Parental Migration, Intergenerational Obligations and the Paradox for Left-Behind Boys in Rural China

ABSTRACT

Drawing on in-depth interviews with caregivers of left-behind children (LBC) in rural China, this article seeks to explore their understanding of migration motives and the social process of taking on care-giving roles for LBC. The authors argue that there are underlying socio-cultural explanations pertaining to economic motives for migration; such as, making contributions to social events (weddings and funerals) in village life, and fulfilling social obligations for left-behind sons’ futures. Parents migrate to save for sons’, but not daughters’, adult lives. Grandparents, particularly on the paternal side, are expected to fulfil social obligations to care for left-behind grandchildren, even without immediate financial returns. These suggest that left-behind boys, and in particular boys cared for by paternal grandparents, may be at greater risk than other LBC, as they may receive even fewer resources in the form of remittances from migrant parents in their early childhood.

KEYWORDS:

• rural-to-urban migration
• left-behind children
• China
• intergenerational exchange

Introduction

Migration in developing countries and deprived areas can relieve poverty at the household level and boost economic growth at the national level (Grindle, 1988; Massey, Alarcon, Durand, & Gonzalez, 1987; McKenzie & Sasin, 2007; Mines, 1981). Many theories have been developed to explain labour migration from an economic perspective (Harris & Todaro, 1970; Jorgenson, 1961, 1967; Lewis, 1954; Stark & Taylor, 1991). Neoclassical economics emphasises the importance of human capital and considers migration as a rational choice to maximise individual net benefits (Sjaastad, 1962; Todaro, 1980). However, this has been challenged by the new economics migration theory that views migration as a strategy to maximise the net gains for the entire household rather than for an individual (Adams & Page, 2003; De Jong, 2000; Stark, 1991). Thus, the main difference between the two perspectives of migration is whether decision making is at the individual level or the household level (Massey et al., 1993). Grasmuck and Pessar (1991, p. 15) suggest that: ‘It is not individuals but households that mobilise resources and support, receive and allocate remittances, and make decisions about members’. Economists tend to focus on the economic nature of migration and assume homo economicus in many economic models. Anthropologists and sociologists, however, argue that economic motives are far from sufficient to predict human movement without accounting for social and cultural contexts that constrain human behaviours (Brettell & Hollifield, 2008). The household rather than the individual as a social and economic unit can help understand the interaction of structure and agency in most societies (Morgan, 1989; Wallace, 2002), which may enable researchers to integrate other socio-cultural factors with economic factors in the study of migration. The household approach can also help interpret the phenomena of temporary migration, split family structures and remittances in migration.

The Chinese context: internal migration and left-behind children

China is a good example for studying temporary migration as the world's largest internal migration has been taking place there, since the reform and the opening up of the economy, which started in the late 1970s (Hu, 2012a), resulting in a massive surplus of rural labourers transferring to cities. At the national level, about 15% of all rural families have at least one member migrating to an urban area (Zhang, 2004). The population of migrant workers was around 268.94 million in 2013, up by 11.0% over that of 2010 (National Bureau of Statistics of China, 2011, 2014). The massive internal migration, mostly of young adults aged 25–49 years old (Hu, Cook, & Salazar, 2008; Wong, Fu, Li, & Song, 2007), has fundamentally altered the Chinese demographic landscape.

The split family structure is common in rural China due to internal migration. The household registration system (known as hukou), which used to act as a strong constraint on individual rural-to-urban movement throughout China (Vendryes, 2011), limits migrant workers’ access to urban welfare entitlements (Gong et al., 2012; Xiang, 2007). Therefore, migrant workers have to leave their children, especially those of school-age, in their rural hometowns (Ye & Pan, 2011), resulting in a social group of ‘left-behind children’ (LBC) in contemporary China. The most recent survey based on the Sixth National Population Census of the People's Republic of China estimates that the overall population of LBC aged 0–17 in rural China in 2010 was around 61 million (accounting for 37.7% of total rural children and 21.9% of all children in China) (All-China Women's Federation, 2013).

In the Chinese context, internal migration has indeed alleviated household poverty in migrants’ rural communities of origin (Huang & Zhan, 2008; Zhu & Luo, 2010). Economic motives for migration are often highlighted (Fan, 2007; Ye, Wang, Wu, He, & Liu, 2013), and seemingly, taken for granted without examination of the underlying explanations. The focus on household economic gains can neglect non-economic sociocultural dimensions that underlie decision-making for migration (Fan, 2007). For example, narrative data from migrant workers suggest that, unlike their previous generation, ‘new generation’ migrants born after the 1980s, participate in rural-to-urban outflow for gaining new experiences and broadening horizons apart from economic motives (Hu, 2012b).

Care-giving arrangements for left-behind children in China

Migration is a complex decision-making process at the household level (Agesa & Kim, 2001; Stark & Bloom, 1985). It concerns not only those who intend to migrate but also those who stay behind. A migrant's decision making often involves considerations for other family members (Rowland, 1994). In particular for migrant parents, care-giving arrangements for their left-behind children may become a major concern. In China, both parent migration accounts for around 46.7% of the 61 million LBC. Over one in five of LBC aged 6–11 share a household with a parent plus grandparents, while one third live with grandparents only (All-China Women's Federation, 2013). Grandparents therefore play a key childcare role in multigenerational as well as ‘skipped generation’ households (Burnette, Sun, & Sun, 2013).

Kinship care of children, especially by grandparents, is a common arrangement in the developing world, and it can be a cultural norm for extended families to play an active role in child rearing (Ma, 2010). Adult children may be motivated to take up migrant work to improve the household's economic status when their own parents are available to look after young children. Those who are left behind in the extended family can constitute ‘a reliable and effective support system’ (Chang, Wen, & Wang, 2011). This is consistent with the group/mutual aid model, in which each household member's capacity is maximised to ensure the success of the household as a whole (Lee & Xiao, 1998). By caring for their left-behind grandchildren, grandparents enhance their migrant children's economic capacity to reciprocate in the form of remittances (Agree, Biddlecom, Chang, & Perez, 2002), which may compensate the grandparents’ efforts as surrogate parents. This form of intergenerational exchange has been called ‘intergenerational contract’ (Croll, 2006) or ‘time-for-money’ exchange, which has been observed in several Asian countries including China (Cong & Silverstein, 2008, 2011; Frankenberg, Lillard, & Willis, 2002; Shi, 1993).

Rationale

Drawing on in-depth interviews with caregivers for LBC in rural China, we explore how caregivers explain migration motives and the social process of taking on care-giving roles for LBC. Most of the literature to date tends to focus on determinants and/or consequences of migration, or the care-giving arrangements of LBC. We argue that decision making for migration and care-giving arrangements are inherently related to each other and are operated at the household level. It is essential to take into account of caregivers’ perspectives, as critical social actors in the migration process they do contribute to migration decision (Cohen & Sirkeci, 2011). Although caregivers may act as an ‘outsiders’ for migration decision-making, their perceptions of migration can be important because of the potential to influence their decision about taking on care-giving responsibilities for LBC, which in turn may affect their own as well as LBC's well-being.

Despite rapid socio-economic changes that China is undergoing, the structure of Chinese families continues to be one of mutual dependence and family obligations still play an important role in family relationships (Qi, 2015). We argue that there are some underlying socio-cultural explanations pertaining to economic motives for migration; for example, making contributions to social events (weddings and funerals) in village life, and fulfilling social obligations for left-behind sons’, but not daughters’, futures as an important driver to motivate parents to migrate. Grandparents, particularly on paternal side, are expected to fulfil social obligations to care for left-behind grandchildren even without immediate financial returns. Children left in the care of paternal grandparents may actually receive fewer remittances from their migrant parents than maternal grandparents. This suggests a paradox of intergenerational obligations for boys in a society where sons are culturally more valued than daughters; because parents migrate to save for sons’, rather than daughters’, adult lives; left-behind boys, in particular boys cared for by paternal grandparents, may have fewer resources in their early childhood than other children in rural China.

Methodology and data

A qualitative study design was used to enable the generation of rich and in-depth information about the complex phenomenon of migration and care-giving arrangements (Creswell, 2006). Qualitative research can ‘capture complexity of family life in the sense of an ongoing stream of behaviour over long period of time that interweaves the perceptions of individuals with behaviours in relationships’ (Wampler & Halverson, 1993, p.189), and produce rich data to provide a good understanding of the intra-household dynamics and/or social processes behind the numbers (Jones & Sumner, 2009).

Research setting

This study was carried out from September 2013 to February 2014 in a rural township in Henan Province, People's Republic of China. Henan Province is located in the northern central part of China, the most populous and traditionally one of the largest migrant-sending provinces throughout China (All-China Women's Federation, 2008, 2013). The annual per capita disposable income of urban residents in 2012 was 19,408 RMB (approximate 1940 GBP) and 7432 RMB (approximate 740 GBP) for rural residents (Yongcheng Government, 2013), which is lower than the national average (24,565 RMB or approximate 2450 GBP for urban residents, and 7917 RMB or approximate 790 GBP for rural residents) (National Bureau of Statistics of China, 2014).

The agricultural township where this study was conducted had a population of 55 thousand (Han Chinese) and included 30 villages under its administrative jurisdiction. Child participants were initially approached through a primary school. This public primary school was free and available to eligible children of school (primary and secondary school) age from nearby villages. Children who participated in this study were from seven geographically close villages.

There were two coal mines providing job opportunities for local people in the area chosen for this study. As described by the participants, jobs from the coal mining industry were often temporary, labour-intensive and only available to young labourers in their 20s or 30s. Permanent job opportunities were limited and people who secured such jobs were financially better off than their village peers.

Since 2011, rural people aged 60 or over have been eligible to claim for a pension allowance, which is 60 RMB [6 GBP] per month in this township. The pension beneficiaries who participated in this study reported that the small amount of financial subsidy had, to some extent, improved their daily lives and reduced their dependence upon their adult children for financial support.

However, it was a fairly small city centre with limited job opportunities, as described by the participants, so the majority of workers had migrated to large cities such as Shanghai, Guangzhou, Shenzhen, and Beijing. They were working in manufacturing, constructing and catering industries. If they returned home, it was mainly for Chinese New Year and occasionally during the harvest season in August.

Data collection

Child participants were recruited at a primary school. The sample was purposive in order to achieve maximum variation (Creswell, 2006) of age, gender, and family structures (left by the mother, the father, or both parents). To be eligible, LBC had to be children who stayed in rural areas with at least one parent having migrated to urban areas for employment reasons. Caregivers for eligible children were invited to take part in this study.

Face-to-face semi-structured interviews were conducted with caregivers, either individually or together with their partners. LBC were present for ten interviews with their caregivers. The interviews took the form of informal conversations in which the interviewer asked open questions about parental migration and care-giving arrangements for LBC using a topic guide. Caregivers were encouraged to elaborate on their answers and to raise additional topics that they considered relevant. Data collection and data analysis were conducted concurrently. Questions became more focused and specified as data analysis progressed and the key themes emerged. Interviews were conducted in Chinese (Mandarin or dialect). Each interview lasted approximately between 1 and 1.5 hours. All interviews were conducted by the first author in participants’ homes and were audio-recorded with their permission and subsequently transcribed verbatim in Chinese by a different person outside the research team. The first author checked all the Chinese transcripts to minimise data loss.

Twenty-one LBC were recruited (Table 1): 12 boys and 9 girls; 10 were cared for by paternal grandparents, 4 by mothers, 4 by one parent and paternal grandparents, 2 by maternal grandparents, and 1 by an aunt; 19 of 21 LBC have at least one brother within household. Table 2 presents the characteristics of caregivers in the study sample. Of the 21 grandparents (17 on the paternal side and 4 on the maternal side), 15 were aged over 60 years old, and 6 were aged over 70. Only two of them went to primary school and the rest had never attended school. The other caregivers (5 mothers, 1 aunt and 1 uncle) had a middle school education or less. Among grandparents, only one paternal grandmother reported that her husband had a pension as he was previously an electrician in his village. Two grandfathers went to the local town occasionally for part-time work in a construction site. Other grandparents relied on agricultural income for a living.

TABLE 1 Sample size and characteristics of left-behind children (N = 21).

	Left-behind children
	Boys	Girls	Total
Number	12	9	21
Age range
6–9 years	4	8	12
10–12 years	5	4	9
Siblings
None	4	0	4
One brother	4	7	11
One sister	4	0	4
Others (two sisters)	0	2	2
Left-behind types
Only mother away	2	0	2
Only father away	2	4	6
Both parents away	8	5	13
Caregivers
Paternal grandparents	5	5	10
Maternal grandparents	2	0	2
Mothers	2	2	4
Fathers	-	-	-
Shared: paternal grandparents and mothers	-	2	2
Shared: paternal grandparents and fathers	2	-	2
Others (uncle or aunt)	1	-	1
Total	21

TABLE 2 Sample size and characteristics of left-behind children's caregivers (N = 28).

Characteristics	Paternal grandmother	Paternal grandfather	Maternal grandmother	Maternal grandfather	Mother	Aunt/uncle	Total
Age Range
30–39 years	-	-	-	-	2	1	3
40–49 years	-	-	-	-	3	1	4
50–59 years	6	-	-	-	-	-	6
60–69 years	5	3	1	-	-	-	9
≥70 years	1	2	2	1	-	-	6
Educational attainment
Illiterate	11	5	2	1	1	-	20
Primary school	1	-	1	-	2	1	5
Middle school	-	-	-	-	2	1	3
High school or higher	-	-	-	-	-	-	-
Total	12	5	3	1	5	2	28

Ethical approval was obtained from The University of Manchester Research Ethics Committee. Informed consent was obtained from participants using either in writing or by verbal audio-recording from illiterate participants. Pseudonyms have been used to preserve participants’ anonymity.

Data analysis

Principles and procedures of the constant comparative methods guided data analysis (Charmaz, 2006), following transcription and entry into the qualitative analysis computer programme Nvivo 10 (QSR, 2012). Concurrent data collection and data analysis occurred with codes and categories being inductively developed from the data. Analysis involved identifying codes and their properties and dimensions, grouping the codes to create categories, systematically comparing and contrasting the codes, and examining the connections between the categories and subcategories.

Data analyses were initially processed in Chinese in order to prevent misunderstanding and to minimize the risk of losing participants’ meanings. It has been argued that the Chinese language is quite complex and it is often difficult to achieve equivalence of meaning in English language (Twinn, 1998). As the analysis developed, data related to emerging themes were translated into English to facilitate review and discussion with co-authors. The translated versions and the original Chinese versions were checked by an independent and bilingual person outside the research team.

Findings and discussion

Migration to ‘make ends meet’

Carers described the economic motives for migration. Three emphasised that migration was a reluctant decision and driven by economic motives in order to meet the basic needs or survival of the household members who remained behind:

No one would ever migrate away if they have enough to eat at home. (Paternal grandmother 5)

If having enough money for a living, no one would ever migrate away. Just like others [non-migrants], they [migrant parents] also would like to be with their children. Who would not want to be with his wife and children? To put it bluntly, it is just for money. (Paternal grandmother 8)

Some carers described the underlying reasons for migration in terms of a need to ensure individual survival, such as ‘get enough to eat’ and ‘making a living’. These carers’ emphasis on meeting the first and basic stage of physiological needs was consistent with motivational theory (Maslow, Frager, Fadiman, McReynolds, & Cox, 1970).

Concerns about the stability and reliability of work were also mentioned. Working away was described as necessary for parents to ‘make ends meet’ and a secure income was compared with working locally:

Local work is not stable. You may have work today but you may lose it tomorrow. Working away is more stable and reliable. (Paternal grandmother 6)

Working away is better than working locally. Most of the times you have work today but perhaps you may lose it tomorrow. But this is not the case if you work away. They [LBC's parents] were unable to make ends meet when working locally. So they both migrate away. (Maternal grandmother 3)

In addition, some grandparents referred to problems in marital relationships as a motivation for migration. Wives and husbands from poor households sometimes fought with each other in the context of limited household resources. Relationship problems were in essence linked to household economic restrictions:

Life is not well off. And my son has no practical skills to earn money. They [son and daughter-in-law] used to fight over this. At last they decided to migrate away for work. You know, they did not get along well with each other at home as they could not make money. (Paternal grandmother 6)

They [LBC's parents] used to fight a lot when they were still at home as they had no money. Now they are away for work, I am not sure whether they still fight; at least, out of sight, out of mind. I need to care for kids for them. (Paternal grandmother 8)

One paternal grandfather who helped his daughter-in-law care for two left-behind girls indicated a non-economic motivations for migration:

He [LBC's father] got some mental problems. When he was around he often beat the kids. The girls were so scared of facing him. I just asked him to migrate away. (Paternal grandfather 2)

The parents’ need to migrate for work due to a lack of land to cultivate, or lack of employment opportunities, can be seen as a ‘push’ factor that drives the rural population to migrate to other places. On the other hand, more employment opportunities and higher salaries ‘pull’ them to migrate to developed regions (Cheng, Shi, & Ma, 2006).

Three carers mentioned that they were not concerned about survival. However, they had to pay contributions to ‘ren qing’ (social and moral norms of reciprocal favour, 人情) and ‘li jie’ (cultural etiquette, 礼节) in order to maintain social relationships in their villages, which caused extra pressure on household finances:

The contribution to one funeral used to cost much less than it is nowadays, no more than 10 RMB [1 GBP]. It is now around 100 RMB [10 GBP]. We were asked to give 500 RMB [50 GBP] for one relative's funeral. We had to do so just as others did. (Paternal grandmother 12)

Anyway what he [LBC's father] earned is enough for our survival. We have farmland to provide grains … You know, “ren qing li jie” [social and moral norms of reciprocal favour and cultural etiquette, 人情礼节] can cost a lot every year, at least 100 RMB [10 GBP] for attending just one funeral. Other families in this village are better off than us. They usually contribute several hundred RMB for weddings and funerals. We can do nothing but to follow others. This has made our situation even worse. (Mother 4)

In both these cases, participants experienced financial difficulties with making contributions to social events (weddings and funerals), but described them as important for social interactions in rural areas. It is likely that these expenses can reduce the household resources allocated to children. A longitudinal study from rural southern China suggests that prenatal exposures to social festivals tends to ‘squeeze’ resources allocated to food and nutrition for poor households, which appears to cause lower height growth in young boys (but not girls) under six years old (Chen & Zhang, 2012). Our data suggested expected contributions to social events are a source of comparison and judgement in the village, and so an additional financial pressure:

Other families in this village make contributions of one or two hundreds to social events [weddings and funerals], it would be inappropriate if you just contribute 50 RMB [5 GBP] as it could make yourself and your host [of the social events] lose face. (Paternal grandfather 2)

Migration to meet social obligations for sons

Another economic motive for migration was to earn and save money for their sons’ adults’ lives, especially marriage. This was prominent during the interviews with all the caregivers of left-behind boys, but not mentioned by caregivers of girls. This may suggest gender disparities in social obligations for boys and girls in the context of rural China, where sons are more culturally valued and receive more resources than daughters (Burgess & Zhuang, 2000; Ebenstein & Leung, 2010). Carers of boys referred to social expectations in rural areas that parents would take responsibilities for sons’ futures:

We have been in debt since we married off our two sons. First, weddings for sons, then feasts for celebrating the new-born grandchildren - both are our responsibilities. (Paternal grandmother 6)

It was worth noting that seven out of nine left-behind girls also had one brother in their family (Table 1). Girls, according to their caregivers’ accounts, were not one of the reasons why parents migrated. Instead, meeting social obligation for sons’ adult lives appeared to be a prominent rationale. This might suggest a distinct gender disparity in investments (referring to quantity and quality of household resources in this study) which favour boys rather than girls within the same households in rural China.

It is not uncommon in traditional patrilineal family system in China for sons to receive more monetary contributions from their parents than daughters. This may be due to the cultural expectation that sons should take the major responsibility for continuing the family line and supporting older parents, while daughters, especially married daughters, do not (Lei, 2013). It has been argued that traditional family values such as intergenerational support and filial piety may have weakened due to China's modernisation and population control campaign (Sheng & Settles, 2006). However, in the vast rural areas where the one-child policy is relatively flexible, the influence of traditional family lineage is still powerful (Liguo, Yingyan, & Xuan, 2014).

Having a son was considered to be associated with enormous expenses in rural areas:

It would cost hundreds of thousand RMB for a boy, like buying him a house, getting him a wife. This is the reality in rural areas. (Maternal grandfather 1)

Parents started working hard to accumulate wealth even when their sons were still young; for example, one paternal grandmother said her son and daughter-in-law started to migrate and save money when their son was only three years old:

They [LBC's parents] are worried about buying house for their son even though he is just three years old. They both could save around 30,000 RMB [3,000 GBP] per year. In this case, it will take about 10 years for them to buy a house. You know, currently the house costs hundreds of thousands RMB, even for the house in the town costing 200,000 RMB [20, 000 GBP]. No one knows where the rate of house will go in ten years. (Paternal grandmother 6)

Evidence suggests that Chinese parents with sons, particularly in rural areas, increase their savings due to the existence of unbalanced sex ratios, in order to improve their sons’ relative attractiveness for marriage (Wei & Zhang, 2009).

Having more than one son could make life even harder. Therefore, having a second son, for some families, served an additional push to parents to ‘migrate again’:

I have been talking to my sister all the time, “you do not have the ability to make money then how come you still give birth to babies, in particularly baby boys - the boys make life even harder”. So may God let the rich have more boys and the poor have fewer. But it does not work this way, does it? (Aunt 1)

He [LBC's father] used to work away when our first boy was around five to six years old. When we had our second child, another boy again, so he had to migrate again to make more money. Two boys, how come he stays at home instead of working away? (Mother 1)

In another case, parents were concerned about how they could afford two houses for their two sons. Therefore, both parents decided to migrate to make money:

His [LBC's] father said they [LBC's parents] had to make money for buying sons’ houses. They migrated away for their sons’ sake. How much money would it cost for two boys, you see? They have been worried sick about this: when will they be able to make enough money for houses? (Paternal grandmother 3)

At least, ‘migration especially both parents away make it possible to buy houses for their sons’. (Paternal grandmother 6)

It was surprising that children's education did not emerge as an important driver for parental migration in rural areas, as was initially expected. Three caregivers (one paternal grandfather, one maternal grandfather, and one uncle) expressed their expectations of LBC's education as a way to secure a better future than their parents’ generation:

You (LBC) need to work hard at school. Otherwise you will be just like your parents. You (LBC) will follow your parents’ paths to become a migrant worker in cities … Education is the only way to change your life. Even if you do not get into college finally, you would probably find a better job as a migrant worker in the cities when you are literate. (Maternal grandfather 1)

Education was an important concern but education per se was not considered as a primary driver of parental migration. This may be partly because primary and secondary education is free and compulsory in China. Caregivers, especially grandparents, did not express concerns about future educational expenses.

Taking on care-giving roles for left-behind children

Paternal grandparents as the first choice

Most of the 12 children with both parents migrating were left in the care of paternal grandparents (N = 10). The remaining two were only left with maternal grandparents as the paternal grandparents were not able to provide help with childcare. A six-year-old boy was originally left in the care of his paternal grandmother but transferred to his maternal grandmother due to her health problems:

He [LBC] was cared for by his paternal grandma when he was from one to four years old. Unfortunately, his paternal grandma got hepatitis B. For his health sake, my daughter and son-in-law took him along when they migrated. However, they both worked and could not manage to look after him. My daughter told me her sufferings and requested me to help her. (Maternal grandmother 2)

This maternal grandmother described how the care-giving roles were transferred. She gave priority to looking after her paternal grandchildren. Her daughters-in-law were an important consideration when she made the decision to also look after her maternal grandson:

You know, I have my own grandchildren [paternal grandchildren] to look after. My sons may have no problem if I look after my maternal grandson but the problem is with daughters-in-law who might not be alright with this. A proverb goes, “Daughters-in-law are not close enough with mothers-in-law from North China to South China (从南京到北京儿媳妇不跟老婆婆亲)”. So I just told my daughter that I can help her but I cannot guarantee how long I can do this for her. (Maternal grandmother 2)

Another left-behind boy and his younger brother, who were cared for by his maternal grandparents, were expected to be looked after by their paternal grandparents in the first place:

If his [LBC's] parents are away just for one or two days, perhaps he can stay with his paternal uncles. But the thing is his parents are away all the year around. In this case, he can stay with his paternal grandmother who is old and ill though. He complained that his paternal grandma's cooking was dirty. So he ended up staying with us. (Maternal grandmother 1)

Four mothers with LBC directly referred to paternal grandparents when asked about care-giving roles of LBC. Two of them mentioned that the paternal grandparents were too old to provide childcare. Another two indicated that the paternal grandparents refused to provide help due to family conflicts:

He [LBC's paternal grandfather] never helped out … He used to say, “We [LBC's paternal grandmother and paternal grandfather] would starve to death if we depend on you [LBC's parents] for elderly care”. In fact, we had three kids and were trying hard to make ends meet for survival, how could we have extra money for you? He even locked the door from the inside once when my kids approached him for food. How could you do this since they are your grandchildren by blood? (Mother 2)

This account suggested that paternal grandparents expected future returns, in the form of elder care from their adult children, when they considered whether to provide childcare or not. This suggests that intergenerational support between grandparents and their adult children in rural China is not a one-way street but follows a bidirectional pattern that resembles a long-term exchange of resources (Li, Feldman, & Jin, 2004; Zhang, 2005).

Three children with only mothers migrating were left behind with fathers who cohabited with the LBC's paternal grandparents. In fact, paternal grandparents were the primary care-givers for the LBC. When the paternal grandparents were too old and/or their health was poor, the maternal grandparents or other relatives, such as an uncle or aunt, often assumed their caring roles instead. For example, a 12-year left-behind boy whose paternal and maternal grandparents were too old to provide help was looked after by his maternal aunt.

Caring for left-behind grandchildren as social obligation

Care-giving practices often involved preparing food, washing clothes for LBC, sending young LBC to school and collecting them after class. Some carers indicated that ‘looking after them [LBC] is very demanding’ as they had to care for more than one child. In two cases, grandparents needed to look after four LBC; ‘it is just like running a kindergarten’. On the other hand, two paternal grandmothers expressed their willingness and pleasure in providing childcare. Otherwise, they would ‘feel lonely’ without the LBC's companionship.

Most grandparents indicated strong feelings of responsibility and obligation towards caring for left-behind children. Our interviews suggested that caring for paternal grandchildren was a social expectation on grandparents:

Daughters-in-law are expected to give birth to children and paternal grandmother is responsible for rearing them, which is quite true in this village or any other village. They [migrant parents] assigned you a task to care for the kids. So it becomes our responsibility to look after them. What can you do about it? (Paternal grandmother 12)

We are doing this [looking after the two left-behind boys] just for our kids’ sake even though we are quite old. What are we going to do? My own paternal grandchildren and my own maternal grandchildren, how come you don't look after them? (Maternal grandfather 1)

These narratives showed strong culture-based responsibility and obligations for grandparents, in particularly paternal grandparents, to help their adult children look after grandchildren. They are the most common caregivers and often the first choice to take on care-giving responsibilities for LBC when both parents are away. Caring for grandchildren in China is considered to be the cultural norm, and thus expected (Burnette et al., 2013). This study found that paternal grandparents in particular might feel obliged to take on care-giving roles. Although they found caring for LBC ‘not easy’ and ‘very demanding’, especially for ageing grandparents or those in poor health, they still encouraged parents to migrate in order to ‘make ends meet’. Adult sons and daughters may be motivated to take up migrant work to improve the household's economic status when parents are available to look after their young children. Those who are left behind in the extended family constitute ‘a reliable and effective support system’ (Chang et al., 2011). These findings are also consistent with the group/mutual aid model, in which each household member's capacity is maximised to ensure the success of the household as a whole (Lee & Xiao, 1998).

In our interviews, one paternal grandmother refused to look after LBC as she claimed that her son was too poor to offer her elder care when she became older. She had more than one son and she chose to look after the children of her second son who was financially better off than her first son. To some extent, grandparents were not always passive recipients and sometimes they made their own decision on whether to provide childcare, or which child they would care for. Therefore, providing childcare, under some circumstances, can be seen as a strategic investment made by grandparents in the hope of maximising the return they will receive in the future (Brown & Poirine, 2005; Cong & Silverstein, 2011; Silverstein, Conroy, Wang, Giarrusso, & Bengtson, 2002; Sun, 2002).

Remittances from migrant parent(s) to non-parental caregivers

In this study, the remittances mainly refer to monetary remittances (for example, money and in-kind goods) rather than social remittances such as ideas, norms, and practices gained through migration and passed on to household members who stay behind (Levitt, 1998). Participants’ accounts gave the impression that the remittances could be small and irregular, in particular for paternal grandparents acting as caregivers. Ten out of 13 paternal grandparents said they did not receive remittances from LBC's migrant parents (see Table 3). Two of three paternal grandparents who did receive remittances described the amounts as inadequate to cover all the LBC's expenses:

My son gives me money irregularly when he gets back home. He asks me to use this money for buying clothes, shoes, food and drinks for them [LBC]. When the money is running out, I spend my own. You know they are my own grandchildren, how can I not love them? (Paternal grandmother 3)

The other day I spoke to my son, “you think you give me 2000-3000 RMB [200-300 GBP] a year, but you have no idea how much money I need to spend, around 10,000 RMB [1,000 GBP]”. He seemed to be all right with it. I almost spend all my income covering the outstanding expenses. (Paternal grandmother 4)

TABLE 3 Remittances from migrant parent(s) to non-parental caregivers (N=16).

	Remittances
	Yes	No	Total
Gender
Boy	4	8	12
Girl	2	2	4
Age range
6–9 years	2	4	6
10–12 years	2	8	10
Caregivers
Paternal grandparent(s)	3	10	13
Maternal grandparent(s)	3	0	3
Other relatives (Aunt)	0	1	1

Two paternal grandmothers showed understanding when their adult children failed to send remittances:

They never gave me money or buy me any clothes. They don't have any money for me since their small business went broke. My son would be able to work in construction fields to earn money if he is in good health condition. They are so poor. I never ask them for money. I don't want to see them suffer a lot. (Paternal grandmother 12)

They [LBC's parents] have no money at all. My daughter-in-law has many siblings so they need to pay many contributions, like weddings, celebrating new-born babies, and so on. Last year I fell down and was sent to hospital. They borrowed money to pay for my medical expenses. So they don't have money. (Paternal grandmother 10)

However, all three maternal grandparent carers said they received remittances from migrant parent(s) (see Table 4) and two described it as ‘enough for the kids’. One left-behind girl and her little brother were cared for mainly by their paternal grandparents and occasionally by their maternal grandparents. Only the maternal grandparents received remittances:

My daughter sent me money to pay tuition fees for my little grandson. You know he is in pre-school and it is not free. And the rest of money is used to buy some food and snacks for children. It is enough for the kids. (Maternal grandmother 3)

TABLE 4 Remittances distribution for non-parental caregivers who received remittances.

Caregivers	Sent remittances or not^a	Amount^b	Enough for LBC
Maternal grandfather 1	yes	Not Specific	-
Maternal grandmother 2	yes, quarterly	About 1000 RMB (100 GBP)	Yes
Maternal grandmother 3	yes	Not Specific	Yes
Paternal grandmother 3	yes, irregularly	Not Specific	No
Paternal grandmother 4	yes	2000–3000 RMB (200–300 GBP)	No
Paternal grandfather 1	only Chinese New Year	Not Specific	-

LBC, left-behind children; RMB, renminbi (the official currency of People's Republic of China); GBP, Pound Sterling.

^aHave his/her parents send remittances back?

^bHow much money roughly has been sent back to you per year?

Although financial returns could cover the LBC's own expenses, we were unable to ascertain whether maternal grandparents received additional compensation for their help with childcare. Given the small sample of grandparents (13 on the paternal side and three on the maternal side), it is not possible to draw a definitive conclusion that maternal grandparents are more likely to be reciprocated financially than parental grandparents. Our findings, however, may shed light on the gender disparities in providing support for parents.

Most paternal grandparents in our study reported that they did not receive regular and adequate remittances from their adult children. The remittances appeared to only cover the LBC's own expenses and did not involve additional compensation for the paternal grandparents’ help. Maternal grandparents reported that they received remittances from their adult daughters which were considered to be ‘enough’ for the LBC. This is consistent with Song, Li, and Feldman's (2012) study suggesting that daughters who received support from their older parents (for example, grandchild care) tended to return more financially than sons. Normally, full-time childcare for maternal grandchildren is rare in rural China, where strong patrilineal family values are embedded and family resources are less likely to be expected to be allocated to married daughters (Chen, Short, & Entwisle, 2000). When childcare is provided for adult daughters, greater expectations of financial returns are expected from them by grandparents (Cong & Silverstein, 2012). Whyte and Xu (2003) have observed a general pattern in provision of support for older parents, in that married daughters tend to provide greater support for ageing parents than married sons. It is possible that even without childcare help, maternal grandparents may still be expected to receive financial assistance from their daughters.

Another possible justification for sons rarely making financial support is due to the ‘long-term contract’ between sons and older parents (Song et al., 2012). The care of the aged is primarily a family responsibility in China (Davis & Harrell, 1993), especially in rural areas without a universal social security system (Lei, 2013). Sons in a traditional family are morally and legally expected to provide support for their parents especially in later stages of life (Ebenstein & Leung, 2010; Whyte & Ikels, 2004). In this case, it is possible that sons will reciprocate, probably in a different way (for example, offering elder care), at a later stage. This life course perspective of reciprocal transfer between inter-generations; that is, ‘earlier investment in the form of future support’ (Thang & Mehta, 2012), may explain irregular and inadequate remittances to grandparents for childcare from adult sons.

It should be noted that the different patterns of intergenerational contracts observed in this study can be applied to daughters as well as sons. The inherent difference may be that contracts with daughters tend to be short-term, while contracts with sons are long-term and may persist into the course of later life (Antonucci, Akiyama, & Birditt, 2004; Yang, 1996). Migration is partly the outcome of meeting social responsibilities for offspring especially sons in rural China. Grandparents, particularly on the paternal side, are expected to fulfil their social obligations to care for left-behind grandchildren even without immediate financial returns. These intergenerational obligations may have inadvertently strengthened family bonds through generations in rural China.

Conclusion

Although the sample is from only one township in rural China, the findings in this study are consistent with those in previous studies in other parts of the country, which found that economic incentives are an essential motivation for internal migration (Cheng et al., 2006; Fan, 2005; Hu, 2012b; Ye et al., 2013). This study sheds further light on some underlying socio-cultural explanations pertaining to economic motives; for example, making contributions to social events (weddings and funerals) in village life, and fulfilling social obligations for left-behind sons’, rather than daughters’, futures. The latter appears to be an important driver in motivating both parents to migrate. When both parents are away, paternal grandparents are often expected to be the first choice for taking on care-giving responsibilities for LBC. Regular and sufficient financial returns are not always available to paternal grandparents. Despite the small size (13 paternal grandparents, three maternal grandparents), the findings of this study tentatively suggest strong culture-based obligations for paternal grandparents to help their adult children take care of LBC.

Despite significant socioeconomic changes in contemporary China, family obligations continue to play an important role in Chinese family ties (Qi, 2015). At least two distinctive flows of intergenerational obligations within a split household due to migration can be observed. One is upwards from grandparents to their adult children, via the provision of childcare for LBC. The other is downwards from migrant parents to LBC, by saving up for their young sons’ adult lives. Preferences for sons, though rooted in traditional Chinese culture, are still prevalent in contemporary rural China (Ebenstein & Leung, 2010). Sons are considered more valuable and often allocated more household resources than daughters (Davis & Harrell, 1993). Chinese parents with sons increase their savings due to the existence of unbalanced sex ratios, in order to improve their sons’ relative attractiveness for marriage, which is particularly true in rural areas (Wei & Zhang, 2009).

The findings of this study suggest that the gender of the LBC can be an important consideration for parental migration in the context of rural China: parents of boys have the additional motivation of migrating in order to save for their sons’ futures. This suggests a paradox of intergenerational obligations, in that left-behind boys may be less likely to receive resources in the form of remittances from migrant parents in a society where sons are culturally more valued than daughters. This is because parents of boys can be more motivated to migrate (than parents of girls) and they tend to save up for their sons’, rather than their daughters’, adult lives. Grandparents, particularly on the paternal side, are expected to fulfill social obligations to care for left-behind grandchildren even without immediate financial returns. Children left in the care of paternal grandparents may end up with fewer resources from their migrant parents. Left-behind boys cared for by paternal grandparents may receive even fewer resources in their early childhood from migrant parents than LBC cared for by maternal grandparents. This lack of resources in early childhood potentially makes left-behind boys at risk of slower growth and development during a critical period of the life course. Moreover, our findings on the gender differences of LBC in the social obligations and remittances from migrant parents, have an important policy implication in the context of rural China, where son preferences are prevalent (Burgess & Zhuang, 2000). Future studies and policy-making on LBC's well-being should therefore consider the variables of the gender of LBC, their relationships with caregivers and the number of boys within families with migrant parent(s).

Several limitations were encountered in this study. We explored the motivations for parental migration and care-giving arrangements from caregivers’ perspectives rather than migrant parents directly. While this enabled us to discover what is important in making migration decisions with the perspective of an outsider, we were unable to explore migrants’ views on their motivations for migration and their care-giving arrangements for LBC. We found that income was a sensitive issue and most participants were reluctant to disclose their incomes. Although some caregivers did volunteer information about remittances from migrant parents and whether the remittances were sufficient for the LBC, we were unable to compare the amount of remittances received by parental grandparents and maternal grandparents. In addition, the participants may not be representative of the community of caregivers for LBC in the area in terms of socio-demographics or views. However, diversity was maximised within the context based on children's characteristics (for example, age, gender, types of parental migration), and negative and discrepant cases analysed (Maxwell, 2012) until theoretical saturation occurred (Charmaz, 2006). Future studies may be able to access children and caregivers with various socio-economic characteristics and elicit further insights from migrant parents directly.

Acknowledgements

We are grateful to all children and their caregivers who gave their time and patience to participate in this study.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Economic and Social Research Council (ESRC) Postgraduate Scholarship and ESRC Overseas Fieldwork funding.