The growing sex ratio at birth (SRB) imbalance and disproportionately high mortality rates among girls in China have stirred extensive scholarly interest in the phenomenon of son preference in recent years.Footnote 1 There are several practices which underpin son preference, one of the most central being patrilocal marriage patterns where women marry into the family of their husband and settle with or nearby their parents-in-law, giving rise to the notion that women lose their belonging to their natal families upon marriage.Footnote 2 Subsumed into patrilocality is filial piety and old-age security arrangements, where sons are expected to provide for old-age care and support.Footnote 3 Relatedly, sons are wanted for their ability to pass on the family line and for their religious and symbolic functions associated with ancestor worship.Footnote 4 These functions have been shaped and established through the intergenerational contract, which is markedly gendered and implicates roles and responsibilities for various family members.Footnote 5 Historically, the intergenerational contract has mostly been a son–parent contract, where sons have been viewed as “unsubstitutable” and primary, and daughters as substitutable and secondary.Footnote 6
Another key factor underpinning son preference is the idea that male manual labour in agriculture is essential, not substitutable and of primary concern for the survival of rural families.Footnote 7 This “heavy labour argument” grew stronger with the household responsibility system (HRS) introduced in the mid-1980s, which granted individual land-use rights to rural families.Footnote 8 The HRS reinforced not only land-use rights being passed down the paternal side of the familyFootnote 9 but also patrilocal marriage patterns.Footnote 10 However, as others have noted, the idea that rural families need sons for their manual labour capacity seems to be at odds with the fact that millions of people are migrating from the countryside to the cities.Footnote 11 The exodus of labour from rural areas, along with migration remittances and the transmission of new ideas and practices from migrants, has had a significant impact on political, economic and social institutions in rural China. Yet, the implications of such changes for son preference remain largely unstudied.
So far, the link between migration and son preference has mostly concerned the impact of migration on rural migrants living in the city.Footnote 12 Based on quantitative data, those studies have focused on the relationship between migration and son preference attitudes, as well as the result of son preference in terms of fertility behaviour. Another study established that women's wage labour, as well as women's broadening experience of the outside world, was associated with a reduced preference for sons in rural areas, suggesting that women's labour migration may attenuate son preference.Footnote 13 Similarly, a recent study found that the lower the ratio of women to men who migrate, the stronger the son preference.Footnote 14 However, the same study also found that son preference is positively associated with the proportion of family members who migrate, implying that both migration and securing the birth of at least one son are strategies to strengthen the patrilineal family. No study to date has made a qualitative investigation into how son preference is affected by migration in rural China. Moreover, few studies have adopted a causal approach, i.e. studying how practices known for instigating son preference become reconfigured through migration.Footnote 15
The purpose of this study is to explore how two practices known for upholding son preference are affected by rural–urban out-migration. First, it will examine the impact of migration on the gender division of labour in agriculture, and how and to what extent migration challenges the “heavy labour argument.” Second, it will address the impact of migration on patrilocality, i.e. the custom that newly married couples settle with or near the family of the groom. Inspired by Croll, I will argue that in order to understand how shifting practices may contribute to reconfiguring son preference in rural China, it is necessary to pay attention to the intergenerational contract and the gender configurations of that contract, particularly with regards to the substitutability of adult sons and daughters vis-à-vis their parents.Footnote 16
Methodology
The study is based on observations, household interviews and in-depth interviews with women and men of different ages in rural China. Interviews were also conducted with local leaders at township and village levels in the field sites described below. As a complement to household and individual interviews, time-use data was gathered to assess gender and generational disparities in the divisions of labour.Footnote 17 The time-use tables were completed retrospectively, based on the interviewees' recall of time use during a typical day and week. The fieldwork was conducted during 22 days spread over three periods in May, June and July 2007. The interviews lasted on average just over an hour.
The field sites
The article builds on data collected in four villages in Anhui province. Anhui lends itself well to exploring how practices known for underpinning son preference in rural China become reconfigured through migration for several reasons: 80 per cent of the population is registered as rural;Footnote 18 there are high rates of rural–urban migration; the SRB was 132 in 2005;Footnote 19 and the infant mortality rate for boys was 26.8 compared to 42.1 for girls in 2000,Footnote 20 suggesting that son preference is strong.
The arable land belonging to the four villages is all flat. Two of the four villages, Village A and Village B, are located in County X in northern Anhui. No village-level data was made available, but my interview data suggest that households were allocated, on average, 1.5 mu 亩 of land per capita.Footnote 21 Households relied heavily on farming, as well as migration – 28 per cent of the residents were out-migrants in County X, 46 per cent of which were women.
The other two villages, Village C and Village D, are situated in the rural areas of County Y, which was recently made into a semi-urban district of a major city in eastern Anhui. Until 2002, residents in these villages were allocated, on average, one mu of land per capita, but owing to rapid industrial development, many households lost most of their land. At the time of my investigation, residents in Village C had, on average, just under 0.5 mu of land per capita. There were several industries within the village, and 21 per cent of residents were out-migrants, 45 per cent of which were women. In Village D hardly any households had land, despite the fact that the village was labelled as rural and its residents mostly held rural residence status. Many factories had been set up in the village, and only 3 per cent of the villagers were out-migrants.
Sampling and interviewee characteristics
In each of the four villages, 12 households were interviewed, which makes a total of 48 households.Footnote 22 The households were selected through purposive sampling in order to cover the broadest possible range of different household types. In addition to gathering information on issues related to divisions of labour, the household interviews focused on the composition of the family, main sources of income, land usage, remittances, local childbearing and marriage customs, care arrangements, family traditions and customs, and the migration history of family members.
In most households, two or more family members were interviewed independently. In total, 85 such interviews were conducted. Sixty-one per cent of the interviewees were women and 39 per cent were men. The contents of the in-depth interviews overlapped with the household interviews but with more emphasis on personal experiences and attitudes, including interpersonal relations with other household members.
Data analysis
With a few exceptions, all the interviews were recorded. The recorded material was subsequently transcribed in Chinese. The written material was then reviewed several times and, as different themes and issues emerged, the material was coded while notes were taken. The coding and the notes were written down in English and particular quotes were translated into English.Footnote 23 In the final data analysis stage, the different themes and issues were compared to look for various patterns, differences and similarities. This was done manually.
Out of the 48 households interviewed, 28 were migrant households,Footnote 24 here defined as households having at least one member migrating.Footnote 25 Eight households had members who were returned migrants, and 12 households were non-migrant households. In other words, migration was, or had been, an important livelihood strategy for the majority of the households interviewed. The various types of migrant households were categorized into different ideal types for the sake of providing conceptual clarity, and these types are used throughout the analysis.Footnote 26
Rural Households in Transition
When looking at the household constellations of the 28 migrant households, it became clear that they differed from the migrant households commonly found in the literature on rural–urban migration in China, namely the “left-behind” woman household and the “lonely” or “empty nester” older person household.Footnote 27 As illustrated in Figure 1, there are six “ideal types” of migrant households, depending on generation, sex and main livelihood strategy, and they were all formed patrilocally.Footnote 28
Ideal type 1, male migrant household, denotes a household where the man of the second generation migrates and the woman of the same generation stays in the village and continues to farm the land. Typically, she lives together with her parents-in-law and children. This type of household is often associated with the feminization of agriculture, which is a process some studies have found to be ongoing to various degrees in China.Footnote 29
Figure 1: Ideal Types of Migrant Households by Generation,* Sex,** and Main Livelihood Strategy***
Ideal type 2 refers to an older peasant household, where the older members farm the land, while both the woman and the man of the second generation live and work outside the village. Often, the older generation looks after the grandchildren.
Ideal type 3 represents an older person household, which resembles ideal type 2 in the sense that both the woman and the man of the adult generation live and work outside the village. However, a major difference is that the older generation in ideal type 3 had withdrawn from agriculture.
Ideal type 4 denotes a housewife household, where the woman of the second generation has withdrawn from agriculture and off-farm work and lives off her husband's income. Sometimes, this type of household also includes the older generation. The adult woman in these families is mainly engaged in home maintenance and the children's schooling. A similar trait in the older person household and housewife household is that the household members who stay in the village are economically dependent on remittances. In a way, these households are “migrant-headed households,” as the migrants are in charge of the financial flows in the household, setting standards for expenditure and living conditions from a distance. This finding echoes previous research, which has shown that decision-making structures are hard to alter, and even though men may spend the vast majority of their time in the city, they still remain the head of the household and make essential decisions concerning longer term strategic and financial issues.Footnote 30
Ideal type 5 refers to a youth migrant household, where both the adult man and woman stay in the village, while the younger generation migrates before getting married.
Ideal type 6 is constituted by a female migrant household, where the woman of the second generation migrates, while her husband stays in the village. Owing to patrilocal residency patterns, this often means that the adult son stays behind with his parents and children.
New Divisions of Labour
The notion that son preference stems from the need for male labour for agricultural work was strong and repeatedly mentioned, in particular by the older generation.Footnote 31 However, as detailed below, the loss of male farm labour through migration and the gain in income through remittances had major implications for the gender and intergenerational division of labour in agriculture. The ways in which new divisions of labour took shape differed widely among the different ideal types of households.
Feminization of agriculture?
The concept of the feminization of agriculture suggests that women replace men as men leave the agricultural sector. Indeed, during my fieldwork I observed how women sometimes undertook farm work, such as distributing pesticide, a task considered to be “men's work.” However, this did not mean that the feminization of agriculture was something universal. Contrary to findings from rural Guizhou of left-behind Miao women who experienced a heavy burden from intensified agricultural work,Footnote 32 in this study only a few women in male migrant households complained of a heavy workload. Similarly, a study by Ye and Wu of left-behind women found that more women reduced rather than increased the size of their arable land (30 per cent and 10 per cent, respectively).Footnote 33
Rather, there were two dominant characteristics with regard to the engagement of adult women in agriculture. First, time-use data revealed that their involvement was limited both in terms of actual time spent and tasks performed. However, when probing into the everyday activities of the different household members, it became clear that some tasks, such as weeding and harvesting vegetables, were not considered farm work (zhongdi 种地 or nonghuo 农活) by the interviewees. Hence, an important aspect of women's engagement in agriculture was at first hidden and it appears as if agricultural work has become incorporated into the domestic sphere and is therefore not so clearly perceived as “work.”Footnote 34
Still, even after taking into consideration the under-reporting of farm work, women's engagement was, in general, limited. This was exemplified in the household of Ling, a woman in her late 20s. Ling's household held 2.6 mu of land, which was used to grow rice and rape. Ling and her mother-in-law undertook the farm work but Ling's main function was to look after her baby son. The fact that Ling's husband had migrated did not mean that she substituted herself for her husband, who never was engaged in farming. Moreover, from the point of view of Ling's economic status in her household, male (off-farm) labour was more important in terms of income than female (farm) labour. This observation is in line with recent research showing that, whereas migration remittances continue to rise, income generated from farming is lagging behind.Footnote 35
The second characteristic with regard to the involvement of adult women in agriculture was that their involvement was fluid, meaning that they were more engaged in agriculture during some periods than during others. However, these periods did not necessarily coincide with the peak seasons in agricultural work, but rather depended on to what extent children needed motherly care. Xiu, a young mother with a toddler, lived with her in-laws while her husband worked in Shanghai. She was ready to leave the village as soon as her son was old enough:
The child will be left back home. I want to work because the economic situation in rural areas is bad. So, the old people can look after the child and we can be outside and earn money. Because already now we have to spend money on the child, and when he is even older we will have to spend a lot of money on him, so we'd better go out and earn some money.
When I visited the same household two months later, Xiu had left and the household, which during my first visit had been a male migrant household, had become an older peasant household.
Aging of the rural workforce population
As the presentation of the six ideal types shows, the vacuum that arises when men – and, to some extent, women – leave farm work is often filled by older men and women (as in older peasant households, and to some degree in male migrant households). This has contributed to a shift in the division of labour, from the adult and older generations working together in the field to the older generation taking on an increasing share of agricultural work.Footnote 36 According to one study from 2008, the rate of involvement of “left-behind” older persons in farm work increased from 10.4 to 14.4 per cent.Footnote 37
Moreover, time-use data revealed that whereas the older generation stepped in and took on an increasing role in farm work and household work, there were important gender differences. Older men did not always engage more in farm work and household activities when adult men withdrew from agriculture. The limited role of older men in agriculture and in household work can be explained by the fact that older men tend to take local ad hoc jobs to contribute to the household income.Footnote 38
Youth in agriculture
One would assume that if male manual labour in family farming was unsubstitutable, young males in their late teens would be active in the fields in their spare time, particularly during summer holidays. However, male youths were largely absent when it came to farming activities in the villages I visited, even during peak seasons. When interviewed, none of the youths indicated that they would like to do farm work. In fact, farming has become associated with something “anti-modern” and belonging to the past. Rather, education is seen as “the way out” of having agriculture as the main livelihood strategy, and most children are encouraged – not least by their migrating parents – to study hard and devote their spare time to homework.
According to Grandpa Song, an older man belonging to an older peasant household who looked after five grandchildren:
Except for some small household work, they [the grandchildren] do not do anything in the house. I would not let them farm the land. If they helped with farming the land and came back with poor results in school, then they would say that it was because of the time they spent in the fields, and I would be responsible. So, I don't want to take the risk of harming their studies. Then the parents would blame me.
When interviewing 12-year-old Little Song, Grandpa Song's eldest granddaughter, there appeared another dimension to the division of labour. Even though Little Song was not expected to work in the fields, she was indeed expected to do household chores. When asked about her time use during a normal week, Little Song explained how she went to a school 20 minutes away by foot, and came home most days at lunchtime to cook for her grandparents and two younger cousins. As well as cooking, she washed the dishes after lunch almost every day and once or twice a week after dinner for the whole family, which also included two older male cousins. She also cleaned the floor in the bedroom and the living room every morning. When comparing time-use data from her male cousin, who was two years older, it became obvious that Little Song did more of the household work. Little Song explained that, “Grandma normally won't ask them [male cousins] to do the dishes and wash the clothes, because they are boys.” Even though Little Song was not involved in farming, it is important to point out that her household work gave her grandparents more time to engage in farming. Hence, migration has an indirect impact on the time use of girls and boys, and this seemed particularly obvious in older peasant households. The fact that the children of migrants do more household work than children in non-migrant families was also pointed out by several of the young people I interviewed.
The “unsubstitutability” of daughters?
Common to all household constellations, except those that were no longer engaged in farming, was that adult daughters – both married and unmarried – had important roles to play in some parts of the agricultural cycle, particularly in transplanting rice seedlings. This was, however, not a matter of substituting for male labour. Rather, this work had been handled by females for generations, although not always acknowledged as “heavy work” and therefore not considered “unsubstitutable.” Still, the nature of transplanting rice seedlings, which involves bending over for a long period of time, could by some standards be considered as “heavy work.”
One of the interviewees, a young woman of 19 called Ping who belonged to a youth migrant household and who had two older migrating brothers, exemplified the fact that daughters have important roles to play in agriculture. She explained that she did almost everything in the field, apart from “heavy work.”
When it comes to farming the land, I do basically everything apart from heavy work … I transplant rice seedlings from the seedling bed to the field, I grow cotton, pick the beans … Basically, the work that needs to be done at home is this. And then I harvest the wheat and cut the rice.
Moreover, the work of transplanting rice seedlings was not only performed by women belonging to the patrilineal family. Grown-up daughters, who according to patrilocal residence patterns had left their natal families upon marriage, also returned to help their brothers or parents with the work of transplanting rice seedlings.
This situation was illustrated in Grandma Liu's family. Grandma Liu belonged to an older person household and had five grown-up children, two daughters and three sons. She made it clear that sons were important for their physical strength, the care they provided for older parents, and for their role in carrying on the family line. However, when asked to describe the type of work and support her children provided, it became clear that, in reality, the situation diverged quite markedly from her idea of the benefits of physical strength that having sons brought. Grandma Liu formally formed a household with her second son, Lao Er. However, since he and his wife were both migrants, she lived with two of their three children. Her third son had migrated with his whole family and Grandma Liu's eldest son, Lao Da, formed a female migrant household and was farming his and his two brothers' land. Since all three of Grandma Liu's daughters-in-law were migrants, the eldest son received help with transplanting rice seedlings and harvesting the rice from his two sisters instead.
Housewifization?
In some instances, the transfer of men into off-farm work also opened up the possibility for women to withdraw from agriculture altogether or in part. Although it is common practice for migrating women to return to the village during pregnancy and for childbirth, not many of the women who had never migrated or who had not re-migrated after childbirth played a great role in agriculture, such as in the housewife household. Several female interviewees of the adult generation also indicated that farm work is tiring and dirty. Hence, in parallel to male manual labour becoming less essential to family farming, there seems to be a trend of ascribing farm work less value and status. Another oft-used explanation for these women's low engagement in farm work was that their parents or parents-in-law were considered more skilled at farming. There was also a lack of rural childcare facilities, which meant that it was hard for women to combine farm or off-farm work with caring for young children.
Moreover, some of the women from housewife households did not seem to think that income-generating work was important to them. In Village D, many households no longer had any land, but rather than migrating or finding off-farm work locally, some women chose to stay at home. Most of them did not consider themselves to be out of a job but felt that their main task was to look after the children. This attitude was particularly common among women who had a son. The fact that women with sons are less likely to migrate has also been noted elsewhere.Footnote 39 Some of the housewives even expressed frustration that their husbands did not have better jobs with high enough incomes.
Commercialization of Agriculture
In line with previous studies, the results from this investigation suggest that remittances are an important source for investment in agriculture.Footnote 40 The possession of cash through remittances means that employing labour on a temporary basis to do intensive and hazardous farm work – tasks which traditionally were performed manually by adult males – has become an option for many rural families. In Village A and Village B, almost all households in the process of harvesting wheat engaged in commercialized farming to different degrees. Some households paid day labourers to harvest the wheat manually, and others commissioned local entrepreneurs to harvest the wheat mechanically.
The increase in the number of harvester machines was made possible thanks to migration remittances, which both directly and indirectly gave several households the purchasing power to buy the machines. In Village B, ten households owned harvesters. One of these machines was owned by Han, a man in his early 40s who came from a non-migrant household. Han explained that he and his sister had borrowed 26,000 yuan each to buy the harvester three years earlier. Back then, they were the third family in the village to buy a harvester. As one village committee member explained:
Before, we collected the harvest by ourselves. Each household gathered friends and the whole family and we helped each other out. Now everybody rents day labourers to collect the harvest with machines. It is much quicker. It used to be only a few families with members migrating to the city that could afford it. Now almost everyone can.
In other words, the outsourcing of heavy agricultural work had started with the migrating families but had then spread to other families. The proliferation of harvesters was also made possible thanks to the affordability and availability of machinery.Footnote 41
Both migrant and non-migrant households had also invested in resources such as irrigation, fertilizers and pesticides to increase agricultural productivity. Decisions to purchase those types of agricultural input were often made collectively at village meetings. One potential disadvantage for the male migrant households was constrained access to information and decision-making processes in this regard. During my fieldwork, the village meeting discussing irrigation and water management I came across consisted solely of male villagers. Although the observation may not be representative of a general trend, it may indicate that “left-behind” women risk losing out on important inputs to increase farming productivity.
Consolidation of land holdings
In Village A and Village B, out-migration contributed to freeing up land. It was particularly the case that the older person households and housewife households did not use their land at all, except for growing vegetables for personal consumption. Instead, these households tended to sub-contract or lend their land to neighbours and relatives. If they were lucky, the plots were located next to the land of the family who leased it, which made it more viable to use machines. However, in reality, as the arable land had been distributed between households based on the quality and the location of the land, each household tended to have many small and scattered pieces of land. In the two villages in northern Anhui, some families had up to 20 different small parcels of land, some several kilometres apart. Therefore, the withdrawal from farming of some families only rarely led to the consolidation of land holdings.
Nevertheless, the opportunity to sub-contract migrant households' land made mechanized harvesting both more feasible and necessary: feasible in the sense that there was more to harvest, and necessary in the sense that there was insufficient manual labour. In one family, the head of an older peasant household, Old Xie, decided to lease the land of his migrating neighbours. In order to manage the land, which amounted to more than 20 mu, Old Xie worked with his eldest son and together they constituted a new type of rural worker that I call rural entrepreneurs. By engaging in larger scale farming, the father and son were able to generate a profit of more than 10,000 yuan a year, and this was enough for the son to get a loan and purchase a flat in the county city.
It is important to note that the way in which farmland is leased or lent tends to follow gendered patterns. For example, the older peasant household of the Song family, presented above, had 19 mu of land altogether (including the land belonging to the migrating sons) but they only farmed 6 mu. The remaining land was leased to Grandpa Song's nephews.
Hence, the mechanization and commercialization processes ongoing in rural China – fuelled by migration and remittances – give rise to new types of rural workers which are different from the family farmer. As suggested by Lanyan Chen, this process is highly gendered: women mostly work as day labourers on farmland managed by men who constitute a professionalized workforce of rural entrepreneurs making use of new technology, increasing efficiency and profitability in the agricultural sector.Footnote 42
Challenging Patrilocality?
Previous studies have shown that patrilocality is positively correlated with son preference.Footnote 43 Judging from data on marriage patterns in County Y, patrilocality has been reinforced in recent years.Footnote 44 In fact, the proportion of uxorilocal marriages has gone down, from 5.1 in 1995 to 3.4 in 2005. As Table 1 further reveals, another trend is that both women and men increasingly marry someone outside of their home county. However, this trend is substantially larger for women, and as women settling far away from their birth place have weaker social and family networks, patrilocality can be said to have increased.
Table 1: Marriage Patterns in County Y 1995–2000
Source:
Statistics provided by local authorities in County Y.
Although it is beyond the scope of this article to establish whether the increase in patrilocality is a result of migration, it is here assumed that there is indeed a correlation. However, when probing into the migration history of non-migrant households, as well as the household constellation of older person households, it became clear that there are also ways in which migration challenges patrilocality. As illustrated in Figure 2, migration gives rise to one additional type of migrant household compared to those presented in Figure 1, namely uxorilocal older person households.
Figure 2: Household Constellations Challenging Patriocality by Generation* and Sex**
Moreover, migration also gives rise to three types of non-migrant households, namely uxorilocal non-migrant households, empty nesters and neolocal households. These households used to be migrant households but transformed into non-migrant households as the migrating family member formed an independent household away from the village. The processes of these household transformations are outlined below.
The data revealed that some youth migrant households transform into uxorilocal non-migrant households. This was exemplified in one household, where the son had migrated for educational purposes and the daughter stayed at home with the parents. When interviewed about the family constellation, the daughter's mother, Grandma Qin, described how they had had a deliberate strategy to “recruit” a husband for their daughter and that she had already formed this plan when her daughter started dating her husband: “[The son's] results were good, you know. I was already thinking that I needed to prepare.” Although Grandma Qin's son migrated for educational purposes, the fact that he migrated out of the village created a situation in which the daughter was provided with the option of marrying a man who would move in with her and her parents. This challenged not only the “heavy labour argument” but also patrilocality. The fact that such large numbers of both male and female youth migrate to the cities means that families who have sons cannot assume that the sons will come back and remain in the village, take over the land and work on the farm.
In another uxorilocal household, which also used to be a youth migrant household, migration played an important but different role. Zhang, a woman in her mid-30s, met her husband when he was working as a temporary migrant worker in her village. Because the living conditions were better in Zhang's village, they decided to settle down there, despite the fact that her husband was the only son in his family. Both these cases exemplify how youth migrant households are vulnerable to the “risks” of losing their son and his manual labour, and how they adopt pragmatic solutions of alternative household formations which challenge the patrilocal norm.
There were other deviations from patrilocality in some older person households, where the grandparents lived together with their daughter's children, representing care arrangements following the matriline. This practice breaks markedly with the belief that the children of daughters do not count as grandchildren. This view was expressed by some older interviewees, who at first would not even mention their daughters and their children, as they did not belong to the patriline.Footnote 45 Interestingly, many grandchildren who lived with their maternal grandparents called their maternal grandmother nainai 奶奶, which is the Chinese term for paternal grandmother. The use of nainai possibly indicates that the grandchildren considered the relationship to be very close – as close as the relationship with a paternal grandmother – and therefore they did not use the term for maternal grandmother (waipo 外婆), which denotes that the relationship is outside the family in a patrilineal sense.
Patrilocality was also challenged through neolocal residence patterns, where migrants from the second generation formed separate nuclear-based households in the cities, detaching themselves from the first generations on both sides.Footnote 46 Whereas the formation of households in the city far away from home would not necessarily mean that intergenerational transfers of social and economic support between sons and their parents would cease, several older persons received no support whatsoever from their migrating sons. For example, Grandma Bu and Grandpa Bao have three grown-up children, two sons and one daughter. They were both open about the fact that they value sons more than daughters. When asked why sons are better, Grandma Bu explained, “Apart from passing on the family line, they also can support you in old age … They can give us money and help us when we are old.” Grandpa Bao added: “The machines that are used for farming cannot be operated by women. Daughters cannot farm the land.” However, the notion that sons are essential for the reasons just quoted was inconsistent with the everyday realities of this elderly couple, who could be labelled as “empty nesters.” In fact, both their migrating sons had made it known that they did not want to come back to the village. This, however, did not disappoint Grandma Bu, who said: “I like them to be outside and work … I want them to develop wherever it is good for them.”
Grandma Liu's family provided another example where patrilocality was partly challenged as a result of rural–urban migration. As mentioned above, Grandma Liu had in the past formed a household with her second son, Lao Er. However, since Lao Er's house was blown over in a storm, she had lived temporarily in her third son's (Lao San) old house. In addition to his old house, Lao San had built a new house with money from his wife's family, who were migrant entrepreneurs. This meant that Lao San's wife had a great deal of decision-making power in terms of how the house was used and by whom, and Grandma Liu was not allowed to stay there. Thus, even though the house was located next to Lao San's birthplace, meaning that it was situated in a patrilocal way, the house was used purely as a holiday home. The fact that the new house was not made available to Grandma Liu indicates that the son–parent contract was indeed relatively weak. Moreover, Lao San's wife had also migrated to Shanghai – not to take up wage labour but to be a housewife and look after the children, which suggests that housewifization is not something taking place only in the countryside. Hence, Lao San was the only breadwinner in his household, which can be viewed as a neolocal family.
Concluding Discussion: Son Preference Reconfigured?
Based on ethnographic data from four villages in rural Anhui, this study has demonstrated how practices known for upholding son preference are shifting as a result of out-migration. It has paid special attention to the practices accorded to the division of labour in agricultural work and patrilocality. Rather than establishing frequencies and generalizable patterns, the article has aimed at contributing to a more nuanced understanding of how son preference is being reconfigured in the villages under study.
The study finds that the intergenerational contract is changing as a result of out-migration in mainly three aspects.Footnote 47 First, the direction of the flow of resources between generations is changing, and the older generation is assuming roles that are more supportive of their children. Second, the content of exchange is shifting, where the older generation offers more unpaid work, especially with regards to childcare and family farming, while the adult generation offers more material and monetary support. Third, the contracting parties are shifting as daughters sometimes also form close relationships and exchange resources with their parents. These changes suggest that beliefs about the functions and benefits of sons do not always translate to actual practices. The most notable discrepancy between the perceived need and actual function of sons can be found in family farming. Another discrepancy is the perceived need of sons for old-age support and the actual support from migrating sons, who sometimes do not send any remittances.
However, although the intergenerational contract can be said to have become weaker in some respects, the findings of this article suggest that such a weakening can take place without substantially challenging the “unsubstitutability” of sons. In fact, the results from this study point at transformations and continuities in the political economy of rural China, suggesting that son preference is becoming reconfigured in new and intricate ways.
The study finds that the extent to which adult sons cannot be substituted differs according to household type. As indicated in Figure 3, the study identifies seven ideal types of migrant households (marked with a white background), and shows that the impact of out-migration on household formation is more complex than the picture of “left-behind” women and “left-behind” older person households that is often discussed in the literature.Footnote 48 Figure 3 attempts to visualize the relative importance of adult sons and daughters to their parents with regards to how much financial support is transferred and labour provided. It gives four main scenarios, where “son unsubstitutable” is the least likely and “daughter unsubstitutable” the most likely scenario to lead to value change regarding son preference.
Figure 3: Changing Intergenerational Dependencies in Migrant Households (HH)*, Rural China
The article concludes that even though men are becoming increasingly substitutable in family farming, the relative importance of sons persists to a large extent for mainly four reasons. First, in the six migrant households marked with a dotted circle in Figure 3, patrilocality remains the main principle for organizing social and economic life. This makes sons largely unsubstitutable, even when they are not physically present, unless the household transforms into an uxorilocal older person household (marked with a solid circle in Figure 3) or a neolocal household. Second, although adult daughters in their place of origin partly take up functions to support their natal families, they compensate for their migrating brothers (and sisters-in-law) but do not substitute for them. Third, with the exception of the female migrant household, origin families tend to be dependent on migrating sons/husbands for financial support because of the low profitability in family farming. Relatedly, women seem to be missing out in the process of agriculture becoming increasingly mechanized and commercialized. Finally, there appears to be a depreciation of women's farm work, fuelled by low profitability in family farming and the fact that it often takes place within the frames of the patrilocal family. This means that even though adult daughters sometimes do farm work for their natal family, their work is not valued in such a way that it challenges the unsubstitutability of sons. Moreover, women's paid work is not only secondary to men's work but is also sometimes considered redundant, as evident from the signs of housewifization identified in this study. This means that practices known for upholding son preference only rarely shift in ways which make daughters unsubstitutable.
Given that the enhancement of women's economic status is negatively correlated with son preference, the out-migration of women seems at first to hold the most promising prospect for undermining son preference.Footnote 49 However, there are four factors to take into account when making such a claim. First, female migration can be an important strategy to strengthen the patrilineal family. Therefore, it is vital to make a difference between the migration of unmarried and married women. Women who migrate in youth migrant households tend to build strong relations with their natal families and continue to support their aging parents even after marriage.Footnote 50 However, the migration of unmarried women can also serve the purpose of strengthening the patrilineal family.Footnote 51 Therefore, it is not evident that youth migrant households contribute to making sons more substitutable, unless the household transforms into an uxorilocal or neolocal household. Second, it is essential to differentiate between older person households which are patrilocal and those which are uxorilocal in nature, where the latter implies a stronger daughter–parent contract, as illustrated in Figure 3. Third, the extent to which migration is a means for women to advance their status largely depends on how well they fare in terms of income and status as migrants. Lastly, it is important to note that while migration may benefit adult women, it may pose limitations for young girls, especially in older peasant households where girls do more household chores.
The final conclusion of this study is that although male out-migration undermines the argument that sons are needed to secure male manual labour in family farming, recent developments in the political economy of the villages observed in this study do not seem to challenge the vital role of male labour as a rural livelihood strategy. Even though female out-migration is common, it only rarely challenges the unsubstitutability of sons, and there seems to be a tendency for female out-migration to be subsumed into strategies which serve to strengthen the patrilocal family, at least from the point of view of those who stay in their place of origin. It is, of course, possible that son preference becomes attenuated through the value change brought about by ideational remittances by migrants who have acquired more gender-equal ideals and values, and through the Care for Girls Campaign.Footnote 52 However, the sustainability of any such value change seems questionable if the practices that underpin son preference do not undergo major shifts.
