants in times of crisis was also critical in this context, with multiple examples of adult migrants returning to live in Kiboga as a result of problems with livelihoods, relationships or ill-health. In the case of illness specifically, it was older women who provided care for migrants. Three older women had sick or disabled migrants living with them and others mentioned that migrants commonly returned for shorter periods to recover from illness or, in other cases, to give birth. None of the individuals in the research sample were able to help in significant ways with housework or in the fields. In common with migrants whose livelihoods or relationships had failed, most were hoping to leave again in future once their health recovered.
Migrant: When I am completely healed I will go to look for work …
Interviewer: Where would you wish to be, between Kiboga and Kampala?
Migrant: I wish to be in Kampala of course.
Interviewer: Why Kampala?
Migrant: Kampala is better than staying in the village. (Sick migrant, HH004)
Both migrants and older women acknowledged the strains this placed on each other, but also the importance of this safety net given the insecurities of urban life.
At the same time, migrants were acknowledged to be critical in providing care and resources where older people fell seriously ill or in cases of crop failure or other crises. Women were aware of particular vulnerabilities in this regard, especially upon the death of husbands where conflicts over inheritance rights were common, even if their rights to land were formally recorded. In one case, the husband died between the first and second round of fieldwork, and a son had changed his livelihood from ‘boda’ (motorbike) driving to agricultural trading to support his widowed mother.
Interviewer: And you said that [he] is no longer riding a boda and now he is here to trade?
Respondent: Yes, he left that one and he decided to come and help me around here.
Interviewer: Did he make that change when his father was still alive or after his death?
Respondent: He decided on it when his father was still alive. And even in nursing him we were doing it together at the hospital bed. (Widow, HH041)
In another, a woman whose husband died many years ago with their new house uncompleted, continued to inhabit an older dilapidated house on the same plot of land. Although the household lacked the resources to finish the work, it was nonetheless inhabited by a fluid cast of children, grandchildren and other relatives anchored around the older woman. Individuals stayed for different lengths of time and in response to both the grandmother’s needs and their own, with ill-health, pregnancy, schooling and livelihood decisions all playing a role. In all these cases, relationships of support were thus framed in terms of mutual dependencies; and these were central to both the positive dimensions of ageing and to the challenges they faced.
For most older women, therefore, complete ‘independence’ from others was not a desirable state even for those who were in reasonable health and relatively wealthy (Fine and Glendinning, 2005; Sevenhuijsen, 2003). Translocal relationships of support were fundamental to their wellbeing; indeed, as everyday social interactions with migrants were less regular, the importance of sending resources was magnified accordingly. Policy discourses, including those on social pensions, that simply aim to reduce ‘dependence’ among older women do not reflect how crucial these translocal interdependencies are to their long-term wellbeing. They are, as Ferguson puts it, desirable forms of dependence (2015: 163) with parallels in, for example, the ‘dispersed dependencies’—encompassing the social, emotional, physical and economic realms—that Easton-Calabria and Herson (2020) argue are crucial to the wellbeing of displaced people in humanitarian contexts but are often obscured by negative framings of dependence grounded in a ‘supposed dependence-independence binary’ (2020: 45).
Some older men also expressed positive thoughts about their grandchildren when asked directly. However, none identified grandchildren as a source of their own wellbeing or as essential to a good quality of life in general without being prompted to discuss them and most were quite negative about the responsibilities involved. During the male FGD, for example, the emphasis was on the inconsiderate way they felt grandchildren had been sent to live with them.
When they get children and have misunderstandings with their wives we are the ones who end up carrying the burden. For example, I have five grandchildren from different women for my sons.
When asked about how they understood wellbeing, older men were notably more focused on maintaining their independence, particularly regarding agricultural livelihoods.
I just hope that God helps me benefit from my garden and also to get over this sick leg and have the energy to move and be a part of a savings group as well as to tend to my gardens very well, then I will be grateful. (Older man, HH021)
As noted above, they were also more critical about relationships with migrants as a whole and many felt the quality of their relationships with their children was comparatively worse.
For us as men we have a bigger challenge than on the side of the women because they are the ones who are closer to the children.… When we as fathers try to talk to the children and get closer to them they are a bit detached from us. (FGD4)
However, older women also noted that men were more likely to reflect negatively on whatever support they did receive.
Respondent: I can appreciate even if my child gives me 500 shillings unlike a man who will look at that as something very small.
Interviewer: Why wouldn’t a man appreciate that?
Respondent: Maybe because he feels it is nothing to him. (Older woman, FGD3)
In responding to the questionnaire, for example, women described receiving more regular support and included smaller amounts that men may not have bothered to mention. This supports findings from other studies in Uganda suggesting older women may be more likely to report receiving financial support than men (Mugisha et al., 2015).
Counter-intuitively, however, men were more likely to mention the importance of support from their family when discussing their hopes or plans for the future, with 4 of 10 older men responding this way to a direct question, while only 1 woman out of 16 did so. Older men were also more likely to provide direct financial support to migrants. Except for in-kind food transfers and small sums to cover transportation costs, all the examples of substantial financial support provided by older people to migrants involved older men, mostly (but not exclusively) giving money to female migrants. Thus, while men may have played down help they were receiving, they still perceived these relationships to be of considerable significance and understood support both to and from migrants as an integral component of them. In this way, aspirations for independence may reflect gendered attitudes to ‘self-sufficiency’ without necessarily revealing the ‘desirable’ (inter-)dependencies upon which even the most visibly independent men relied (Easton-Calabria and Herson, 2020; Ferguson, 2015).
For research on informal social support systems, this could lead to problems in assessing the relative support provided to men and women. Women may overstate the support they receive, while men downplay it.
I think for us, the women, we just accept and decide to cover up for our children even if they don’t support us, we can say they do. Unlike the men who just expose them plainly. (Older woman, HH044)
Given the added burden of care most older women have, any assessment of the translocal support to which men and women have access should give careful consideration to this dynamic, especially if evidence seems to suggest women have equal or even greater access to resources. For example, in a review of research on social networks and resilience in rural communities, Rockenbauch and Sakdapolrak (2017) call for a ‘translocal social network perspective’ to understanding the capacity of households and communities to forge resilient livelihoods. However, while acknowledging the importance of power asymmetries within these dynamics, they do not consider how gendered responses from participants which downplay or even exaggerate this support might influence or distort research findings on the networks themselves.
These gendered aspirations to ‘independence’ also accord with other evidence in the region which suggests that older men may struggle to align care work with societal norms of masculinity (Mudege and Ezeh, 2009; Munthree and Maharaj, 2010) and that ‘common narratives of masculinity give primacy to time, money and energy spent on leisure activities in old age rather than on care’ (Schatz and Seeley, 2015: 1192). Compounded by normative framings of masculinity that position men as familial ‘breadwinners’ with primarily financial responsibilities, rather than providers of emotional or physical care work, this may make both the giving and receiving of care psychologically challenging for many older men, especially where it coincides with a decline in their ability to ‘provide’ financially (Schatz and Seeley, 2015). This model of masculinity aligns with the narratives of independence described above, even where it does not match the reality of ongoing patterns of support and mutual interdependence which bind older men and migrants together.
It is not possible here to explore further gendered dynamics in detail, but I will conclude by noting that certain groups faced particular challenges. This includes widows, many of whom spoke of a ‘double burden’ they had taken on where traditional male responsibilities were added to their existing female roles.
Some of us we don’t have our husbands, they died so we are widows, all the responsibility formerly used to be undertaken by our husbands and now rests on our shoulders. (Older woman, FGD2)
The three female interviewees who received no support but nonetheless provided a lifeline to migrants in crisis were also particularly vulnerable. For example, all three reported very significant difficulties during a delay in SCG payments preceding the second round of fieldwork. Analogous to these were nine men in the questionnaire who reported having limited or no communication with migrants. They were not selected for the interviews, so detailed analysis of this group is not possible here. However, during brief discussions, it was apparent they were mostly in-migrants living without any form of support or connection to their family, locally or further afield. Sadly, one of these men had committed suicide by the second round of data collection. Beyond the translocal households that are the subject of this article, these highly vulnerable categories of isolated women and men would be an important area for further research.
Notably, all three women and seven of the men not receiving support were from tribes other than the Baganda. As the typology above illustrates, this highlights how significant migration histories and ethnicity was in understanding networks in these translocal households. Baganda households were wealthier on average with more land, more substantial translocal links and local sources of support and stronger political connections. For example, while the men who provided financial support to migrants were from all ethnic groups, the two examples where migrants were given land to work on or sell to finance their migration plans were both politically-connected Baganda older men. Similarly, many Baganda households had notably stronger support networks at destination, particularly in Kampala, and these households included the only examples of migrant family members with formal sector jobs and university educations.
Clan networks beyond immediate kin were also important in this regard, although most interviewees highlighted their role in burials (and to a lesser extent marriage and other cultural events) rather than daily lives or times of crisis. The exceptions to this were all Baganda individuals, who emphasized the importance of clan relationships in terms of mutual support, access to plots of land and maintaining social networks. In contrast, non-Baganda interviewees described the significant difficulties of maintaining these connections at distance.
There are some disagreements in some of our clans. For example, when I went back to my roots in Ankole after my father passed away, the clan members refused me any rights on the property so I never went back. (Male participant in FGD)
Although some practices were adapted to their new location, such as adopting the equivalent Baganda name for their own totem animal, the clan nonetheless had diminishing relevance for most of these interviewees.
We do not have a particular clan we are attached … because our parents died a long time ago. They were the ones who would have told us the clan to be in. (Older woman whose parents migrated from Rwanda, HH034)
Finally, it is worth reiterating the degree to which many migrants relied upon older people for help in establishing themselves, in caring for their children and for practical and emotional support in periods of personal difficulties. Interviews with migrants highlighted some tensions with older people, sometimes relating to the decision to migrate in the first place, but more often regarding the sharing of resources for childcare, health services and everyday needs. Concerns were also expressed by migrants about the ability of older people to cope with caring roles and in maintaining productive livelihoods as they age further. However, in most cases it was the strength of the emotional as well as practical bonds within translocal households that was most striking during interviews. As demonstrated above, these translocal bonds and the support channelled through them are not evenly distributed across households nor are they static in nature. Nonetheless, while these ties may be reshaped and sometimes disrupted in profound ways by contemporary trends, improved communication technology and increasing precarity are in many cases strengthening rather than undermining the value of translocal interdependencies for migrants as much as for older people.
V. Conclusion
This study showed how gendered, translocal interdependencies underpin many households headed by—and centred around—older people in Kiboga and that these are critical to the wellbeing of both older people and migrants. However, it also demonstrated that these households are diverse in character and occupy a ‘spectrum of translocality’ ranging from coherent, consistently supportive translocal households, to those characterized by less unified and less reliable—but often highly persistent—linkages. Importantly, the gender, relationship status, ethnicity and political connections of older people were key factors in explaining the position of households along this spectrum.
Older women face particular vulnerabilities, especially widows and those who are living alone. However, they also derive greater emotional satisfaction from these translocal interdependencies alongside their material benefits. Older men, in contrast, are more openly ambivalent about their relationships with migrant family members and the support they both give and receive, even as they acknowledge the importance of these translocal relationships as they age further. Migration is part of everyday life in Central Uganda. However, while almost all older men and women were in-migrants at one stage of their life, those who moved from further afield are often less well supported both locally and translocally in their old age. This affects them directly, for example, in access to land and political influence, but also indirectly through the opportunities, resources and connections their migrant children and grandchildren can access.
Thus, not all households are equally vulnerable, but policy debates about ageing, care work and the links between formal and informal social protection in the context of migration, frequently overlook these intersecting dimensions of privilege or vulnerability. For example, attempts to promote ‘synergies’ between informal or traditional social protection and formal programmes must be sensitive to the communities within which they operate. Thus, supporting traditional clan or tribal structures through formal ethno-development organizations in this regard—as has been suggested at times in Uganda (Kyaddondo and Mugisha, 2014) —would be a highly inequitable way to proceed in the context of multi-ethnic Kiboga, unless great care was taken to address rather than exacerbate existing gendered and ethnic inequalities in doing so.
At the same time, the role of older people in caring for the children of migrants is frequently mentioned, most often in policy debates about the cumulative effects of HIV and rural–urban migration on intergenerational caring roles. However, there is often little practical consideration of the material, emotional and relational consequences for older people in development policies or programmes. For example, evaluations of the SCG frequently note the positive impacts on children within pensioner households (e.g., Gelders and Athias, 2019). While the behaviour of individual pensioners in supporting children is laudable, particularly grandmothers as ‘anchors’ of these households, this should not detract from a recognition of the challenges they face in doing so. The use of the pension to cover school fees for the children of migrants is indicative of a wider set of problems that Uganda faces in relation to ‘universal’ childcare and schooling, and one the onus of which should not be on older people within the community to address.
Finally, this article has also shown that older women may overstate, and older men understate, the translocal support to which they have access, with important consequences for their own relational and material wellbeing as well as for development programming targeted at older people. A gendered, relational framing was essential in understanding these dynamics of support within intergenerational, translocal households. This suggests that policy and research on ageing and development in Africa—and other developing country contexts where translocal households are common—should aim to better reflect the key role played by older people within these gendered webs of translocal interdependence, while also acknowledging—and seeking ways to address—the many challenges they face.
Footnotes
Acknowledgements
This article was first presented at a panel on ‘Intergenerational boundaries and migratory borders’ at the RGS-IBG Annual Conference in 2021 and benefited from the helpful comments of participants in the session. It draws on my doctoral research and thus benefitted from the insights and support of Tanja Bastia, Armando Barrientos and Sam Hickey. Comments on later drafts by Penny Vera-Sanso also helped to improve the article substantially. Finally, I would like to acknowledge the important role played in collecting the data by my research assistant, James Bamwenda Tayebwa.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
This work is supported by the Economic and Social Research Council (ESRC).