Rural-urban migration in Vietnam and China: gendered householding, production of space and the state

The transition from state socialism to market socialism in Vietnam and China has been characterized by unprecedented rural-urban migration. We argue that this migration is integral rather than incidental to the gendered reproduction of state and society. A review of the emerging literature on trans-local householding explores the process whereby the reflexive engagement of the state and the household remakes rural-urban differentiation in ways that are deeply gendered and classed. As such, state regulation and control of migrants are part of a process of reconfiguring state-society relations in which the production of space and the symbolic valuation of ruralness and urbanness have become a central trope.


Introduction
In the latter half of the last century, both Vietnam and China underwent two major social transformations that have fundamentally reconfigured the relationship between rural households and the state (Kerkvliet and Selden 1998;Perry and Selden 2010;Whyte 2010). The first was the socialist revolution followed by successive policies of land reform and collectivization between 1945 and the 1970s. The state initially redistributed agricultural land to peasant households and later reorganized land and labor for large-scale production through agricultural collectives. A socialist welfare system was established, which went hand in hand with state interventions in family life. Equity was promoted in both productive and reproductive spheres through policies targeting gender discrimination, gerontocratic control and elite privilege (Phạm 1998;Hershatter 2007). The second transformation, which started in the late 1970s in China and the late 1980s in Vietnam, saw a redeployment of land to peasant households together with reduced state control of labor, pricing and accumulation, creating a market economy managed and directed by the state, namely market socialism. In both countries, the state has since exercised less intrusion into the daily life of ordinary people and their mobility, while reinstating the household as the basic unit of welfare provision. While the household economy existed only at the margin of the previous economy, the household now regained its primary role in production and reproduction.
Liberalization and market integration have brought about enormous economic and social changes in both countries. Gender differentiation has sharpened as welfare provisions designed to support equity have been dismantled and responsibility for social provisioning has been reallocated to households. There has been a dramatic decline in absolute poverty while the high growth rate has been sustained over a long period, despite fluctuations and environmental problems. Wealth accumulation and urbanization have taken place at a faster rate than ever before in history; in China, the number of millionaires has reached hundreds of thousands and China's mega urban centers are emerging as global cities (Hsing 2010;Jacka and Sargeson 2011). The second transformation has been characterized by much greater population mobility, internally and internationally. In the last three decades, hundreds of millions of Chinese and millions of Vietnamese have been migrating from rural areas to urban centers as hundreds of thousands others have gone overseas (World Bank 2008a) 1 as contract and illegal workers (Chu 2010;Dang et al. 2010;Nyíri 2010). A large number of others have emigrated to study and for leisure (Dang et al. 2010;Nyíri 2010;Fong 2011), becoming part of global circuits of labor, ideas and lifestyle.
Our focus is on rural-urban migration, which we argue is integral rather than incidental to market socialism, and its role in the ongoing reconfiguration of state-citizen relationships in Vietnam and China. Until recently, studies of rural-urban migration have focused mostly on rural-urban inequalities, gendered experiences of migrants in the city, and the life of 'left-behind' family members. These phenomena were often treated in isolation from each other rather than as components of a long-term and dynamic process of social change in which the nature of rural households is redefined. A false dichotomy between migrants and the left-behind obscures the processes in which households are iteratively organizing and deploying labor for multiple purposes in response to shifting systemic conditions while recreating gender and generational relations.
In contrast, an emerging literature on Vietnam and China examines the ways in which households straddle rural and urban areas, with generations of men and women simultaneously undertaking care and livelihood activities from different locations. We review this literature in order to link rural-urban migration with changing state-citizen relations under market socialism. We argue that rural households' multi-locational functioning interacts with state visions of development, control mechanisms and discursive categories in ways that marginalize migrants but simultaneously render migration increasingly imperative for market socialism and for household reproduction. Our analysis contributes to the existing literature on the ways in which migration is integral to the restructuring of state-society relations. Whilst one strand of this work focuses on international migration (Ong 1999;Silvey 2004;Porio 2007;Chan and Tran 2011;Pearson 2012), we develop insights about rural-urban migration that complement those for other places, such as Latin America and Africa (Hojman 1989;Lawson 1998;Francis 2002). Vietnam and China represent particular cases as transitional economies in which the state, despite liberal-inspired reforms, has been seeking to shape economic development, rural-urban migration and household relations to a much greater degree than other political regimes.
In the next section, we introduce the theoretical lens for the review which combines two concepts, namely Douglass's (2006) concept of householding and Lefebvre's (1991) theory of the production of space. We then proceed to analyze the changing forms of rural-urban inequalities rooted in the rural-urban relationship of the previous eras. While the forms have altered, the relationship remains unequal, with significant bearings on the social status of rural migrants. We argue that it is from rural migrants' reflexive engagements with this institutional context that translocal householding emerges, and we will examine how this produces a life 'in between' that is deeply gendered. Finally, we explore how this life 'at the margin' is producing ambivalent spaces for householding and for negotiating citizenship across different locations. This differentiated process invests rural(ness) and urban(ness) with meanings about gender and class. 2 Householding and the production of space Douglass (2006; develops the concept 'global householding' to analyze the social reproductive dynamics that are emerging in Pacific Asia in response to rapid changes in the world order. It refers to the 'ways in which the processes of forming and maintaining household through time' (Douglass 2012, 4) evolve in the face of mobility, population ageing and neo-liberalism. According to Douglass, 'all the key dimensions of the life of the household, including marriage/partnership, bearing, raising and educating children, managing daily life, earning income and caring for elders and non-working members ' (2012, 4) have to be understood as projects sustained across multiple locations and through varied social formations. He argues that neo-liberal policy regimes promote ideological functions of households that sustain gender, class and racial differences for the sake of capitalism. Meanwhile, '"transitional economies" such as Vietnam and China [ … ] maintain the authority of socialism but are dismantling public support for household reproduction while dispossessing rural and lower income households of land and other assets' (2012,10). His conceptualization thus also enables analytical attention to the householding processes that are occurring through rural-urban migration.
In post-reform Vietnam and China, the rural household has been reproducing its labor and gendered social relations across and between places and spaces (Oakes and Schein 2005;Jacka 2012;Nguyen forthcoming). Unlike the 'global household' that Douglass analyses, their spatial practices are heavily circumscribed by the sovereign state and firmly embedded within specific historical and cultural contexts. Lefebvre's notion of 'production of space ' (1991) is therefore apt to capture this interaction. According to Lefebvre, space is produced through a dialectical interplay between 'the perceived, the conceived, and the directly experienced ' (1991, 246). These constitute spatial practices by human beings, lived spaces (spaces of representation), and the ways in which they are conceived by state actors and the spatial imaginary of the time (the representation of space). Space is 'politically instrumental in that it facilitates the control of society, while at the same time being a means of production by virtue of the way it is developed' (349). It is central to the 'reproduction of the social relations of production', including the hierarchical ordering of locations and class structures (349). The urban-rural trope therefore must be analyzed as a complex social construction that is fundamental to capitalism's relations of domination.
In Vietnam and China, the shifting social construction of spatial categories has been intrinsic to the changeover from state socialism to market socialism. This process, we show, is political in the ways in which it reconfigures gender and class relations via means of reproducing 'the rural' and 'the urban' for market socialism. Combining Lefebvre's conception of space with Douglass's notion of householding enables us to analyze the reflexive processes in which rural-urban migrants' strategies for sustaining the household interplay with state governance of market socialism. In particular, it allows us to explore how they are inscribed in the production of space and the ongoing restructuring of rural and urban identities.

Changing social relations of production
Although state socialism was aimed at creating a classless society, we argue that it in fact elaborated pre-existing differences between urban and rural areas by privileging the party membership, its leadership and urban workers over the rural peasantry. Rather than creating entirely new class relations, market socialism in turn remade the social relations of state socialism in ways that reify rural-urban differentiation. This section elucidates this argument by tracing the changing social relations of production in China and Vietnam over time.
The size of the economy aside, there are a number of differences between China's and Vietnam's experiences with state socialism. First of all, collectivization was more farreaching in China in terms of scale, duration and the actual outcomes for productivity and rural social organization. Despite the Great Leap Forward policy with which collective production was brought to the extremes, Chinese cooperatives, which were larger and more industrially oriented, provided the conditions for rural entrepreneurship to evolve (Kerkvliet and Selden 1998;Ye et al. 2010). The cooperatives in northern Vietnam were, in contrast, more fragmented and had to be maintained alongside fighting the American war in a divided country. The household in Vietnam had somewhat greater autonomy in agricultural production than its counterpart in China, where household labor was managed and controlled to a greater degree. 3 Secondly, state interference in social reproduction was stronger in China than in Vietnam, where the Vietnamese state, especially at the local level, was less able and less willing to do so Kerkvliet and Selden 1998;Kerkvliet and Marr 2004;Koh 2004;Kerkvliet 2005). State socialism sought to reshape social reproduction through propaganda and incentives, social surveillance, social criticism, re-education, violence and imprisonment, but the will and reach of the Chinese state were far greater than those of the Vietnamese state. Key elements of social reproduction such as marriage and child bearing, child rearing and socialization, and peasant/worker entitlements were more strictly regulated in China than in Vietnam. An example is the draconian application of the one-child policy in China, compared with the two-child policy in Vietnam that was pursued mainly through discriminatory entitlements and social pressure.
These differences influenced the ways in which regional development took place following the reform, which began a decade earlier in China than in Vietnam, where actual restructuring only started in the late 1980s. The rural entrepreneurship emerging from the former Chinese cooperatives made it possible for the government to implement a policy to urbanize rural areas, the Township and Village Enterprises program or TVEs (O'Connor 1998;Perry and Selden 2010), together with developing core urban metropolitan areas and opening up its coast to global capital. This dual-track urbanization, alongside policies to control migration, was aimed at limiting urban growth in existing urban areas by simultaneously attracting the rural populations to smaller towns and diverting migrants to the periphery of major cities where they could be accommodated by local farmers (McGee 2009;Hsing 2010).
In contrast, Vietnam's urban development in the early reform period was less planned by the state and was instead more driven by the 'popular sector' comprising individuals and households taking advantage of the ambiguity in urban property rights and support by local governments (Leaf 1999;Leaf 2002;Koh 2004;Koh 2006). Whereas in China, rural industrialization took off, especially in and around coastal cities such as Guangdong, Beijing, Jiangsu, Shanghai, Zhejiang and Fujian, rural industries in Vietnam remained fragmented (O'Connor 1998), with industrialization and urbanization mainly concentrated in the Hanoi and Ho Chi Minh city regions. In Vietnam, there was a marked increase of the urban informal sector, including street vending and small-scale household businesses, whereas in China, rural townships, especially along the east coast, were able to promote greater-scale industrialization with rapid population and housing growth (O'Connor 1998;Leaf 1999;Leaf 2002;McGee 2009). Despite these initial divergences, both states have since the 1990s promoted urbanization and global integration, encouraging foreign investment and adopting a modernizing vision of development (Zhang 2001;McGee 2009;Hsing 2010;Harms 2012). The urban populations in both countries have increased significantly in the last three decades and continue to rise (Table 1), albeit at a faster pace in China.
Even with the rural industrialization program that draws more than 80 million rural laborers, China has been witnessing the rural-urban movement of between 100 and 200 million people at any time, making up 80 percent of the construction and 50 percent of the service workforce (Nielsen and Smyth 2008;van der Ploeg and Jingzhong 2010;Whyte 2010). On a smaller scale but accelerating, internal migrants in Vietnam numbered more than 6 million as of 2009, excluding the potentially huge number of unregistered migrants that were not counted by the census (United Nations 2010; GSO 2011). The great population mobility is foregrounded by an enduring rural-urban disparity. In both countries, urban income remains double that of rural areas (Đặng 2008;Sicular, Ximing et al. 2010) and urban populations have visibly better access to infrastructure, social services and social protection (Nielsen and Smyth 2008;Whyte 2010;Le et al. 2011). According to a conventional narrative, social inequality was deeply entrenched in prerevolutionary Vietnam and China; then it was lessened during state socialism, only to reemerge under the market economy. While socialist policies were aimed at erasing the difference between social groups and regions, the rural-urban division was in fact a product of the socialist modernizing project (Fforde and De Vylder 1996;Fesselmeyer and Le 2010;Whyte 2010). This 'hierarchical ordering' of geographical locations was indeed central to the 'reproduction of the social relations of production' (Lefebvre 1991) under state socialism and remains essential to the functioning of market socialism. Urban and industrial development was (and still is) prioritized over agriculture and the rural sector (Đặng 2008;Davis and Feng 2009;United Nations 2010;Whyte 2010) in both countries. The state used agricultural production to finance urban industries through price curbs on agricultural products and inflated prices of industrial commodities, including inputs for agriculture. While a state system of social services was set up in the countryside to provide education and healthcare, the rural populations never had access to the social securities that urban state workers and cadres were guaranteed, including provision of food, jobs, education and welfare (Kuruvilla et al. 2011). Finally, the household registration system, hukou in Chinese and hộ khẩu in Vietnamese, which was started in China in the early 1950s and later imported to Vietnam, practically prevented peasants from migrating elsewhere to seek alternative employment. The system ties rural and urban citizens to their household registration in a particular location, which is connected to land entitlements and access to social services (see Hardy 2001;Wang 2005 andLe et al. 2011 for background information on the household registration systems in China and Vietnam).
Summing up, the rural-urban disparity had existed before the revolution in both countries and to eradicate it was one of the goals of socialism. However, policies in practice made second-class citizens of the rural population, despite the socialist state's emphasis on the peasants' major role in national construction in both countries and, for Vietnam, in national defense. This has implications for understanding the class structure in both countries today. Rather than constituting a legacy of the previous 'social relations of production' (Lefebvre 1991), the post-reform rural-urban disparity has evolved in complex ways out of institutional processes in which old and new governing techniques interact with the market.

The state and the regulation of rural migrants
In both countries, the household registration (hukou/hộ khẩu) is no longer an instrument for the state to prevent mobility and allocate socialist provision. As an institutional boundary and commodity, it now facilitates a dualistic economy in which rural migrants are excluded from urban citizenship at different levels. While allowing for easier acquisition than before, the current household registration acts as a mechanism to control resources, limit social rights and expectation, and reinforce class and subject formation (Wang 2005;Lê and Khuất 2008;Luong 2009;Nyíri 2010;Zhao and Howden-Chapman 2010;Le et al. 2011). It has become an expensive commodity that can be bought, either directly from the state in highly sought-after urban centers of China (Wang 2005), or through bribes and other means (Hardy 2001;Hardy 2003). While rural people are no longer prevented from migrating, the opportunity to legally settle in the city is selectively narrow and accessible only to the most competitive migrants who are better educated, better connected or better off economically. As well, China has started to experiment with harmonizing rural and urban hukou in a small number of wealthy cities such as Chengdu and Beijing as a basis for reforming the local social security system. This potentially generates new spatial politics between insiders and outsiders of richer regions that exclude both rural and urban migrants from poorer areas (Shi 2012).
If the household registration previously kept rural people immobile in the service of socialist construction, it now makes them mobile for the sake of urbanization and capitalist industrialization. 4 Their 'floating' is perpetuated because they cannot easily settle in the city legally, while remaining in rural areas has become a template for failure, especially for young people (Ngai and Lu Huilin 2010;Fang 2011). Their exclusion from urban social citizenship, including access to jobs and urban social services (Nielsen and Smyth 2008;Davis and Feng 2009;Le et al. 2011) practically separates the labor of migrant workers from its social reproduction, which must take place in their home place. This creates a low-cost and flexible labor force for urban services and industries. With the household registration, rural laborers have thus been made instrumental for the state's successive projects of socialist construction and marketization through their status as second-class citizens in both periods.
Three decades after the reform, poverty remains predominant in certain rural areas, especially in western and central China (Davis and Feng 2009;Whyte 2010) and the central coast and mountainous regions of Vietnam (Taylor 2004;World Bank 2006;Badiani et al. 2012). Nevertheless, the rural-urban division has been reconfigured in both countries; rural poverty has declined whereas urban poverty has become more entrenched (Badiani et al. 2012;Cho 2013). 5 In both countries, urban workers have borne the brunt of lay-offs and urban residents are increasingly represented amongst the urban poor while 'industrial jobs, far from being a secure way out of poverty, remain unstable and underpaid' (Porta and Beresford 2012, 78). Some rural areas have been able to capitalize on their geo-political advantage for development, especially those along the east coast of China and the Southeast of Vietnam. As well, not all urban citizens are economically better off than rural citizens or migrants, a significant proportion of whom have become upwardly mobile. In China, 11 percent of the wealthiest income decile live in rural areas (Davis and Feng 2009, 10). In Vietnam, the latest census indicates that about two thirds of registered rural-urban migrants have 'high living standards', greater than the proportion of urban residents with the same living standards (GSO 2011). The rural-urban distinction has also been blurred by the incorporation of rural villages into urban centers alongside the citification of rural villages (Guldin 2001), and the ways in which diverse groups of migrants have become part of the urban society, with or without hukou/hộ khẩu (Chen et al. 2001;Zhang 2001;Koh 2006;Anh et al. 2012;Zavoretti 2012).
There has also been a shift in state policies in recent years towards increased investment and social transfers to rural areas, such as grain subsidy in China or old-age allowance in both countries (World Bank 2008b; Jacka and Sargeson 2011), coupled with the removal of agricultural tax and levies. China, followed by Vietnam, has instituted a program for Building the New Countryside, with plans for land consolidation, technological investment, and improvement of rural infrastructure and services, partly in response to rural discontent from land dispossession (Hsing 2010;Gillespie 2013). 6 The actual outcomes of these programs remain unclear (Nyíri 2010;Whyte 2010), while rural people continue to be on the move and their movements remain controlled by state instruments 7 such as the household registration. 5 A large part of the urban poor are migrants (c.f. Liu and Wu 2006). 6 About 50-60 million peasants have been dispossessed of their land for industrialization and urbanization (Hsing 2010). In Vietnam, hundreds of thousands farmers have also lost their agricultural livelihoods on the same account. 7 According to Goldstein (2006), the average single migrant needs four types of identification to work legally in Beijing: a national identity card, a temporary residence permit (issued annually at a cost), a work permit (also with yearly cost) and a health card issued by an appointed hospital. In Vietnam, the requirements are limited to the temporary residence card issued by the local police.
Not only is the hukou/hộ khẩu an administrative device, it is also a powerful metaphor for the discursive division between rural and urban people, a symbolic barrier that migrants have to cross to become desirable citizens. The second-class citizenship of the rural populations is indeed a product of the discursive production of categories and knowledge actively promoted by the state (Zhang 2001;Yan 2003;Harms 2011;. During state socialism, the peasants were constructed both as backward and ignorant and as powerful and revolutionary, both as examples of the hard-working socialist laborers and as in need of reform (Rato 2004;Jacka 2006). In Vietnam, the mobilization of the rural populations for the American war also relied on the construction of the heroic peasant. This construction has now has given way to an image of the uncivilized peasant whose presence pollutes the urban space (Nguyen 2012;Turner and Schoenberger 2012). In China, peasants are considered 'lacking in suzhi' while in Vietnam they are seen as 'having a low dân trí'. These keywords refer to the general quality of the population, encompassing professional skills, educational level, knowledge of law and appropriate cultural and social practices.
In China there is extensive literature on the suzhi discourse, which reveals the emerging ways in which people construct their class subjectivities in line with the state's governing goals (Kipnis 2007;Yan 2008;Jacka 2009;Sun 2009a;2009b). According to Andrew Kipnis (2007, 393): Since the early 1980s, government workers and analysts have increasingly used notions of suzhi to argue for all manner of policy. Any sort of 'development' project can be described in the language of raising the population's quality. Chinese educational reformers, intrigued by English-language educational theories that go under the name of 'competence education', translated this term as 'suzhi education', and managed to get suzhi education inscribed as a guiding national policy for the twenty-first century (Kipnis 2006). Human resource managers in both the public and private sectors justify recruitment and salary decisions in terms of suzhi. Rural cadres justify their own leadership positions in terms of their suzhi being higher than that of the peasants around them, and, of course, urbanites discriminate against rural migrants for their lack of suzhi.
In Vietnam, the term 'dân trí' or 'trình độ dân trí' has not reached the level of usage as the term suzhi in China. Yet in both countries, state discourse and practices related to population quality are closely linked to a form of governmentality that emphasizes the responsibility of the individual for self-development and self-governance. Similar to China, Vietnamese 'state and non-state actors have over the past decade become increasingly interested in the projects of self-cultivation and value creation that resonate both with the needs and anxieties of the market place and with continuous socialist genealogies' (Schwenkel and Leshkowich 2012, 386). According to Zhang and Ong (2008, 11), post-reform governance in China is characterized by the coexistence of self-interest promotion and socialist control, of neo-liberal logics and national sovereignty, which they term 'socialism from afar', where 'micro-freedoms coexist with illimitable political power'. In Vietnam, a parallel form of state governance is demonstrated by Nguyen-Vo (2008) in her political economic analysis of commercial sex, and contributors to the special edition on neoliberalism by Schwenkel and Leshkowich (2012). These studies indicate that the state has increasingly recast its interventions in social relations and citizen's life through means of technocratic guidance and scientific expertise.
In this late-socialist governing orientation, suzhi/dân trí is part of a biopolitics that differentiates between good and bad subjects, depending on people's ability to meet the demands of the market and their loyalty to the state. Rural migrants are cast as undesirable citizens, lacking in quality and potentially rebellious, who must reform themselves under guidance of state or market agents while providing for their own needs. Migrants' low suzhi/dân trí is contrasted with the higher urban quality of civility, law-abidingness and sophisticated consumption, which urban residents are able to self-cultivate with private resources (Jacka 2009;Sun 2009a;2009b;Schwenkel and Leshkowich 2012). The assertion of this hierarchical ordering of personhood alongside the ordering of geographical locations is again not new. The peasants, deemed backward and poor, had been the target of civilizing (wenming/văn minh) missions by colonial governments and nationalist movements in the two countries before socialism (Gourou 1955;Han 2005). Underlined by 'a fundamental inequality between the civilizing center and the peripheral peoples on which it acts' (Friedman 2004, 714), the discourse of civility nowadays increasingly indicates that 'the privatization of space is explicitly linked to efforts to protect propertied elites from the perceived threat of an unruly, uncontrollable public' (Harms 2009, 193).
Despite its modifications, hukou/hộ khẩu thus remains an institutional instrument to enact the differential civil, political and social rights that are part of rural and urban citizenship (Shi 2012). According to Fong and Murphy (2006, 7), 'by focusing on personal transformations as key to citizenship rights, state leaders could blame their inability to guarantee social rights such as education and satisfactory living conditions on individual failings and dispositional inadequacies'. As an example, the urban Chinese educational system excludes migrant children from the better-quality public schools on account of hukou, high fees and discrimination. They then turn to migrant schools that are understaffed, lacking in facilities and providing inadequate schooling for higher education (Wang 2006), which reinforces the construction of their low suzhi. This valorization of citizenship is embraced by the migrants, who, however, are 'dismayed by how this valorization devalued their own citizenship status' and seek to 'capitali[ze] on the flexibility afforded by their marginality' (Fong and Murphy 2006, 6-8). Citizenship, broadly conceived as membership in a community, therefore, is also about the right to belong through recognition and responsibility, or cultural citizenship (Marshall 1964). This results not only from 'struggles between individuals and states, but also from individuals' engagement with local and global discourses and economic systems that are not necessarily controlled by the state' (Marshall 1964). However coercive, a state vision of development, or 'representation of space' in Lefebvre's terms (1991), is constantly renegotiated through the ways in which people appropriate spaces for their everyday purposes. The following section focuses on the rural migrants' negotiations with institutional and discursive structures through transnational householding, which creates social spaces that cut across different categories, imagined and real.

Translocal householding: life in between
One of the most important tropes for the migration literature in China is that of the dagongmeiyoung female and single migrant workers, who are part of the floating population (liudong renkou), a category that includes people living and working in the city without permanent registration. Among others (Lee 1998;Pun 2005), Fan (2002Fan ( , 2003Fan ( , 2004a discusses the emergence of a migrant factory labor regime characterized by precariousness, vulnerability and exploitation that the dagongmei are subjected to. The dagongmei's labor is necessary for urban centers, but their citizen status renders them outsiders to the urban society. Similar migrant labor regimes to those in mainland China are visible in analyses of Vietnamese factory workers (Angie 2004;Nghiem 2004;Nguyen-Vo 2006;Nghiem 2007;Bélanger and Pendakis 2009;Bélanger 2010), where a docile female labor force is constructed and deployed for capitalist production. The literature on factory daughters contributes powerful accounts of their negotiations with exploitative and alienating employment structures. However, this strand of literature pays insufficient attention to the linkage between their mobility and the translocal householding that involves migration and return, multiple trajectories of household members, and the ways in which the household's gendered reproduction spans regions and places.
Nearly as much time has passed since the reform as the time under state socialism, and rural-urban migration 8 is no longer external to either the countryside or the city in both countries. More than a generation of migrants have left the country and come back; quite a few have left for good. Studying return migration in rural Jiangxi of southern China, Murphy (2002) analyses the diverse migration trajectories of households and individuals that eventually lead to their differential standings in the village. Neither temporary nor contingent, Murphy (2002, 45) points out, these movements have been internalized in village life: [ … ] through migration and return, individuals and families use, reproduce, and reconstitute values and resources in their efforts to attain goals. This creates a continual feedback mechanism whereby migration and return become institutions internal to the villageinstitutions in which both migrants and non-migrants participate, and institutions that interact with the outcomes of other processes of change.
Most of the early migrants in China and Vietnam have now reached the age when they are likely to be grandparents. Many retired migrants are now caregivers for their grandchildren and the elderly in their household, as their children migrate (Judd 2010;Nguyen forthcoming). In the 1980s and 1990s, it was common for one spouse to migrate at a time while the other stayed home with the children. In China, young women often migrated for some years for work before returning to get married (Fan 2004b) whereas married women were usually pressured to return by domestic duties (Lou et al. 2004). In Vietnam, it was common for married women to migrate into urban small trade (Resurreccion and Ha 2007) on account of the greater availability of such work in urban Vietnam (McGee 2009) and the flexibility of the work that allows them to perform familial duties while migrating (Vu and Agergaard 2012a;2012b). Since the 2000s, more young married couples migrate together in both countries, either bringing their children along or leaving them in the country with the grandparents (Fan 2011;Locke 2012;Locke et al. 2012).
In her recent works, Fan (Fan 2008;Fan and Regulska 2008;Fan and Wang 2008;Fan 2009;Fan 2011) has moved away from viewing the hukou as the main reason that keeps rural migrants from settling in the city. In contrast to her earlier structurally driven accounts (2002, 2003, 2004a), Fan and her colleagues are now more emphatic about the agency of migrant households. 'Split-households' featuring migrants as single persons, couples or a couple with children, they argue, are an active strategy of maintaining a combination of rural and urban livelihoods in order to deal with the insecurity of urban life and work. In straddling the city and the country, these households are able to 'make the best of both worlds' (Fan 2008, 13), namely the income opportunities in the city and the extended support system and a sort of fallback position in the country. The household thus represents a source of security, and gendered translocal householding is a mechanism to strengthen this security.
This focus on the agency and strategies of rural household has also been central to recent works by other China scholars (Judd 2010;May 2010;van der Ploeg and Jingzhong 2010;Jacka 2012). Jacka argues for the utility of the householding concept by Douglass (2006) in examining the 'strategies and processes through which rural households create and reproduce themselves ' (2012, 2). In the northwestern rural community she studied, local people have diverse strategies for combining migrant work, agriculture and care that vary according the household's labor capacity and its life cycle (see also Chen and Korinek 2010). The fluidity and flexibility of householding are premised on the cooperation between household members, but internal conflicts are equally part of the process. The conjugal bond has become more fragile due to the mobile nature of householding, despite its increasing importance for the household economy (Judd 2010;Jacka 2012). Young people's migrant work brings them closer to the family because of greater uncertainty, as it establishes their independence, which is both liberating and frightening for them. Young female migrants returning home for marriage tend to experience a clash between the perceived personal freedom enabled by migration and patriarchal relations in village life (May 2010;Ge et al. 2011;Zhang 2013).
The mobility of young men and women shapes householding dynamics as much as it is shaped by them. May (2010) analyses the contradictory parental motives behind the mobility of sons and daughters in a rural northeast village of China. Whereas sons are encouraged to accumulate savings for an eventual return to marry a local woman who could later take care of them in old age, daughters are actively motivated to seek marital opportunities with urban men through investing their earnings in appearance and lifestyle. The parents aim to access the social provision that urban people enjoy through their daughters' potential marriage alliance, which in fact does not occur frequently, alongside the traditional care provided by the dutiful daughter-in-laws. These contradictory aspirations, she shows, are bound to lead to disappointments among the young men in finding the right wives and among the young women seeking more urbane and less patriarchal marriages. This indicates that emerging strategies for old-age care and provision against the shortage of social security are indeed grounded in age-old notions of filial piety that are highly gendered.
Translocal householding processes are also surfacing in recent migration research on Vietnam. The emphasis has similarly shifted to the ways in which the rural household reflexively embraces migration and return as part of diversified livelihoods during its life cycle (Resurreccion 2005;Zhang et al. 2006;Resurreccion and Hà 2007;Pham and Hill 2008;Truong 2009;Agergaard and Thao 2011;Hoang 2011a;2011b;Locke et al. 2012;Vu and Agergaard 2012a;2012b;Vu 2013;Nguyen forthcoming). These authors likewise demonstrate that the migration of individuals is embedded in household gender and intergenerational relations while their migratory experiences are intimately connected to the well-being of their family members. Care and parenting are shown to be essential to migration, which migrants often frame as parenting rather than as a departure from it . As argued for China (Murphy 2002(Murphy , 2004Zhang 2013), translocal householding in Vietnam is fraught with emotional dislocations and adjustment and belonging issues, as shown in studies of marital dissolution among migrants (Locke et al. 2013) and life-work negotiations of factory workers (Bélanger and Pendakis 2009).
While the China literature pays greater attention to young female migrants, Vietnam authors focus more on the migration of married women, the accompanying household rearrangements and their implications for gender identity. The initial uneasiness created by married women's departure has been now significantly reduced; their absence from home no longer requires justification (Resurreccion and Hà 2007;Truong 2009;Hoang and Yeoh 2011;Vu and Agergaard 2012a;2012b;Nguyen forthcoming). What emerge are the 'caring man' and the 'empowered woman' who move back and forth between city and village life, earning income and sustaining their family (Hoang and Yeoh 2011;Locke 2012;Locke et al. 2012;Vu and Agergaard 2012a;2012b). Yet gender identity remains constructed around women as the dutiful and caring wives/mothers, and around and men as 'pillars of the family'. What male and female migrants do may be challenging conventional familial arrangements, yet they still seek to perform and cast their practices in ideal traditional terms of the patriarchal family (Vu and Agergaard 2012a;2012b;Nguyen forthcoming). Similar dynamics are documented in the sexual life of couples separated by migration in which traditional norms of sexuality are invoked even when migrant men have transgressed them (Nguyen et al. 2011).
In contrast to the media panic about the 'left-behind' children and the elderly, abandoned and uncared for, care is central to the diverse arrangements of the translocal household. While the emotional dislocations and vulnerabilities of the children and the elderly, many of whom are caregivers themselves, have to be taken into account , Mummert 2010Hoang and Yeoh 2012), rural households in fact continually adjust migration to the care needs and caring capacity of their members (Fan and Wang 2008;Fan 2009;Nguyen forthcoming). The decision to migrate, to return or to remain in the countryside has as much to do with care as with household pursuits of livelihoods. Migration patterns, i.e. circular or permanent, over long or short distances, and the chosen type of migrant work, are closely related to local ideas regarding how care should be provided to dependents and household negotiations over the care needs of members. These needs shift over time together with changing notions of the appropriate caregivers and caring practices (Mummert 2010).
Finally, migration has facilitated changing aspirations in family life and work. Rural migrants nowadays increasingly include young people born after the reform, for whom labor migration has become an inevitable trajectory (Fang 2011;Ngai and Lu 2010). For many of these new-generation peasant-workers, return to the countryside is becoming difficult. Ngai and Lu (2010), for example, show that the attempts to go back to agricultural work by Chinese young men in Shenzhen and Dongguan fail without the support of their parents who expect them to move away from farm work. In Minh Nguyen's recent field research in rural North Vietnam, she also observed a strong aspiration among farmers for their children to be 'free from the paddy field', through urban work, education or marriage. As the population ages, young people's movement out of agriculture will affect rural social support, which is currently maintained primarily by middle-aged men and women, many of whom are former migrants (Judd 2010;Nguyen forthcoming). Their aspirations and strategies to have their children liberated from farm work and rural life inadvertently help drain care resources from the countryside, undermining intergenerational support in the future. Their children, meanwhile, continue to join the urban precariats. As the second generation of migrants, their struggles are, however, distinct from those of the previous generation, especially with an emerging class consciousness fueled by anger and resentment (Ngai and Lu 2010). These partly explain their enduring identification and connections with the home place, even when they do not plan to return (Myerson et al. 2010;Anh et al. 2012). As such, translocal householding will continue to be part of life for rural people, even though its dynamics will shift together with new patterns of population mobility. In the next section, we connect translocal householding with the production of space in the major urban centers of the two countries.
At the margin and on the edge: space, marginalization and power So far, we have argued that both because of and despite structural conditions, translocal householding produces a relatively autonomous space through the circulation of individuals and families between the city and the country. This space bridges the rural and the urban worlds, rendering the rural-urban distinction porous and fuzzy for people involved. In this section, we show that operating at the margin of urban society, physically and metaphorically, rural migrants create other spaces that transcend the rural-urban distinction, at times precisely through practices that reproduce them.
The floating population (liudong renkou) remains central to scholarship on internal migration in China (Lee 1998;Ma and Xiang 1998;Solinger 1999;Zhang 2001;Zheng 2003;Zheng 2004;Jacka 2006;Pun 2005;Zheng 2007;Yan 2008;Sun 2009a;2009b). Rather than being concentrated in factories, migrants increasingly work in the urban service sector. As domestic workers, bar hostesses or sex workers, they are subject to labor regimes that are similarly precarious and alienating while also being stigmatizing because of the association with servility and promiscuity (Zheng 2004;Zheng 2007;Yan 2008;Sun 2009a;2009b;Otis 2012). This trend has also been observed among Vietnamese migrant laborers, the lao động ngoại tỉnh (Henriot 2012;Nguyen 2012;Arnold 2013). Vietnamese labor migrants, however, remain more employed in 'popular sector' activities (nghề tự do) such as street vending, junk trading and household businesses that are less predominant in major Chinese cities (Mc Gee 2009), except for domestic service. In Vietnam, the female street vendor carries the same analytical weight for the rural-urban relationship as the migrants who are staffing industrial zones (Higgs 2003;Resurreccion and Hà 2007;Mitchell 2008;Mitchell 2009;Truong 2009;Turner and Schoenberger 2012;Arnold 2013). The greater presence of women migrants in Vietnam, who account for more than 50 percent (GSO 2011) of migrants in contrast to 30-40 percent in China (Jacka and Sargeson 2011), might also be correlated to the greater prevalence of popular-sector activities in Vietnam that are more accessible to women of all ages, especially married women.
Depictions of rural migrants tend to reinforce an image of abjectly poor and marginalized people who move around passively, doing whatever comes their way. In fact, most migrants come to the city with specific ideas about what they are doing and a network of support that facilitates their entry and operations (Ma and Xiang 1998;Zhang 1999;Resurreccion and Hà 2007;Agergaard and Thao 2011), while many have become economically better off than urban people of the laboring class (Zhang 2001;Nguyen 2013). Rural networks guide men and women to different kinds and destinations of migrant work, depending on local norms governing male and female behavior (Hoang 2011a;2011b). They often develop out of a particular connection to a certain trade or activity in the home place, such as junk trading in Nam Định of Vietnam (Mitchell 2008) and in Sichuan of China (Ma and Xiang 1998).
On account of these networks, migrants are commonly concentrated in certain areas of the inner city or in the urban edge of the metropolises, where about three quarters of the Chinese liudong renkou reside (Leaf 1999;Zhang 2001;Liu and Wu 2006;Hsing 2010;Hao et al. 2011;Harms 2011). They form enclaves of migrants from various native places practicing particular trades or working in the factories located in the region. Many commute daily to the inner city to work (Zhao and Chapman 2010). Migrants' life there is closely connected to that of local villagers, peasants-turned-landlords who enjoy considerable income but remain culturally inferior to urban people (Liu and Wu 2006;Siu 2007). To the urban middle class and urban governments, these 'villages in the city' (zhengchongcun) represent a mass of crime and poverty, which corresponds with the construction of peasants as backward and dangerous. In both countries, urban governments seek to cleanse these spaces of the perceived unruliness and disorder through eviction for construction and industrial development (Zhang 2001;Harms 2009;Hao et al. 2011;Harms 2011). To the migrants, these places are as much sources of livelihoods and sociality as spaces dominated by powerful people in their rural network and property-owning local residents (Zhang 2001;Siu 2007). The following accounts of the migrant communities in the urban edge of Beijing (Zhang 2001) and of Ho Chi Minh city (Harms 2011) provide further insights into these dynamics. Zhang (2001) documents the rise and fall of a migrant enclave in the southern suburb of Beijing, the Zhejiangcun or Zhejiang village, a vibrant community of migrants from Zhejiang province who are renowned for business savviness and their traditional trade in garments. Informed by Lefebvre's theory of space (1991), Zhang pays special attention to the ways in which the migrants struggle to appropriate urban space to accumulate wealth and 'gain control over communal lives and economy ' (2001, 9). Spatial practices in Zhejiangcun, however, also reveal the exploitation of migrant tenants and waged workers by the housing bosses and business owners from the same migrant network, while indicating a high level of gender differentiation. Wealthy women are removed from production to be confined to the home whilst middle-and low-income women's primary role in garment production is devalued on account of being performed at home. Young female migrant workers are subject to their migrant bosses' control, often disguised in the 'part of the family' discourse. Although deeply embedded for years in patronage ties with the local government, Zhejiangcun was finally demolished to make way for urban development, to reemerge in other locations that are similarly marginal but suitable for their economic operations. Zhang's account reveals the ways in which migrants, through maximizing their networks' advantages and working with the state clientele politics, are able to create autonomous social and economic spaces against the state, albeit uncivic ones. It also suggests the degree to which the Chinese state is willing to impose its visions of modernity and engineer social change, which, however, are tempered by the migrants' spatial practices.
Erik Harms analyzes the spatial and social development of Hoc Mon district on the periphery of Ho Chi Minh city, focusing on the everyday practices of different groups of residents, including the locals, rural migrants and people moving out from the inner city. He demonstrates that these practices defy the rural-urban dichotomy in a social space that is neither rural nor urban, as the city 'cuts across the rural landscape and creates new forms of space that are both rural and urban at once ' (2011, 45). This is a space betwixt and between that appears both dangerous, associated with disorder and 'social evils', and a space of opportunities, where export factories and property-owning opportunities abound. Similarly to Zhang (2001), Harms stresses the production of meanings partaken by the state and the people themselves that maintain the rural-urban division in the urban edge, despite the blurring of physical boundaries. He argues that existence in such a between-category space is characterized by a combined sense of exclusion and power. This combination explains why people continue to reproduce the division through their performance and movements between the edge and the inner city, between the city and the country. Residents of the urban edge, he shows, strategically deploy rural and urban identities for social and economic purposes in different social settings, thus simultaneously contradicting and reproducing these ideal categories.
Similar dynamics of space and power are prevalent not only in the urban fringes (Leaf 2002;Siu 2007), but also among groups of migrants who mingle more with the urban populations in the inner cities, such as street vendors, sex workers and domestic workers.
Nguyen (2013)'s analysis of a migrant community specialized in junk trading in Hanoi suggests that they have capitalized on their traditional niche in the junk trade to develop a profitable recycling economy employing tens of thousands of people. Their spatial practices are similar to those described by Zhang (2001) and Harms (2011), except that they continually move around the city, making use of urban spaces that are temporarily available. These practices afford them flexibility while making them ever more transient and liminal, consolidating the urban perception of rural migrants as unruly and dangerous.
For these migrants, the performance of a rural identity is essential to their daily transactions with urban people. Like the female street vendors from the edge of Saigon who don a rural attire in order to market their goods as rurally authentic in the inner city (Harms 2011), junk traders in Hanoi highlight their inferior status to appeal to the compassion of urban people (Nguyen 2013). Performative strategies are also common among personal service workers such as bar hostesses or sex workers. As much as their 'female body is being packaged in relation to the growing consumer market' (Chen et al. 2001, 25), they also tailor a variety of personas to appeal to different customers to maximize their earnings (Zheng 2004;Zheng 2007;Kay Hoang 2010). Such performances are often accompanied by their consumption practices modeled on dominant cultural symbols (Zheng 2007;Otis 2012). Indeed, urban-oriented consumption is a common strategy among female migrants in China, who see it as a way to improve their suzhi and reduce their marginality (Sun 2009a;2009b;Yan 2008). As Otis (2012, 150) writes: Through their commodity-enabled, self-imposed self-alterations women workers adapt to the tastes, preferences, and expectations of urban consumers. They spontaneously conform to their symbolic labor, the presentation of an aesthetic tied to the body, to urban gender and class expectations.
The success of this strategy is doubtful, since it is precisely through consumption that the urban middle class establishes its social distinction (Zhang 2010;Nguyen-Marshall et al. 2012). Migrants' spatial practices therefore work like double-edged knives (Harms 2011), both challenging their peripheral position by 'creating a parody of their constructed image and, at the same time, reinforc[ing] and reproduc[ing] hegemonic asymmetrical power relations and their marginality' (Zheng 2004, 88).

Conclusions
Rural-urban migration in Vietnam and China has been greatly transformed alongside momentous spatial and social transformations since the two countries shifted to a market economy, albeit to differing degrees and with regional variations. These transformations have reconfigured the rural and urban division. While material boundaries may be receding, symbolic boundaries are expanding, as the state increasingly seeks to govern through categories that construct rural and urban populations as occupying varied levels of development. They form 'idealized conceptions of spatial order', which social actors simultaneously internalize and play off by 'practices, strategies and socially embedded negotiations' in order to 'make space for everyday life' (Harms 2011, 238). Through their spatial practices, rural migrants have created relatively autonomous spaces to sustain families, carry out their economic activities and foster sociality and belonging across places. Since the state in both Vietnam and China continues to keep a strong hold on its citizens as it advances capitalist development, these spaces remain subject to its interventions and control, at times highly intrusive and destructive.
Translocal householding has become central to rural-urban migration. The previously rural household now operates between the city and the country, as household members migrate and return over their life course, performing gendered familial duties from both locations. Migration and return thus must be seen in the context of the rural household reshuffling and strategizing in order to reproduce itself in response to social and systemic changes. Individuals may migrate out of desires for freedom and independence, but their mobility is necessarily part of gendered household strategies to accumulate wealth and secure social provision that are continually adjusted to household members' changing care needs and capacities. This has become more imperative on account of the emphasis on the household as the primary unit of reproduction and guarantor of individual well-being. Translocal householding, however, is malleable to ruptures, instability and internal contradictions, especially when migrant household members assume identities that are not necessarily favorable to the householding project, which remains embedded in patriarchal relations and ideas.
Operating at the edge of cities and on the margin of urban society, rural migrants have also created spaces that concurrently blur the rural-urban binary and discursively reconstruct it. Local residents and rural migrants on the urban fringe are viewed as an obstacle for the state's civilizing project. Yet they effectively defy official and popular categories imposed on them by creatively carving out their own social and economic space, notably through performative strategies and consumption. Although they recognize 'how seemingly pure categories are shifty symbols, with meanings that change according to contexts' (Harms 2011, 76), their spatial practices reproduce these dominant categories while creating new patterns of social and gender differentiation.
We conclude that bringing 'class back in' to analyses of agrarian change (Bernstein 2010) enables an examination of the ways in which rural and gender differentiation feed into class formation processes through rural-urban migration in post-reform China and Vietnam. The rural-urban distinction will become less significant, as rural labor not only straddles between the city and the countryside, but also between categories of work, joining the 'classes of labour' (Bernstein 2010). However, this process will not erode the marginality that rural migrants experience, because they are most likely to occupy the lower rungs of future 'classes of labour' and, as such, precariousness is likely to remain a defining feature of their existence.