1 Introduction

Subjective well-being assumes a central role in social scientists’ theoretical and empirical understanding of social inequality. Since subjective well-being refers to people’s multidimensional perceptions of their lives, including both a relatively stable general judgment and more momentary affective evaluations (moods and emotions), it is well acknowledged to be an essential complement to standard of living as an indicator of social well-being wellbeing (as reviewed by Fleurbaey 2009). Therefore, longstanding literature has focused on subjective well-being as an important indicator of development outcome of interest, examining the effects of a wide array of economic, cultural, and social characteristics in international comparative studies (Diener et al. 1993, 1999, 2009). Moreover, in recent decades, social scientists have invested considerable research effort in unraveling the micro-level determinants of subjective well-being, seeking a better understanding of the social group differences in subjective well-being by various demographic, psychological, and socioeconomic traits (Ballas and Tranmer 2012; Easterlin 2001; Veenhoven 2008; Helliwell and Putnam 2004; Kahneman and Deaton 2010; Bellani and D'ambrosio 2011).

Within the massive subjective well-being research, a fast-emerging subset has concerned the disadvantaged attainment of happiness associated with migrant status. Although existing scholarly knowledge of the decision-making process of migration generally sees migrants as rational actors who seek to maximize their pecuniary utility (the relevant literature is duly influenced by Harris and Todaro 1970, e.g., Massey et al. 1993; Hare 1999; Chiswick 1999; Zheng et al. 2020), studies pertinent to subjective well-being generally yield a rather pessimistic finding. On the one hand, migration per se is found to be a stressful event (Magdol 2002; McCollum 1990); on the other, generally positive labor market outcomes associated with migration do not necessarily lead to a better experience of the quality of life, especially in the long run (Ek et al. 2008; Nowok et al. 2013; Hendriks et al. 2016). Due to the difficulties in adapting to the different social circumstances and establishing new social ties in the destination cities, the “unhappy migrants” phenomena has caused concern because of its substantial policy implications (Diener and Seligman 2004).

We look to the Chinese rural-to-urban migration to examine the important association between migrant status and subjective well-being. Although “rural migrant workers” have been receiving heated analytical attention, scholars have particularly emphasized their disadvantaged objective well-being in city life and their chances of overcoming institutional hurdles of the hukou registration system (Chan and Buckingham 2008; Huang et al. 2017; Li 2006; Shen 2017). Relatively few analytical efforts have examined their subjective well-being, and even fewer are devoted to analyzing the underlying mechanisms resulting in these subjective stratification outcomes (Knight and Gunatilaka 2010; Liu et al. 2017a, b).

Our research context of China offers several crucial advantages. Foremost among these is the unprecedented magnitude of the Chinese migration process. According to the last four censuses, the size of the “floating population” in China has surged from 21 million in 1990 to 261 and 376 million in 2010 and 2020, respectively; and most of these are aforementioned rural migrant workers (National Bureau of Statistics in China 1993, 2011, 2021). Understanding how these rural migrants evaluate their lives is of vital theoretical and policy importance because it is not only a matter of happiness for almost 5% of the world population, but also a profound source of social inequality in Chinese society.

Second, the loss of subjective well-being associated with migration status might be pronounced in the Chinese context, given that as one of the most fundamental redistributive systems—the household registration (hukou) system—still functions today, endowing rural migrants and urban locals with essentially unequal social benefits (Solinger 1999). A growing body of literature suggests that this institutional discrimination against rural hukou status has given rise to multidimensional social exclusion. A notable case is residential segregation, referring to the fact that rural migrant workers tend to inhabit affordable dwelling units that are geographically separated from urban neighborhoods, most typically in those “villages in the city” (chengzhongcun) that are located either in the remote outskirts or crowded downtown segments (Liu et al. 2018; Zhu 2016). Occupational segregation is also prevalent in that rural migrant workers suffer from substantially disadvantaged access to elite and rewarding occupations, due to their lack of advanced educational credentials, vocational skills, and, probably most fundamentally, urban hukou status (Zhang and Wu 2017; Meng and Zhang 2001). If the segregation in residence and occupational attainment could be considered as the direct objective outcomes of hukou discrimination, an indirect long-term effect points to the psychological feeling of being excluded: with the substantial impacts of hukou status on people’s everyday life, rural migrant workers are treated as de facto second-class citizenry, and many of them suffer from a feeling of vulnerability, a lack of social support, and a sense of belonging (Li and Rose 2017; Zhong et al. 2017). Our research context of Chinese internal migration is, therefore, valuable, enabling us to examine empirically the impact on subjective well-being of social exclusion under institutional barriers.

Third, a distinguishing feature of Chinese rural-to-urban migration is its lack of long-term commitment, hence the name “floating population” (Liang and Ma 2004). Commonly, Chinese rural migrant workers move back and forth strategically to juggle urban employment opportunities with their farm job and family responsibilities. This temporary aspect of Chinese domestic migration has significant implications on the subjective well-being of rural migrant workers. On the one hand, social integration into destination cities tends to be more difficult for short-term migrants than their long-term counterparts, as the building of identity, trust, and belonging usually takes a long period (de Haas and Fokkema 2011). On the other, despite the fact that return migration is prevalent and also typically temporary (Zhao 2002), scarce literature has investigated the extent to which return migrants are integrated into or excluded from their origin rural communities and the subjective well-being outcomes of temporary return migration.

Using data from the China Labor Dynamics Survey, this study examines the role of social exclusion on the subjective well-being of both current migrants and return migrants. We find that the temporary nature of migration behaviors in China due to the institutional constraints of the hukou registration system has brought rural migrant workers into confrontation with an awkward dilemma: they are socially excluded not only in destination cities but also in origin villages and, therefore, suffer from impaired subjective well-being whether they stay in the city or return to the countryside. In particular, our empirical results suggest: (a) the economic gains of rural-to-urban migration come at the expense of subjective well-being, and, moreover, this psychological cost persists even after migrants’ return to their villages of origin; (b) the negative evaluation of rural migrants workers on their life is highly correlated with their generally low self-identified social class and their sense of unfairness in socioeconomic status attainment; (c) social exclusion plays an important mediating role between migration status and subjective well-being. In particular, the lack of a social support network, the feeling of vulnerability, and the deficiency in the sense of community all contribute to rural migrant workers’ disadvantaged subjective well-being; (d) the difficulties of getting integrated not only hamper the subjective well-being of migrant workers in urban cities but also of peasants who have returned to the countryside.

This article broadens the literature by explicitly studying the role played by social exclusion in the linkage between migration status and subjective well-being. Specifically, we attempt to identify the mechanism through which migrants are deprived of their subjective well-being, paying extra attention to the effects of social support and the feeling of vulnerability and belonging. We also take into account the temporary nature of Chinese rural-to-urban migration to compare not only current migrants but also returned migrants with their non-migrant rural counterparts.

The remainder of the paper is organized as follows: Sect. 2 offers a brief retrospective of the existing literature and based upon that, the hypotheses we propose. In Sect. 3, we describe our data, sample, the construction of pertinent variables, and the specification of models. In Sect. 4, we present the relevant descriptive statistics and empirical results. Section 5 summarizes the empirical findings and offers a discussion on policy implications.

2 Literature Review and Research Hypotheses

The concept of subjective well-being (hereafter, SWB) refers to people’s evaluation of their own state of life, including life satisfaction, mental health, and affect (Diener 1984). While many earlier studies mainly focus on the objective quality of migrants’ life (Solinger 1999; Li 2006; Démurger et al. 2009), in recent years, scholars have given rural-to-urban migrants’ subjective well-being increasing attention (Knight and Gunatilaka 2010; Cheng et al. 2014). Among the studies looking at the factors that affect SWB, some center on individual-level factors such as socioeconomic status, migrant remittance (Akay et al. 2014), and relative deprivation (Knight and Gunatilaka 2010; Jin 2016), while others pay attention to contextual factors such as the neighborhood social environment (Wen et al. 2010; Liu et al. 2017a, b).

Although many studies have demonstrated the disadvantages of rural migrant workers relative to urban locals in terms of SWB attainment, only a handful compare SWB between migrants and non-migrants among rural citizens (Li et al. 2007; Jin 2016). Given that migrant workers and urban local residents may differ not only in migration experience, but also in hukou status, childhood environment, and cultural adaption, extra caution is required if we are to attribute the observed group differences to migration status per se. This study instead defines its analytical sample as rural citizens born with agricultural hukou: we first compare rural-to-urban migrants’ SWB with that of their rural non-migrant counterparts; and then, to reveal the lingering effects of migrant behavior on SWB, we further incorporate returned migrants into our comparative framework.

Not only is the examination of SWB an important stratification outcome of interest, but this research also attempts to estimate the mediating effect of social exclusion on the relationship between migration status and SWB. However, what makes our empirical strategy challenging is that social exclusion is a multidimensional concept and is only ambiguously defined in the existing literature. With various definitions in different disciplines and contexts, social exclusion refers to a variety of group-specific disadvantages such as systematic lack of resources and opportunities, denial of rights and services, the inability to participate in social relationships and activities, and the difficulty in realizing full citizenship.

The concept of social exclusion can be traced to Max Weber and to Frank Parkin’s theory of “social closure,” which denotes a process by which social groups seek to monopolize economic resources and opportunities and to maximize their rewards (Parkin 1979, p. 44). Developed in 1970s against the background of European integration and the decline of welfare states, social exclusion has been attracting attention from social observers and policy makers, gradually becoming a fundamental policy discourse that has drawn heated theoretical and empirical debates (Silver 1995; Gordon et al. 2000; Sen 2000). For example, the European Union considers reducing social exclusion as an important social policy agenda item and explicitly requires its member countries to take measures to limit or prevent it (Atkinson and Da Voud 2000). Against this background, the intergroup exclusion across many social domains has been well investigated in European countries, yielding compelling empirical findings that enrich our understanding of the institutional determinants of negative consequences on life chances and the quality of well-being (Levitas et al. 2007).

Due to its multidimensional conceptualization, measuring social exclusion empirically could be difficult. The existing scholarly knowledge of social exclusion has generally explored both individual- and the aggregate-level measures. At the aggregate-level, social exclusion is usually proxied by the level of inequality or the proportion of people at various disadvantaged statuses (Levitas 2006). At the individual level, social exclusion is measured both from an objective perspective—with indicators such as poverty, unemployment, and school dropout to describe deprivation of rights and access to social resources—and from a subjective perspective, with the sense and the experiences of being segregated lying in the locus of the examination. There is also a growing interest in the study of the association between aggregate- and individual-level social exclusion (Chakravarty and D’Ambrosio 2006).

Regarding rural-to-urban migration in China, given the indubitable significance of the household registration system in shaping rural migrants’ well-being in particular and social inequality in general, existing studies have overwhelmingly centered on the influence of hukou status on economic segregation (Tang 2002; Yang 2013). Specifically, a massive literature has comprehensively described rural migrant workers’ disadvantages in the urban labor market, including their underprivileged occupational attainment, income, social services, working conditions, among other aspects (Huang et al. 2010; Zhan 2011; Yue et al. 2013). Essentially, scholars inquire about the extent to which rural migrant workers can overcome the institutional hurdle of the hukou system, seize the chance of moving upward on the urban social ladder and settling down in cities (Zhan 2011). Our research rather attempts to join a handful of recent studies pertaining to the psychological aspect of social exclusion, paying attention to either the experiences of exclusion in social relations, intergroup interactions, and social participation, or the perceptions of social exclusion, including senses of being discriminated against or lacking feelings of belonging.

Therefore, the concept of social exclusion in our research mainly manifests a state of psychosocial marginalization, in which rural migrants find it difficult to integrate into the mainstream social community. As we will show in the following analyses, social exclusion not only bears important implications on the formation of social inequality, but also on migrants’ objective living standards and their subjective evaluation of their lives. Our research is thus relevant along with many surveys and studies focusing on rural migrants’ disadvantaged mental health status (Zhong et al. 2013) and, in particular, the mediating role of social exclusion (Jin et al. 2012; Lin, et al. 2011). For example, Lin et al. find that the experience of discrimination and perceived social inequity are strongly associated with migrants’ mental health problems (Lin et al. 2011). Likewise, Zhang et al. suggest that the experience of being discriminated against directly affects migrants’ evaluation of the quality of their lives through the large, psychologically perceived expectation-reality discrepancy (Zhang et al. 2009).

Previous studies are mainly concerned with migrants’ social integration into destination cities, paying less attention to the issue of integrating returned migrants into their home villages. It is often neglected that upon return from a city of destination, peasants need to be reintegrated into the original rural society and, this reintegration is not always successful. On the one hand, the social and economic environment to which migrants return may have changed considerably, especially against the background of rapid rural development in China. On the other hand, migrants could be acculturated to the cities of destination, making their modes of life, values, preferences, and perception different from their fellow countrymen. Although return migration is an under-researched area, still a longstanding literature has suggested that the adaption between those who have returned and those who remain at places of origin cannot be taken as granted (Nisrane et al. 2017). To fill in this research gap, we enrich the previous comparative framework and compare both current migrants and returned migrants with their rural non-migrant counterparts throughout the analysis. Based upon this framework, we try to examine both the configuration of rural migrants’ perceptions of social exclusion and its mediating role on SWB. In particular, our analysis aims at investigating the following research hypotheses:

Hypothesis 1.

Both current migrants and returned migrants tend to have lower subjective well-being than their rural non-migrant counterparts.

Hypothesis 2.

Social exclusion plays a mediating role between migrant status and subjective well-being. In particular, both current migrants and returned migrants experience social exclusion in their destination cities and home villages, respectively, lowering their subjective well-being.

3 Empirical Strategies

3.1 Data

We use data from the China Labor Dynamics Survey (hereafter, CLDS) conducted by Center for Social Survey at Sun Yat-Sen University in 2012. This survey implements the probability-proportional-to-size (PPS) sampling method and uses a rotating sample to reduce the problem of panel attrition and selection bias (Wang et al. 2017). Beginning in 2012, the survey was conducted biennially, collecting comprehensive data at the individual, family, and community levels, focusing on a wide array of information about the existing labor force (aged 18–64 years), including education outcomes, working experiences, migration, health, and social and political participation. The multilevel feature of CLDS enables researchers to examine how the individual stratification outcomes are influenced by overarching social contextual variables.

We define our analytical sample as respondents who were born in the countryside with rural hukou status, and we exclude those elite rural migrants who have experienced hukou mobility from agricultural to non-agricultural.Footnote 1 Our research subjects are, therefore, rural citizens who still hold their agricultural registration status. As described in the above sections, given the temporary nature of Chinese rural-to-urban migration and the prevalence of return migration, we divide these rural citizens into three groups according to their migration status: (a) current migrants, referring to those who are now migrant workers living in the urban destinations, returned migrants; (b) those who had migration experiences before but now resided in origin villages; and (c) non-migrants with no residential mobility. Our final analytical sample covers 7,274 rural laborers in 2012. Noting that migration status rarely changes in a short period of two years, we only use the CLDS 2012 data as a cross-sectional sample. The empirical results presented in this paper are qualitatively identical if we employ the CLDS 2014 sample for robustness checks.

3.2 Variables

3.2.1 Subjective Well-Being

We proxy for subjective well-being mainly by self-assessed happiness. To explore the intrinsic configuration of SWB, we further incorporate two measures of self-reported social status and perceived social fairness. This multidimensional measuring strategy is premised upon the reasoning that rural migrants’ evaluation of the quality of their lives depends on the extent to which they believe they have climbed up the urban social ladder and consider this social selection process to be fair. All three measures bear important policy implications on social well-being, inequality, and stability.

The self-assessed happiness is measured using the question “in general, how happy do you consider your life is.” The original answer is on a 1–6 scale that we re-scaled to a 1–10 interval to facilitate comparison and interpretation.Footnote 2 The variable of self-reported social status is measured by the question as to “which rung of the social ladder do you think you are on,” which is a continuous variable ranging from 1 to 10. Rural citizens’ perception of fairness is captured by the question of “given your efforts, do you think your current quality of life is fair,” to which the respondents’ answers are also scaled 1–10.

3.2.2 Social Exclusion

We adopt a relatively narrow definition of social exclusion and measure the concept with three variables: social support network, vulnerability to crime, and sense of community. The social support network is measured by a set of questions inquiring about the number of friends from whom the respondent can demand financial, psychological, and cognitive help. The original answers are categorical; however, for the sake of interpretation, we arbitrarily compute the mean of these answers, and, hence, our measure of social support network is a continuous variable, with which a higher value represents a stronger supporting network. The employed proxy is on a 1–5 scale, with one denoting no socially supporting network and five indicating a broadly supporting structure with approximately 16 or more close friends. Using alternative categorical measures does not change our empirical results qualitatively.

We then proceed to the measure of vulnerability. The CLDS inquiries into five different experiences of victimization in the past 12 months, including being robbed, assaulted, extorted, and threatened. Since the likelihood of victimization is generally low in a nationally representative sample, we create a dummy variable, with one denoting the experience of any of these violent crimes and zero otherwise. Sense of community refers to people’s identity as a community member and their feelings of belongings to the neighborhood. The CLDS inquiries about respondents’ acquaintances, cooperation, and trust of others in their neighborhoods. To get a comprehensive measure, we summarize the answers to these questions by taking the mean to proxy for rural citizens’ general sense of community. The variable is on a 1–5 scale, with one denoting none or an extremely weak sense of community and five indicating an extremely strong sense. Our results stay robust if we employ factor analysis to construct an index for sense of community.

3.3 Models

To understand the effect of migration status on rural laborers’ SWB and to explore the mediating effect of social exclusion in the above-mentioned association, we employ a comparative framework that incorporates current migrants, returned migrants, and their rural non-migrant counterparts to estimate empirically a series of group differences in SWB with and without the control of social exclusion measures. Our analysis considers most of the controls used in the previous literature, including various demographic and socioeconomic characteristics. As a key to evaluate how social exclusion mediates the association between migration status and SWB, we focus on the changes in the effect of migration status on SWB before and after the three measures of social exclusion are added into the regression models.

4 Empirical Results

4.1 Descriptive Statistics

Table 1 shows the descriptive statistics of all the dependent, independent, and control variables. Based on the 2012 CLDS sample, approximately 65% of the respondents have never migrated to urban areas, while the current migrants and returned migrants account for 23% and 12% of the sample, respectively.

Table 1 Descriptive statistics of relevant variables
Full size table

We only present the descriptive statistics on SWB and social exclusion measures based upon the whole sample, without looking at the group differentiations by migration status. We will show the relationship between migration status and social exclusion rigorously using regressions in the following section. For SWB measures, the overall mean happiness, fairness, and status scores in 2012 are 6.106, 5.743, 3.968, respectively, all on a 10-point scale. These numbers suggest that although rural citizens are moderately happy, they tend to consider themselves to belong to the underclass. The average social support network score of rural citizens is 2.796 in the 2012 CLDS sample, denoting that the respondents approximately obtain social support from four to six close friends. In terms of vulnerability to crime, the sample means are 0.117 in 2012, suggesting that approximately 12% of the rural respondents have experienced relevant violent crimes. Regarding the sense of community, the sample means of 3.630 in 2012 indicates that rural citizens generally acquire a moderate sense of community.

As for the control variables, the mean age of our rural sample is approximately 44 years old in 2012. 5% of the rural citizens in our sample are Communist Party members. Based upon the CLDS samples, it seems that rural citizens are undereducated. Approximately only 9% of them have a high school education or equivalent, and only about 2% of them have a tertiary education or above.

4.2 Baseline Results: Migration Status and Objective Well-Being

To gauge the implications of our analysis better, we begin our estimation by examining the economic returns of migration status for reference. We take the logarithm of income to achieve more stable estimates. Table 2 reports the pertinent results, suggesting a stable estimation of the statistically significant effect of migration on workers’ objective well-being across models. Model 1 regresses the logarithm of income on both the migration status and control variables such as gender, age, Party membership, and education, to evaluate the net effect of migration status on income. Note that the results are robust if we estimate the reduced-form associations without any control variables. According to Model 1, based on the CLDS 2012 sample, migration status significantly increases rural citizens’ income; however, such economic return decreases as the migrants return to their home villages. In particular, the economic returns to current migrants are five to six times larger than those to the previous migration status, namely, to those who had ever migrated but returned to their home villages. Moreover, while the economic returns to current migration status are statistically significant, the premium obtained by returned migrants is statistically insignificant. In a word, there is a short-term gain of objective well-being associated with migration decisions. Although a consensus acknowledges that migration improves income (Zhu 2002; Cai 2003), our analysis reminds us that this advantage of objective well-being does not persist after migrants return to their home villages.

Table 2 OLS regressions predicting log income and subjective well-being
Full size table

Not only does the income premium of migration not persist, our analysis suggests that this short-term economic return comes at the expense of SWB. The pertinent evidence is shown in Models 2–4 in Table 2. Specifically, we attempt to ask, for rural migrants, whether the increased income translates to a better life experience? Models 2 to 4 provide suggestive evidence based upon the sample from CLDS 2012.

Models 2, 3, and 4 of Table 2 report the regression results predicting self-assessed happiness, perceived social status, and a self-reported sense of fairness, respectively. The model specifications across three columns are similar, comparing both current migrants and returned migrants with their rural non-migrant counterparts, considering gender, age, Party membership, education, and income as control variables. The results show that current migrants have significantly lower self-assessed happiness, self-identified social status, and sense of fairness than rural non-migrants, holding individual-level characteristics constant. In particular, current migrants are generally more likely to report that they are living an unhappy life and consider their quality of living to be not fair relative to their labors. They also tend to place themselves in a lower rung of the social ladder than their rural non-migrant counterparts. These estimates are robust to the inclusion of a series of individual-level characteristics as controls. Regarding the effect of income, it seems like economic gains may counteract some of the negative effects of migration on self-assessed happiness and self-recognized social status, but there is no evidence of an offsetting effect on perceived fairness.

Our unreported auxiliary results suggest an intrinsic configuration of SWB—the three proxies we employ are highly interdependent. While higher self-recognized social status leads to a better perception of fairness, the perception of fairness may then be translated into greater self-assessed happiness. A suggestive finding is that rural migrant workers’ negative evaluation of their lives is substantially due to their sense of deprivation—they recognize themselves as the social underclass, and they perceive the social stratification process as unfair. The results of a factor analysis show that ratings of the three statements are positively correlated, as we expected; however, any single statement cannot absorb the information in the other two. The three statements represent three different dimensions of how respondents evaluate their life.

Moreover, unlike the short-term positive effect of migration on economic well-being, we find that the migration experience has a lingering negative effect on SWB, shown by the statistically significant negative coefficients of returned migrants across models. Specifically, returned migrants hold significantly lower self-assessed happiness and fairness than rural non-migrants. Therefore, our analysis of the effect of migration status on social well-being altogether suggests a rather pessimistic picture: rural migrant workers are likely to sacrifice their long-term SWB to trade for a short-term economic return. We further cross-check the robustness of the above-discussed results. Based upon the sample from CLDS 2014, the empirical results are qualitatively identical to what we find in Table 2.

In sum, rural-to-urban migration results in a higher income but a lower SWB. This finding could shed some new light on the existing theories explaining the relationship between objective well-being and SWB. One seminal theory comes from Easterlin (2001), who contends that happiness is a function of income and aspirations, with the former playing a positive role and the latter a negative role. Knight and Gunatilaka (2010) further argue that migrants’ unhappiness may be due to an unrealistic expectation for the urban lifestyle and income promotions. Following this line of thinking, our empirical results suggest that when rural citizens migrate to urban cities, they might have adopted an urban frame of reference. The resulting sense of deprivation undermines their feeling of fairness, self-evaluation in class stratification, and, finally, happiness. Rural migrants are deprived of SWB, striving for a higher income at the cost of a prolonged low level of happiness.

4.3 Mechanisms: The Mediating Effects of Social Exclusion

What are the underlying mechanisms buttressing this tradeoff between objective and subjective well-being? We look to social exclusion for a possible answer. Given the hukou registration system and the temporary nature of Chinese rural-to-urban migration, we speculate that not only the current migrants are socially excluded from the urban communities, but also the returned migrants are inadequately integrated into their home villages. Therefore, in Table 3, we pay extra attention to the mediating effects of social support networks, vulnerability to violence, and sense of community on the SWB of both current and returned migrants.

Table 3 OLS regressions of subjective well-being on migration status and the experienced social exclusion
Full size table

We begin our analysis by examining the role of social support networks in Models 1–3 of Table 3, using the CLDS 2012 sample. It is shown that social support networks have a statistically significant positive effect on every dimension of SWB. Comparing the coefficients of migration status in Table 3 with those from Table 2, we find that the magnitudes of these effects are slightly reduced, suggesting that the disadvantages of current and returned migrants on SWB are moderately explained by their lack of social support networks. The underlying policy implication seems to be that through promoting the social support network for rural migrants, we can slightly improve their SWB status. Interestingly, this mechanism functions for both current migrants and returned migrants, implying the disintegration of migrants in both urban and rural communities.

Likewise, Models 4, 5, and 6 suggest that migrants’ vulnerability to victimization plays a mediating role similar to a social support network. Again, comparing the coefficients of migration status in Table 3 to those from Table 2, we find that the effects of migration status on SWB are also moderately mediated by the vulnerability to violent crime. Those who are more vulnerable to violent crime and have had more experiences of victimization tend to have a lower SWB, where this vulnerability is also an important source of migrants’ disadvantaged SWB status.

The last three columns of Table 3 centers on the mediating role of the sense of community in the association between migration status and SWB. Comparing the results in Table 3 to those from Table 2, we find that the negative effects of the migration experience on SWB decrease substantively after controlling for the sense of community, and for current migrants, such is particularly the case. The disadvantaged SWB of current and returned migrants are strongly explained by their lack of a sense of community. For both current migrants and returned migrants, those who have a weaker sense of community may feel more excluded and, thus, tend to have a lower SWB in all three aspects. Our results imply that to enhance rural migrants’ SWB, policymakers should help them overcome the institutional barriers of the hukou registration system to better integrate into the local communities. The results in Table 3 are robust if we employ an alternative sample from CLDS 2014.

We also conduct a formal test of the mediating role of each measure of social exclusiveness, and report indirect effect, direct effect, and the share of the indirect effect, respectively in Table 3. The results suggest a more important mediating role of sense of community than the other two. Integration and reintegration to the residential communities are therefore of vital importance to the SWB of both current and returned migrants.

4.4 Auxiliary Results: Migration Status and Social Exclusion

We further provide auxiliary results on how migration status affects social exclusion in Table 4. Results based upon CLDS 2012 suggest that, compared with rural non-migrants, both current migrants and returned migrants are significantly more vulnerable to violent crime, feel less belonging to the community in which they reside, and have a narrower social support network. It is worth mentioning, though, that the group difference on social support networks between returned migrants and non-migrants is statistically insignificant based upon CLDS 2012 sample.Footnote 3

Table 4 OLS and logistic regressions predicting the experienced social exclusion
Full size table

Taken together, the results we show in Tables 2, 3, and 4 may contribute to a more comprehensive understanding of the social exclusion toward rural migrant workers and the lingering SWB costs. Even though current migrants are socially excluded from their urban communities, we surprisingly find that migrants are also poorly integrated into their home villages, even after their return to the countryside. Our results suggestively show that returned migrants face difficulties in restoring their social support networks, and, therefore, they feel a lack of belonging to their rural community of origin. Moreover, they are more likely to be the targets of violent crime than non-migrants, living in their own hometowns. All the above results remind us not to overlook the lingering social exclusion that migrants face after returned to their home communities.

5 Concluding Remarks

Using data from the China Labor Dynamics Survey, this study examines the role of social exclusion on the subjective well-being of both current migrants and returned migrants. Our analysis yields four empirical findings: (a) the economic gains of rural-to-urban migration come at the expense of subjective well-being, and, moreover, this psychological cost persists even after migrants return to their villages of origin; (b) the negative evaluation of rural migrants workers on their lives is highly correlated with their generally low self-identified social class and their sense of unfairness in socioeconomic status attainment; (c) social exclusion plays an important mediating role between migration status and subjective well-being. In particular, the lack of social support networks, the feeling of vulnerability, and a deficiency in the sense of community taken together contribute to rural migrant workers’ disadvantaged subjective well-being; (d) the difficulties of getting integrated not only hamper the subjective well-being of migrant workers in urban cities but also of peasants who have returned to the countryside. In general, our results imply a pessimistic picture of rural migrant workers’ well-being: to achieve a short-term economic return, they have to suffer from a lingering deprivation of social integration and happiness.

Our empirical findings bear important policy implications. Given that migrant workers suffer from social exclusion both in cities and in their home villages, decision-makers should pay more attention to the policy tools that enhance social integration. In particular, we suggest that policymakers invest their greatest efforts to improving the sense of belonging and the social support networks of migrants, not only in destination cities but also in villages of origin (Kassiola 2017).