The return intentions of China’s rural migrants: A study of Nanjing and Suzhou

ABSTRACT

A considerable proportion of China’s rural-to-urban migrants today are not able to formally settle in one location. Whether these migrants stay in the city or return to their rural homes has important ramifications for the country’s demographic and economic changes. Using data from a recent survey in Nanjing and Suzhou, this article explores the settlement and return intentions of rural migrants. It is found that although most rural migrants do not intend to settle permanently in the host city, a considerably large group of them are inclined to return to a local town or city in their home region rather than to their rural origin. Such prospective urban returnees share many similarities with members of the floating population but express distinct preferences and concerns. For both prospective rural and urban returnees, family obligations play a decisive role in the formation of the return decision. The findings suggest that return migration, on the one hand, is a trade-off between livelihood in the rural origin and the urban destination and, on the other hand, aims to maximize utilities based on one’s family background and resources, which may drive rural migrants to settle in urban areas close to their rural homes.

Introduction

In the process of rapid industrialization, migration features a one-way movement, with major streams of migrants leaving the home countryside for urban areas. Nevertheless, the dominant stream of outmigration is often accompanied by a counterflow (Lee, 1974). After living in the city for a certain period of time, a portion of rural migrants may choose to return for a variety of reasons, from life-changing chance opportunities to limited socioeconomic capabilities, and from lifestyle preferences to family obligations (Constant & Massey, 2002; Gmelch, 1980; Simmons & Cardona, 1972). Culture may also play a role; in China, migrants and sojourners, guided by the old belief of luo ye gui gen (i.e., fallen leaves return to their roots), are expected to return home eventually. Since the 1978 reforms when Chinese farmers started leaving home for urban jobs, they retained their rural homes and expected to return (Wang & Fan, 2006). Evidence suggests that a large proportion of early migrant workers have indeed returned home, especially those who were unable to sustain their urban employment due to old age and/or family obligations at home (Wang & Fan, 2006; Zhao, 2002).

More recently, the intention of return has gained popularity among younger migrants. The notion of “Escaping Bei[jing] Shang[hai] Guang[zhou]” aroused much empathy among rural and urban migrants in big cities due to the high living costs and restless lifestyle. With the slowdown of economic growth and the surplus of labor in industrial sectors, return migration is on the rise (Chan, 2010). A growing number of younger migrants have begun to consider returning home a desirable option.

Return migration in China has been extensively studied (Démurger & Xu, 2011; Wang & Fan, 2006; Zhao, 2002). Evidence suggests that discriminatory urban policies engendered by the hukou (residency status) system have discouraged rural migrants from settling, and family obligations have often motivated them to return (Wang & Fan, 2006; Zhao, 2002). Nevertheless, migrants who intend to return do not necessarily go back to their home village. Instead, many of them choose to settle in a local town or city in the home region and close to the rural home. On the one hand, settling in an urban area gives them access to better employment opportunities and living standards absent in the rural home. In the meantime, family obligations, such as taking care of elderly parents and/or children, are possible given the proximity of rural and urban homes. On the other hand, living costs in these urban areas are considerably lower than those in big cities. Such considerations drive migrants to return but not necessarily to their home village. However, research on China’s return migrant workers typically defines return as returning to the home village, and the selection bias in the conceptual framework and data collection only includes those who return to the home village and inevitably excludes those who return to an urban area in the home region that is considered home in a broader sense. Given the sheer size of China’s migrant population and the likelihood of returning to their home regions’ numerous destinations other than the home village, the migrants’ decisions about returning and where to settle could have tremendous implications for the country’s sociodemographic changes.

To bridge the gap, this article explores the disparities in the return propensity of rural migrants between returning to the home village (i.e., rural returnees) and returning to a local town or city in the home region (i.e., urban returnees). It contributes to the existing literature on return migration by including alternative return pathways followed by a large proportion of rural migrants. The research, which theoretically exhausts the options of return migration and empirically includes in the sample prospective return migrants of all kinds, can provide more thorough insights into the patterns of return migration and their impact on relevant rural and urban societies. The article is organized as follows: The second section reviews the literature with respect to return migration in China. The third section introduces the data and methodology. The fourth section presents the results of a statistical analysis to illustrate the driving factors of different types of return migration. The final section concludes.

Literature review

Return migration

In general, people are motivated to migrate when economic returns elsewhere are higher than those at home. Neoclassical economics consider migration a cost–benefit evaluation to maximize earnings (Sjaastad, 1962; Todaro, 1976). However, the new economics of labor migration views migration as a household strategy to overcome market failures at the origin (Stark, 1991). Based on an understanding of migration, return migration could be explained from several theoretical perspectives. The human capital approach considers that returnees are rejected by the destination due to their relatively poor human capital (Lindstrom & Massey, 1994). The social capital perspective highlights the role of social networks in the migration process. For instance, social capital accumulated in the place of destination increases the opportunity cost of returning, whereas social capital in the place of origin increases the benefits of returning (Constant & Massey, 2002). Nevertheless, the life cycle approach views return migration as a consequence of critical life stages, such as marriage, childbearing (Borjas & Bratsberg, 1994), and retirement (Rogers, 1990). Likewise, the family strategy perspective emphasizes the need to fulfill family obligations in the decision making of return migration, such as taking care of young children or elderly parents (Tiemoko, 2004).

Empirical studies have examined the determinants of return migration. In market economies, economic conditions in the host country, such as an economic recession and low employment rates, are a major cause for return migration (Massey, Goldring, & Durand, 1994). At the same time, improved economic conditions in the origin, such as rising employment opportunities and wage levels, also drive people to return (Dustmann, 2001; Lindstrom, 1996; Newbold, 1996). For instance, Lindstrom (1996) found that the increase of capital investment in Mexico shortened the trip spells of Mexican immigrants in the United States. Newbold (1996) indicated that the rise of employment and disposable income in the home country of immigrants to Canada increased the likelihood of return migration.

In addition to macrolevel circumstances, the human capital of migrants, such as language fluency, educational attainment, and migration experience, may contribute to the return decision (De Jong, 2000; Dustmann, 1999; Lindstrom & Massey, 1994). Lindstrom and Massey (1994) demonstrated that Mexican emigrants who lack human capital are more likely to return. In addition, migrants who have lower wages and lower social status in the host city are more likely to return (Constant & Massey, 2002; Gmelch, 1980). In addition, social networks (Gmelch, 1980; Niedomysl & Amcoff, 2011; Tiemoko, 2004) and family obligations in the place of origin and discrimination in the host society encourage return migration (Gmelch, 1980). For instance, Niedomysl and Amcoff (2011) observed that some migrants in Sweden return to live closer to relatives and friends. Tiemoko (2004) identified that African emigrants to Europe or North America are likely to return when influenced by relatives and friends in the home country. Sometimes, return migration could be a deliberately planned event rather than the outcome of a failed career (Newbold, 2001). For instance, based on data collected in Canada and Australia, Newbold and Bell (2001) found that return migration caused by failure is overstated; instead, a large share of return migration is preplanned, which is a part of the career development plans of individuals. Economic opportunity in the place of origin is a particularly important determinant for preplanned return immigration (King & Newbold, 2008).

Rural migrants and return migration in China

In China, rural migrants are in a disadvantaged position in host cities, given that the hukou system permits unequal access to urban resources and routine discrimination against rural migrants (Chan & Buckingham, 2008; Y. Liu, He, & Wu, 2008; Lu & Song, 2006; Nielsen & Smyth, 2011; Wang & Fan, 2012). Rural migrants usually take the labor-intensive and low-paying jobs that are despised by local residents (Lu & Song, 2006; Roberts, 1997; Wong, Li, & Song, 2007). In addition, rural migrants typically suffer from poor living conditions (Huang, 2003) and have little chance of becoming homeowners in the city (Huang & Clark, 2002; Wu, 2004). They also face various difficulties and higher risks given the lack of social welfare (Guan, 2008; Seeborg, Jin, & Zhu, 2000; Xu, Guan, & Yao, 2011) and have a different lifestyle and concerns compared to local urban residents (Ye & Wu, 2014).

Nevertheless, the situation has been gradually changing. Following the central government’s “new-style urbanization” campaign in the early 2010s, some local governments have carried out reform measures to further relax the hukou system. In 2016, the “residence card” policy was introduced in some cities to allow eligible rural migrants to enjoy basic urban public services such as public education and medical care. In addition, rural migrants’ educational attainment and professional skills have improved considerably in the recent decade, leading to higher human capital to enable a long-term urban livelihood (National Bureau of Statistics, 2016). On the other hand, economic conditions and living standards in rural areas have also improved substantially over the years (Long, Zou, Pykett, & Li, 2011), driving many rural migrants to return. These changing circumstances may alter rural migrants’ opinions about where to settle and hence their attitudes toward return migration.

Studies on return migration in China have focused on those who have returned to home villages. Such return migration closely correlates with certain socioeconomic attributes of rural migrants. For instance, Zhao (2002) found that return migrants tend to be older, married, better educated, and solo migrants with the spouse staying in the home village. Wang and Fan (2006) observed that migrants’ institutional and social inferiority in the host society and family needs in the rural home drive migrants to return. In general, unfavorable conditions in the host society including a lack of job opportunities and family separation as well as improved employment opportunities and an intention of reuniting with the family in the rural home encourage return migration (Démurger & Xu, 2011; Murphy, 1999; Wang & Fan, 2006; Zhao, 2002). Return migrants’ knowledge and experience gained in the city benefit their employment or entrepreneurial activities in the home countryside, contributing to the revitalization of rural economies in less developed regions (Démurger & Xu, 2011; Ma, 2001; Murphy, 1999).

Rural land is another influential factor for return migration in China. Hare (1999) found that a land endowment motivates return migration in Henan. Wang and Fan (2006) indicated that the availability of land resources is a determinant for return migration in Sichuan and Anhui. In addition, government policies contribute to return migration. For instance, Solinger (1999) suggested that rural migrants maintain circular migration to counteract institutional constraints. Zhao (2002) stated that many city governments’ “anti-migrant” policies increase the rate of return migration somewhat. Murphy (1999) emphasized the effect of rural policies on attracting migrants back to their hometown.

An intrinsic problem with the analysis of return migration is the selection bias associated with the sample of return migrants. Such a sample, usually collected in the place of origin (e.g., one village or a collection of sites), includes only those who have migrated and returned and tells little about the possible structural differences between those migrants who have returned and those who have not. There are two means for solving this problem. First, a longitudinal research design could allow for a sufficient time span to examine the life trajectories of migrants and identify the factors that drive return migration. This research design requires prolonged data collection of an intrinsically mobile group, and it often covers a relatively small sample. These drawbacks and difficulties lead to the second research design, which examines prospective returnees and stayers rather than actual returnees. In other words, it aims to identify the intentions of migrants to stay or return, though such intentions might not be realized eventually. For instance, based on a survey in Fujian Province, Zhu (2007) found that the settlement intention of migrants is negatively associated with unstable employment, low earnings, low levels of social insurance, and expected risks in cities. Individual-level attributes, such as gender, age, marital status, education attainment, income, occupational type, and place of origin, and city-level attributes, such as administrative status and the population size of the destination city, influence the settlement intention of migrants in the cities. Yue, Li, Feldman, and Du (2010) highlighted the cohort disparity between the young generation of migrants and their older counterparts. The former group is concerned with their social capital and the socioeconomic conditions in the place of origin, whereas the latter group is driven by age, occupational skills, and family obligations. In addition, the affective emotion of migrants and access to housing also play an important role in their stay/leave intentions (Du & Li, 2012; Z. Liu, Wang, & Chen, 2017).

In addition, existing research tends to focus on the return to the countryside home. Other types of return such as returning to a local town or city in the home region—a main settlement option for many rural migrants—have been neglected. In addition, the changing socioeconomic circumstances have rendered the decision of return migration and settlement intention increasingly complex and deserving of thorough investigation.

Destinations of return migration

Institutional and social constraints in the city and family obligations at home often drive rural migrants to return (Wang & Fan, 2006; Zhao, 2002). However, due to a lack of adequate farmland and the negligible income from agriculture, returning to the rural home to engage in agricultural activities generally cannot provide a standard of living that is satisfactory (Hare, 1999). As a result, rural migrants would rather keep circulating between urban and rural areas and have agricultural income only as supplementary to the bulk of income earned in the city (Hare, 1999). The younger generation of rural migrants, who have more schooling but little farming experience, are generally not willing to return home to do farm work (Tang & Feng, 2015; Yue et al., 2010). Instead, they prefer to take a more “decent” nonagricultural job in the countryside when they have to return (Yue et al., 2010; Zhao, 2002). With the return of migrants, capital, skills, and social networks brought back to the rural home may drive entrepreneurial activities and create jobs (Démurger & Xu, 2011; Ma, 2001; Murphy, 1999).

Return migrants are not necessarily going back to the rural home; cities and towns near the rural home are also popular options. However, little academic attention has been paid to return migration related to destinations other than the rural home. Cities and towns in the home region have different characteristics and circumstances from the major cities that attract and accommodate massive amounts of rural migrants (Hao & Tang, 2017). In general, these urban destinations provide fewer employment opportunities, lower wages, and inferior urban infrastructure and services than big cities. However, the living costs and degree of competition and work-related stress are considerably lower (Chan, 2014; Hao & Tang, 2017; Tang, Hao, & Huang, 2016). Though small cities are generally less attractive to rural migrants (Tang & Feng, 2015; Zhu & Chen, 2010), the dramatically rising living costs (especially housing prices), unstable employment, and discriminatory policies in large cities increasingly motivate rural migrants to leave for home or alternative migration destinations. Compared to the countryside home, public amenities and services, such as education and health care, are better in urban areas. That is largely because of the uneven distribution of public resources between rural and urban areas and the concentration of large infrastructure and public facilities in urban areas. Consequently, urban areas in the home region are considered by many prospective returnees as an ideal place to settle down. Nevertheless, among the prospective returnees, what drives their destination choice is yet to be understood and will be examined next.

Data and methodology

Study area

To examine the settlement and return intentions of rural migrants, a survey was conducted in 2015 in Nanjing and Suzhou—two major migration destinations in Jiangsu Province (Figure 1). The two cities were selected for two reasons. First, Nanjing and Suzhou, with 6,587 km² and 8,657 km² of area, respectively (Jiangsu Provincial Bureau of Statistics, 2016), are major cities in the Yangtze River Delta, one of China’s most developed regions. Abundant employment opportunities and relatively high earnings attract rural migrants from different provinces in China. In 2015, of the 8.24 million urban inhabitants living in Nanjing, 1.73 million were migrants; in Suzhou, among the 10.62 million inhabitants, approximately 40% (3.95 million) were migrants (Jiangsu Provincial Bureau of Statistics, 2016). Second, the two cities have different economic structures. Nanjing, the capital city of Jiangsu is a traditional industrial base for many sectors, including petrochemical, machinery, and textile industries. With the economic restructuring in recent decades, the tertiary industry already accounted for 57% of the economy in 2015. Suzhou is an important city for foreign direct investment and export-oriented industries. The secondary industry still accounts for 49% of the gross domestic product and employs a large number of migrant workers (Jiangsu Provincial Bureau of Statistics, 2016). The disparities in the industrial structures of the two cities determine some distinct characteristics of the two cities and their labor forces. In sum, the presence of large numbers of rural migrants in the two cities and the structural difference of their labor markets provide a large and diverse sample for the exploration of the settlement and return intentions of rural migrants.

Figure 1. Jiangsu Province in China (left) and Nanjing and Suzhou in Jiangsu (right).

Data and survey

Using a stratified sampling strategy, four inner-city districts (Xuanwu, Gulou, Qinhuai, and Qixia) and two suburban districts (Luhe and Jiangning) were selected from the 11 districts of Nanjing (Figure 2), and three inner-city districts (Gusu, Huqiu, and Wuzhong), one suburban district (Wujiang), and one county-level city (Kunshan) were selected from the five districts and four county-level cities in Suzhou (Figure 3) for drawing the sample. The sixth national census conducted in 2010 was adopted to determine the sample size for each administrative unit (i.e., district- or country-level city) proportional to its migrant population.

Figure 2. Map of Nanjing.

Figure 3. Map of Suzhou.

The census, which includes all residents living in the cities for more than 6 months, indicates that most migrants engage in manufacturing, construction, and service sectors. Respondents were randomly selected in 18 industrial areas and 8 construction sites from a collection of 50 industrial areas and 30 construction sites that have the highest concentration of migrants. To sample service sector workers, 22 urban neighborhoods that accommodated abundant economic activities were selected. For each site selected for sampling, no more than 25 rural migrants were sampled from the official registration records. The overall response rate of the survey was 92%. The final valid sample size was 1,065, with 441 respondents in Nanjing and 624 respondents in Suzhou. The sample size accounts for approximately 0.02% of the total migrant population in each of the two cities.

In the questionnaire survey, each respondent was asked to select a place where they intended to settle among four exhaustive options: (a) the city they currently live in, (b) their rural home, (c) a local town or city near their rural home, or (d) other places. Because the number of respondents who chose other places was only 50, this group was excluded in the statistical analysis. The final sample includes 1,015 respondents.

In addition to the questionnaire survey, a total of 40 rural migrant respondents (18 in Nanjing, 22 in Suzhou) were identified according to their personal attributes (20 males and 20 females; aged between 16 and 65) and employment types (20 in manufacturing or construction sectors, 16 in service sectors, and four in other sectors) for in-depth interviews. These interviews were carried out during and after the survey in 2015 and 2016. Each interview lasted about 45 to 60 min, recorded by taking notes. These interviews were used to understand the decisions made at different stages of migration as well as the considerations that drive future migration intentions.

Theoretical model

To explore the reasons for different return intentions, a theoretical model was created to include three aspects of independent variables, namely, geographic features, sociodemographic attributes, and rural landholdings of rural migrants. Different from most prior studies that focus on the socioeconomic factors of return migrants (Hare, 1999; Wang & Fan, 2006; Zhao, 2002), the inclusion of geographic factors takes into account the characteristics of the origin and destination of rural migrants. The sociodemographic attributes of rural migrants include age, gender, marital status, educational attainment, and working and living conditions. The variable age is considered by the cohort differences between the born-after-1980 generation and their older counterparts because distinct cohort characteristics are more significant than the effect of age (Y. Liu, Li, & Breitung, 2012). The born-after-1980 generation generally has less agricultural experience, better human capital, and a stronger desire for urban settlement than their predecessors (Tang & Feng, 2015; Yue et al., 2010). Education attainment and occupation represent the human capital of rural migrants. Whether one has children or parents living in the origin location highlights the role of family obligations in migration decisions. Marital status (including whether living with the spouse) and living conditions (including whether living with relatives or friends) indicate the family composition and social networks of rural migrants at the destination. The rural landholdings of rural migrants in the home countryside involve both farmland and rural housing land. According to prior research (Hao & Tang, 2015; H. Li & Zahniser, 2002; Roberts, 1997; Tang et al., 2016), rural landholdings are closely related to migration and settlement decisions. Consequently, rural landholdings are hypothesized to influence return migration as well as the return destination.

Floating or return

Floaters, rural returnees, and urban returnees

According to the intended location of permanent settlement, rural migrants are classified into three groups: (a) rural returnees who are prospective return migrants intending to return to their home village, (b) urban returnees who are prospective return migrants intending to settle in a local town or city near their home village, and (c) urban floaters who are migrants intending to stay in the current city. Different from urban settlers who are defined as those who have a formal local urban hukou (Chan & Buckingham, 2008; Hu, Xu, & Chen, 2011), urban floaters do not hold a local urban hukou, although they may have intentions of urban settlement. Only when rural migrants are granted a local urban hukou are they officially considered local urban residents with equal rights to urban resources. As a result, for rural migrants without a local hukou, their intentions of permanent settlement or even homeownership in the host cities do not equate them with urban settlers.

The three subgroups of rural migrants exhibit different patterns in terms of geographic, sociodemographic, and landholding attributes (Table 1). Among all respondents, more than half (52.3%) migrated from regions outside Jiangsu. In terms of the location of urban settlement, 53.4% of respondents live in the inner districts of Nanjing and Suzhou, whereas 46.6% live in either suburban districts or subordinate county-level cities. They tend to be married males and have an average age of 34. Most of them have attained secondary or tertiary education (86.5%) and earn a salary lower than 4,000 yuan per month (57.2%), in contrast with the average urban salary of 6,579 yuan in Nanjing and 6,055 yuan in Suzhou (Jiangsu Provincial Bureau of Statistics, 2016). Most rural migrants work in labor-intensive manufacturing or service sectors (87.7%) and live in either rental housing (57.8%) or rent-free dormitories (21.7%). The majority of respondents (83.1%) have parents in home villages, whereas 27.2% of respondents have dependent children in the rural home. In terms of rural landholdings, 46.7% of respondents have farmland, and 74.3% have rural housing land.

Table 1. Descriptive statistics of rural returnees, urban returnees, and urban floaters.

Variable	Category	Total	Rural returnee	Urban returnee	Urban floater
Region of origin	Outside Jiangsu	52.3	60.1	49.4	40.8
	Within Jiangsu (=ref.)	47.7	39.9	50.6	59.2
Level of destination area	Inner-city district of prefecture-level city	53.4	49.0	60.0	55.4
	Suburb district of prefecture-level city or county-level city (=ref.)	46.6	51.0	40.0	44.6
Gender	Male	57.6	66.7	43.4	53.3
	Female (=ref.)	42.4	33.3	56.6	46.7
Age	≤35 (=ref.)	58.2	57.5	58.7	59.2
	>35	41.8	42.5	41.3	40.8
	Mean (years)	34.0	34.5	32.6	34.1
Marital status	Married and together	43.2	43.1	38.7	47.0
	Married not together	16.6	21.2	16.6	8.7
	Single (=ref.)	40.2	35.7	44.7	44.3
Education	Primary school and below (=ref.)	13.5	17.5	7.7	11.2
	Middle school	56.7	66.7	53.6	41.8
	College and above	29.8	15.8	38.7	47.0
Monthly income (yuan)	<4,000 (=ref.)	57.2	58.9	54.0	57.5
	4,000–6,000	30.3	31.9	31.5	26.5
	>6,000	12.5	9.2	14.5	16.0
Occupation	Worker (=ref.)	53.1	60.5	49.8	42.9
	Service personnel	34.6	33.9	29.8	39.7
	Clerical personnel	12.3	5.6	20.4	17.4
Living duration	<0.5 year (=ref.)	7.1	7.1	8.6	5.9
	0.5–2 years	37.3	35.9	41.4	36.6
	>2 years	55.6	57.0	50.0	57.5
Living condition	Self-owned housing	17.8	14.5	8.1	31.4
	Housing of relatives and friends	2.7	2.0	5.5	1.4
	Rent-free dormitory provided by workplace	21.7	20.4	29.8	17.0
	Rental housing (=ref.)	57.8	63.1	56.6	50.2
Parents at origins	Have parent at origins	83.1	82.9	88.9	79.1
	Have no parent at origins (=ref.)	16.9	17.1	11.1	20.9
Children at origins	Have children at origins	27.2	34.5	28.4	13.9
	Have no children at origins (=ref.)	72.8	65.5	71.6	86.1
Farmland	Have no farmland (=ref.)	53.3	50.4	55.7	56.8
	Have farmland	46.7	49.6	44.3	43.2
Rural housing land	Have no rural housing land (=ref.)	25.7	20.2	30.6	32.1
	Have rural housing land	74.3	79.8	69.4	67.9
Number		1,015	496	232	287

Of the 1,015 respondents, 496 (48.9%) are rural returnees, 232 (22.8%) are urban returnees, and 287 (28.3%) are urban floaters. Compared to urban returnees and urban floaters, a larger proportion of rural returnees come from outside Jiangsu and live in the suburban districts of the prefecture-level cities or country-level cities. Rural returnees are less educated and paid less than urban returnees and urban floaters. Most rural returnees (60.5%) are manufacturing workers, whereas only 5.6% are clerks compared to urban returnees (20.4%) and urban floaters (17.4%). Though both urban returnees and rural returnees chose to leave the current host city, in quantitative terms, urban returnees are more similar to urban floaters than to rural returnees. For instance, 38.7% of urban returnees and 47.0% of urban floaters have attained college or above, compared to 15.8% of rural returnees. Larger shares of urban returnees (14.5%) and urban floaters (16.0%) earn a salary of at least 6,000 yuan more than that of rural returnees (9.2%). These results suggest that, in general, rural returnees are more disadvantaged than urban returnees and urban floaters in terms of education, income, and occupational skills.

In addition, the situations of family ties and rural landholdings in the place of origin vary among the three subgroups. Higher proportions of rural returnees (34.5%) and urban returnees (28.4%) have children in the home villages than urban floaters (13.9%). With respect to rural housing land, 79.8% of rural returnees have rural housing land, compared to lower proportions of urban returnees (69.4%) and urban floaters (67.9%).

Intention of return migrants

To further explore the disparities among migrants with different settlement intentions, we explore the association between each variable and the settlement intention when the other variables are controlled. Multinomial logit regression and analysis of variance (ANOVA) are used to analyze the data. To supplement the statistical analysis, qualitative data from in-depth interviews are used to interpret the patterns identified by statistical analysis.

Table 2 reports the results of the multinomial logit model that identifies the variables that lead to the different intentions of the prospective rural returnees, prospective urban returnees, and prospective urban floaters. The group of prospective urban floaters is chosen as the reference group for the dependent variable. The estimated coefficients indicate the effect of each independent variable on the likelihood of returning to the home village or to an urban area near the home village in comparison to floating in the city. The estimated model is significant at the .001 level.

Table 2. Multinomial logit regression for settlement intention of rural migrants.

Independent variable	Rural returnee	Urban returnee
Geographic factors
Region of origin
Outside Jiangsu	0.504** (1.655)	0.295 (1.343)
Inside Jiangsu (=ref.)
Level of destination area
Inner-city district of prefecture-level city	−0.413** (0.662)	0.258 (1.294)
Suburb district of prefecture-level city or county-level city (=ref.)
Socioeconomic factors
Gender
Male	0.160 (1.173)	−0.808*** (0.446)
Female (=ref.)
Age
≤35 (=ref.)
>35	−0.268 (0.765)	0.219 (1.245)
Marital status
Married and together	−0.047 (0.954)	−0.339 (0.713)
Married not together	0.177 (1.194)	−0.320 (0.726)
Single (=ref.)
Education level
Primary school and below (=ref.)
Middle school	−0.179 (0.836)	0.560 (1.751)
College and above	−1.603*** (0.201)	−0.031 (0.970)
Monthly income
<4,000 (=ref.)
4,000–6,000	0.164 (1.178)	0.318 (1.375)
>6,000	−0.066 (0.937)	0.569 (1.766)
Occupation
Worker (=ref.)
Service personnel	−0.346 (0.707)	−0.436 (0.646)
Clerical personnel	−0.747** (0.474)	0.369 (1.446)
Living duration
<0.5 year (=ref.)
0.5–2 years	0.091 (1.095)	−0.027 (0.974)
>2 years	−0.034 (0.967)	−0.203 (0.817)
Living status
Self-owned housing	−0.578** (0.561)	−1.309*** (0.270)
Housing of relatives and friends	0.436 (1.546)	1.441 (4.226)
Rent-free dormitory provided by work place	−0.305 (0.737)	0.434* (1.543)
Rental housing (=ref.)
Parents at origins
Have parent at origins	0.275 (1.316)	0.440 (1.552)
Have no parent at origins (=ref.)
Children at origins
Have children at origins	0.698** (2.010)	0.929*** (2.533)
Have no children at origins (=ref.)
Rural land factors
Farmland
Have no farmland (=ref.)
Have farmland	−0.075 (0.928)	−0.076 (0.927)
Rural housing land
Have no rural housing land (=ref.)
Have rural housing land	0.395* (1.484)	0.107 (1.113)
Constant	1.479***	0.162**
Chi-square	290.229***
−2Log likelihood	1,751.44
Nagelkerke R^2	0.384
Sample size	1,015

Note. Odds ratios in parentheses. Reference category for the dependent variable: urban floaters.

*p < .05. **p < .01. ***p < .001.

Compared to urban floaters, rural returnees are more likely to migrate from regions outside Jiangsu, and they tend to live in the suburban districts of a prefecture-level city or county-level city. These results suggest that both rural origins and urban destinations influence the intentions to return or stay. Migrants from outside Jiangsu are comparatively more disadvantaged, and those who migrate within the province can benefit more from social capital and cultural background. Consequently, interprovincial migration increases the likelihood of returning to the rural home. In addition, migrants who live in suburban districts or county-level cities have relatively less access to good employment opportunities and urban services and hence are less likely to stay in the city than those who live in inner urban districts.

According to prior studies, the young generation of rural migrants is more inclined to settle in urban areas rather than return to the countryside, which is related to their higher human capital and weaker sense of attachment to the rural home (Hao & Tang, 2015; Tang & Feng, 2015; Yue et al., 2010). Nevertheless, the cohort difference has no significant effect in the model. The changing circumstances may have altered the considerations of young migrants. Since the turn of the new millennium, the central government of China has launched a series of policy guidelines and campaigns, such as “new countryside development,” “urban–rural coordination,” and “beautiful countryside.” Rural areas, especially those in the coastal regions, have been substantially transformed (Long et al., 2011). Many advantages of rural areas, such as a better natural environment, lower living costs, and less stressful lifestyle, are attractive to some young migrants who are dissatisfied with the polluted air and water, high living costs, and fast-paced urban life in the city. Some of them expressed their opinions: “I like the life in my rural hometown, more relaxed than in big cities. And I don’t agree that cities are better than the countryside. The infrastructure and living environment in my hometown have been improved, and I could live in a much bigger house than in the city” (Mr. Li, age 25, manufacturing worker, migrated to Suzhou from rural Shandong Province, personal communication, August 29, 2015).

According to the model results, educational attainment is significantly associated with the return decisions of rural migrants. Rural migrants with a college education are considerably more likely to stay in the city than return to the rural home. The result is consistent with previous research that suggests that inferior human capital prevents some rural migrants from integrating into the urban labor market and society (Borjas, 1989; Wang & Fan, 2006). Likewise, clerks are more inclined to live in the current host city rather than return to the rural home, suggesting a positive effect of higher employment and social status (Zhu & Chen, 2010). For those who take office jobs, the countryside is less attractive because equivalent employment opportunities and amenities are absent there. The insignificant effect of living duration in the host city indicates that a longer migration spell does not necessarily motivate people to settle. This is because the effect is likely opposite for two types of rural migrants. Those migrants who are more marginalized in the host society might be more discouraged to settle as the duration of residence gets longer, whereas those migrants who have more chances to integrate might develop the opposite intention.

Rural migrants who own housing in their host city are more likely to be urban floaters than rural returnees. Among the migrant population, homeowners usually have more human and financial resources than others (Wu, 2002). Most rural migrants are excluded from affordable housing options; hence, they can only buy commodity housing to become homeowners, which suggests that homeowning migrants have considerably stronger purchasing power than their peers (Wang & Fan, 2012). In addition, for rural migrants, owning an urban home usually implies a strong motivation to stay in the city. Consequently, the group is less likely to return to the home countryside. However, family separation drives rural migrants to return. Rural migrants who have children in the place of origin are particularly more likely to return. The result is consistent with other research findings that fulfilling family obligations, such as marriage and childcare, is an important reason for return migration (Wang & Fan, 2006).

In terms of rural landholdings, the farmland has no significant effect, whereas rural housing land is a positive determinant in the intention of returning to the rural home. The result contrasts with previous research that highlights the importance of agricultural employment for return migration (Hare, 1999; Zhao, 2002). Today, agricultural production seldom is a crucial driver for rural migrants to return, largely because of the negligible income from working in the fields. In addition, the young generation of rural migrants generally have no farming experience (Yue et al., 2010), which reduces the likelihood of considering farmland as a means of employment or social security. Even for those who consider farmland as a safety net when they experience difficulties in cities such as unemployment or illness, many of them would return to a local town to find nonagricultural jobs (B. Li, 2006; Nielsen, Smyth, & Zhai, 2010).

In contrast, rural housing land is still considered by most rural migrants as a safe haven. On the one hand, the rural home is a place where rural migrants can always return when encountering difficulties. On the other hand, owning a house at the home countryside maintains the sense of security and autonomy, which is essential for those rural migrants who lack access to resources and power in the city. Because rural citizens have the exclusive rights to ownership of rural land and housing, the dream of “fallen leaves returning to their roots” could be fulfilled for those who retain their rural hukou. As a rural migrant mentioned, “I want to return to my rural hometown when I am older. … I am not going to do farm work then, just build a house on my housing plot and spend my twilight years” (Mr. Li, age 40, bookstore owner, migrated to Nanjing from rural Anhui Province, personal communication, April 18, 2016).

As for urban returnees, gender, living status, and family separation are important driving factors. Men are less likely than women to become urban returnees. That is probably because female migrants are more apt to return at certain stages of their life cycle, such as marriage and childbearing (Fan, 2000). For them, a local town or small city close to the rural home is an ideal place to balance employment and family obligations. Homeownership is strongly associated with the intention of settling down, because living in dormitories motivates migrants to return. Dormitory tenants are mostly manufacturing or construction workers, who have limited human capital and are less likely to settle in the host city. In addition, according to the survey, most manual workers are from less-developed regions. Despite fewer and inferior employment opportunities and lower standards of amenities, much lower living costs and fewer institutional restrictions in the urban areas of origin motivate these “blue collar” migrants to return.

Rural migrants who have children in the home village are 2.5 times more likely to become urban returnees than urban floaters. Urban areas at the origin are ideal places for those who have dependent family members because they can enjoy many urban resources such as urban jobs and amenities and at the same time benefit from low living costs and abundant local social networks in fulfilling their family obligations. Many rural migrants shared such considerations: “I plan to return to the county town near my home village, because my parents, wife and children are all in the village” (Mr. Wang, age 42, decoration worker, migrated to Nanjing from rural Shandong Province, personal communcation, August 25, 2015); “The housing price in Nanjing is too high. I want to open my own hairdressing shop in the city near my rural hometown and live together with my parents” (Mr. Xiao, age 20, hairdressing apprentice, migrated to Nanjing from rural Anhui Province, personal communication, April 29, 2016).

In addition to the attributes examined in the regression model, the three groups of rural migrants have distinct subjective perceptions and attitudes related to their urban life. A set of one-way ANOVA tests was employed to explore these subjective factors. The tests use the answers to a set of questions about their self-evaluation on the living experience in the city. The answers are based on a 5-point scale ranging from 1 (least satisfied) to 5 (most satisfied). The results are shown in Table 3. A significant p value indicates strong disparities between the three groups in terms of each aspect.

Table 3. ANOVA statistics of self-evaluations of rural migrants.

Aspects of evaluation	Rural returnee		Urban returnee		Urban floater		F-value	Sig.
	Mean	SD	Mean	SD	Mean	SD
Working pressure	2.68	1.44	3.00	1.14	3.09	1.26	10.358	.000
Living cost	2.41	1.35	2.50	1.19	2.65	1.24	3.266	.039
Dialect proficiency	3.63	1.21	3.73	1.08	3.84	1.13	3.098	.046
Living custom	3.58	1.12	3.58	1.11	3.74	1.08	2.413	.048
Being respected	2.94	0.96	2.94	0.93	3.18	0.99	1.272	.281
Contact with local citizens	3.72	1.11	3.77	1.07	3.87	1.10	3.251	.041

Among the six aspects tested, statistically significant differences between the three groups were identified in “working pressure,” “living cost,” “dialect proficiency,” “living custom,” and “contact with local citizens.” In general, rural returnees had lower scores than other subgroups, indicating that migrants who are least satisfied with their working and living conditions in the host society are most likely to return to the rural home. The result also confirms that rural migrants who are not adapted to the urban society are likely to return to the original rural community (Wang & Fan, 2006). For urban returnees, the scores are generally higher than those for rural returnees but lower than those for urban floaters. This indicates that despite the similarities between prospective urban floaters and urban returnees, language, customs, living conditions, and social interactions still prevent urban returnees from fully integrating into the host society. Consequently, they are motivated to move to the urban areas of their origin, where dialect and customs are familiar and higher levels of quality of life and employment opportunities are available. As Du and Li (2012) indicated that affective emotion significantly influences the stay–leave intention of migrants. Our study further identifies that such factors not only affect the intention of staying or returning but also exert an effect on the destination choice of return.

Conclusions

Return migration is an important component of China’s rural–urban migration. Different from the existing literature, which focuses on the prominent counter flow from urban destinations to rural origins, this article distinguishes two main streams of the counterflow: return migrants to the home village (rural returnees) and return migrants to a local town or city near the home village (urban returnees). Through examining the intentions of rural migrants in Nanjing and Suzhou, prospective rural returnees, urban returnees, and urban floaters are identified and their socioeconomic attributes are compared. Given that the survey looks at prospective return migration rather than actual migration, it is worth noting the possibility that actual migration flows of the migrant population may differ from the patterns identified in the analysis. Nevertheless, the significant associations between socioeconomic attributes and return intentions suggest some important and dissimilar driving factors of the two main types of return migration.

A larger share of rural migrants intends to become rural returnees (48.9%) than urban returnees (22.8%) and urban floaters (28.3%). Among the rural migrant population, prospective rural returnees tend to be those migrants who are the most disadvantaged and least satisfied with their urban circumstances. They usually have lower-than-average human capital, social resources, and subjective well-being among their peers. When they encounter difficulties in settling in the host city, they resort to the safe haven of their village of origin. In contrast, prospective urban returnees share more similarities with urban floaters. Both groups are apt to make a better life in urban areas, though family obligations and attachment to the origin motivate prospective urban returnees to move to the home region. For both rural and urban returnees, however, homeownership in the urban destination and family members in the rural home significantly influence their return intentions, indicating the importance of the “assets” in both urban and rural sites.

Existing literature on return migration focuses on the narrow definition of return: returning to the home village. By including the migrants who intend to return to a local town or city near the rural home, we expand the definition of return migration. In addition, by examining the settlement and return intentions of rural migrants in the host cities, we theoretically exhaust all possible outcomes of the rural migrants’ further geographic mobility. Consequently, we can compare migrants who intend to return with others, instead of examining the group of return migrants alone. The findings suggest that the decision to return is a trade-off between rural and urban livelihoods, but it also aims to optimize utilities based on the socioeconomic attributes, preferences, and the family background of migrant households. Through comparing rural returnees and urban returnees, this article broadens the understanding of the decision making of return migrants in the Chinese context. In such decisions, the agency of rural migrants is manifest in two aspects. First, despite the difficulties and sacrifice they suffer, most rural migrants, whatever their return intentions, continue to strive in the cities to make a livelihood. Their individual presence and actions have been collectively changing the economic and political landscape of the cities. Second, their diverse intentions in terms of the settlement place reflect the growing desire and capacity of rural migrants in changing the “make money in the city and return to the rural village” model. A growing number of rural migrants, although they do not necessarily settle in the cities that currently host them, are determined and capable of leaving their rural roots for good.

China’s rural migration has changed the social and policy context. The implementation of “rural–urban coordinated development” guidelines has improved the living environment and amenities of many villages in recent years, driving a growing number of return migrants. In addition, new employment opportunities emerged in some rural areas, where industries were established to exploit local resources, such as a scenic landscape, fresh agricultural products, and traditional customs and culture. Different from previous studies that highlighted the effect of agricultural activities on return decisions (Hare, 1999; Ma, 2002; Zhao, 2002), the function of the rural economy is weakening, whereas the role of rural housing remains significant. This is especially the case for the young generation of rural migrants, who have less farming experience. In addition to the impact of institutional controls and market forces, the personal preferences of rural migrants, such as lifestyle choices and preferred social environment, have played an increasingly important role in determining migration and settlement decisions. Most research suggests that the institutional and social inferiority of rural migrants in the city compelled them to return (Wang & Fan, 2006; Zhao, 2002), yet this study finds that a proportion of rural migrants actually voluntarily return to their home village or a smaller local town in pursuit of a better quality of life. These new circumstances emerged in recent years, and in some areas they not only reflect the agency of individual rural migrants in the changing context but may suggest some future development trends of internal migration in China.

In the 2010s, the central government of China proposed the “new-style urbanization” policy, highlighting urban–rural integration and coordinated development among urban settlements of different sizes. A series of reform measures were introduced in both urban and rural areas, such as the residence card and the rural property right system (General Office of the CPC Central Committee & General Office of the State Council of PRC, 2015; The State Council of PRC, 2014). These measures aim to grant rural migrants more rights to their rural assets and access to urban public services. In addition, the policy agenda has emphasized the development of smaller cities and towns. Improved infrastructure, investment environment, and employment opportunities and relatively low living costs have become increasingly attractive for rural migrants. These urban destinations will certainly provide rural migrants with alternatives to the large cities for settlement based on their personal and familial circumstances.

To facilitate the mobility of the migrant population, policy reforms are in demand to further reduce the institutional barriers and narrow the gap between different regions in terms of the economic vitality and the living environment. In addition, more responsive and differentiated policies should be formulated to address the fact that different subgroups of rural migrants have different aspirations and prospects for their future life. Channels for social mobility should be broadened and diversified to facilitate the agency of migrants with different capabilities. More equitable urban policies are crucial to allow rural migrants to be socially mobile, and a more untrammeled and flexible policy environment for migration (and return) will enable rural migrants to be more geographically mobile in the pursuit of personal aspirations and a better life. These measures will contribute to the well-being and social development of the migrant population to achieve the goal of “people-oriented urbanization.”

Acknowledgments

We thank the anonymous reviewers and the editor for their constructive comments.

Additional information

Funding

This research is supported by the National Natural Science Foundation of China (41401175, 41401167), the Natural Science Foundation of Jiangsu Province (BK20141325), Hong Kong Baptist University (FRGII/15-16/026), and the Fundamental Research Funds for the Central Universities (020914380032).

Notes on contributors

Shuangshuang Tang

Shuangshuang Tang is an Associate Researcher at the Research Centre of Human Geography, School of Geographic and Oceanographic Sciences, Nanjing University. Her main interests are migration, urbanization policies, and regional planning.

Pu Hao

Pu Hao is an Assistant Professor in the Department of Geography at Hong Kong Baptist University. His research interests include urban sociospatial dynamics, urban economic geography, migration, and urban housing. His current research projects investigate unplanned commercial establishments in Chinese cities and the geographic and social mobility of rural migrants.