Social, Cultural, and Civic Reintegration of Returning Rural Migrants in China: A Multidimensional Perspective

Abstract

Understanding the reintegration of returning rural migrants in China is crucial due to the large scale of return migration and its associated challenges. While existing research has largely focused on economic reintegration, this study broadens the scope to include social, cultural, and civic dimensions. Using data from the China Labor-force Dynamics Survey (CLDS) 2016 and employing multilevel ordered logistic regression, the research uncovers the following key patterns: (i) Determinants differ largely across dimensions; (ii) The roles of the same determinants can also differ significantly across dimensions; and (iii) There are significant community-level variations across dimensions. The findings emphasize that success in one dimension, such as economic reintegration, does not necessarily translate into success in others. Moreover, complex interconnections between dimensions reveal positive, negative, and non-linear relationships, underscoring the multidimensional nature of reintegration. These insights highlight the importance of considering multiple dimensions to fully understand the reintegration processes of returning migrants.

1. Introduction

1.1. Return Migration

The scale of global migration has reached unprecedented levels, with 281 million international migrants reported in 2020 [1]. This number was surpassed by internal migration within China alone, with 376 million internal migrants by late 2020 [2]. This massive movement of people often involves temporary migration, leading to significant patterns of return migration [3,4,5].

Estimates suggest that 20% to 50% of international migrants return to their country of origin within five years [6]. Similarly, China has experienced substantial return migration of internal migrants since the 1990s [7,8]. While migrants with higher education levels are more likely to settle permanently in urban areas, the majority of rural migrants, often with lower educational attainment, tend to return home after accumulating savings, skills, and work experience [8,9,10]. However, return migration is not merely a reversal of the migration process. Scholars increasingly emphasize that returning is a complex and often contested process, shaped by emotional expectations, power relations, and structural barriers [11,12,13]. It can involve disillusionment, identity renegotiation, and a sense of social or economic failure. This is particularly evident in contexts where reintegration is constrained by policy regimes or local dynamics. In the Chinese context, the hukou system not only makes long-term settlement in cities more difficult and encourages return but also reinforces returnees’ marginal status [14]. The hukou system is a household registration system unique to China, determining an individual’s official place of residence and associated entitlements, including access to social welfare, education, healthcare, and employment opportunities. By restricting migrants’ access to services and rights in urban areas, the hukou system shapes both the migration and return migration decisions. Specifically, it can transform what would elsewhere be voluntary return migration into a forced or constrained decision due to institutional and structural barriers. As such, return does not guarantee successful reintegration.

1.2. Reintegration

Reintegration, a vital process for return migrants after they return home, is often challenging. Returnees face economic hurdles, including difficulties in finding employment, achieving financial stability, and utilizing acquired skills [15,16,17,18]. They also encounter social challenges, such as rebuilding networks and facing exclusion and discrimination, which can affect their health and mental well-being [15,19,20]. Cultural difficulties also arise as migrants adapt to local customs and navigate cross-cultural conflicts [16,21]. Additionally, returnees may struggle with civic engagement and accessing social services and rights [22,23,24].

Importantly, reintegration is not a self-evident or linear process. As scholarship emphasizes, it is often fragmented, contested, and incomplete. Rather than simply “going home”, return migrants must renegotiate identity, belonging, and opportunity within shifting personal and structural constraints [11,12,13]. Emotional dislocation, social alienation, and institutional barriers may persist long after return, shaping reintegration in unpredictable and unequal ways [11,12,13]. Successful reintegration, though challenging, is meaningful for both returning migrants and their home societies [25]. For migrants, it safeguards rights, preserves dignity, ensures livelihood security, fosters community belonging, and prevents remigration [23,26,27]. It also promotes well-being, societal engagement, and utilization of acquired knowledge, skills, and capital [28,29]. In addition, successful reintegration facilitates access to social services, employment, and economic opportunities [18,22]. Conversely, poor reintegration can lead to social exclusion and marginalization [15,22,30].

For home communities and national development, successful reintegration catalyzes local growth through productive investments and skill application [18,29]. It can also revitalize local labor markets and contribute to national development [31,32]. Particularly, highly skilled returnees stimulate economic growth by transferring skills and resources, establishing businesses, creating jobs, and supporting economic activities [16,21]. This process transforms “brain drain” into “brain gain” [21].

For rural returnees in China, significant challenges persist despite potentially fewer reintegration issues. A substantial rural–urban gap in well-being aspects like education, occupation, earnings, and life perceptions creates diverse challenges upon return [33]. These challenges are not only individual but also structural, as the hukou system continues to institutionalize rural–urban disparities in access to education, employment opportunities, and social services. Economically, migrants’ acquired skills often do not align with farming. Many then choose to pursue self-employment [34,35,36]. Socially, while migrants gain human and financial capital, they may lose social capital at home [37]. Culturally, migration can alter values; thus, returnees will need to adapt to rural culture [38]. Civically, rural–urban differences in governance views [39,40] and migrants’ expanded place attachment [41] may result in distinct civic engagement behaviors among returnees compared to rural residents.

The importance of reintegration in China parallels international contexts, particularly because the hukou system uniquely restricts internal migrants’ access to services, social welfare, and economic opportunities in urban areas, resembling structural barriers and policy restrictions that international migrants often face abroad. Similar to international return migration, successful reintegration in China not only benefits returnees individually but also positively affects their home communities through entrepreneurial initiatives, investments, and the transfer of skills and resources acquired elsewhere [7,42,43].

1.3. Research Question

Despite comprehensive research on return migrants’ reintegration, a significant gap exists in the context of internal migration in China. Existing studies primarily focus on economic reintegration, particularly self-employment or entrepreneurship [7,35,36,43,44]. Chen [34] extends this to income reintegration (i.e., income gaps between return migrants and rural residents). However, non-economic dimensions of reintegration remain underexplored.

This research examines the determinants of social, cultural, and civic reintegration among returning rural migrants in China, as well as the interconnections among these dimensions. Figure 1 presents the study’s conceptual framework, which illustrates how major determinants identified in the literature, including demographic traits, socioeconomic status, migration experience and plans, and institutional context, jointly shape reintegration across the three non-economic dimensions. The framework also reflects the mutual interdependence of social, cultural, and civic reintegration, as indicated by the bidirectional arrows connecting these domains. These arrows highlight that reintegration is multidimensional and interactive, such that improvements in one domain can potentially influence outcomes in others. To my knowledge, this is the first empirical study in the Chinese context to explore non-economic reintegration dimensions and explicitly test their interrelationships. This multidimensional approach offers a more complete understanding of reintegration, which is especially important given the substantial and growing number of returnees in China [2,8].

This study also offers insights applicable to broader international contexts. Rural-to-urban migrants in China face similar institutional, economic, cultural, and social barriers as international migrants [45,46]. The hukou system creates circumstances comparable to those of international migrants [47,48]. Cultural distinctions, including language differences, further parallel international migration experiences [49]. Given these parallels, the findings of this study have broader relevance for understanding return migration and reintegration globally.

2. Literature Review

2.1. Exploration of Different Dimensions of Reintegration

Scholars have increasingly recognized reintegration as a multidimensional process [23,28,50,51]. This framework guides the present study’s examination of non-economic reintegration in China.

Economic reintegration has received the most attention in the existing literature. Studies have primarily focused on labor market outcomes, exploring the probability of returnees finding a job, becoming self-employed or entrepreneurs, and evaluating their working conditions [18,21,30,32,52,53,54]. Others examine financial dimensions including income, savings, asset ownership, and sustainable livelihoods [16,17,27,55,56]. While economic aspects are crucial, other non-economic dimensions of reintegration have received less attention. Social reintegration, which involves rebuilding social networks and relationships, has been explored by some researchers [16,19,21,24,27]. Cultural reintegration, focusing on adapting to cultural norms and practices, has been examined even to a lesser extent [24]. Civic reintegration, which encompasses participating in civic life and accessing public services, has also been studied, albeit less frequently [22,23,24]. This gap is particularly notable for China’s internal migrants.

2.2. Determinants of Reintegration

Studies on reintegration in various contexts suggest four main groups of determinants, including demographic traits, socioeconomic status, migration experience and plans, and institutional factors. Among demographic traits, age, gender, and marital status are shown to shape reintegration outcomes, with older and female migrants facing more challenges [19,32]. In China, these traits also influence outcomes through selection processes [34].

For socioeconomic status, both human capital, such as education level and skills, and financial capital, such as income and savings, significantly impact reintegration [16,19,21,27,28,50,52,53]. In general, human capital and financial capital are expected to facilitate reintegration. Employment status and opportunities also matter [16]. In the case of China, we have also observed that human capital, financial capital, and employment status can explain reintegration outcomes [7,34,36,43].

Migration experience and plans matter significantly. Migration experience includes migration duration and types of migration (i.e., legal vs. illegal, economic vs. political) [28,29,50,53,56]. Migration duration is largely related to skills gained abroad [29], while types of migration are largely related to whether it is a forced or voluntary return, which in turn largely influences reintegration [15,51,53,56]. Migration plans include whether they plan to remigrate [23,27,32]. In the context of China, the importance of migration experience and types of migration (i.e., temporary vs. permanent migration) has also been identified [34,36,43,44].

Finally, institutional factors such as home conditions, including employment opportunities and government policies, can significantly influence reintegration [16,17,18,21,22,26,28,30,31,32,51,55]. Similarly, in China, employment opportunities are crucial for return migrants’ reintegration [7].

Although not studied as an independent determinant in this paper, the hukou system’s impacts are inherently captured within the four groups of reintegration determinants examined—demographic traits, socioeconomic status, migration experiences and future plans, and institutional factors. These determinants already reflect the critical pathways through which the hukou system influences reintegration outcomes, making its explicit inclusion as a separate determinant redundant within the existing analytical framework.

3. Data and Methodology

3.1. Data

This research is based on the 2016 China Labor-force Dynamics Survey (CLDS) data. The CLDS dataset is nationally representative and notable for being one of the first rotation panel surveys in China. It provides a wealth of information across three key levels: individual, family, and community. The publicly available CLDS includes waves from 2012, 2014, and 2016, but only the 2016 survey contains the full set of variables required for this study; later waves (e.g., 2018) have been collected but are not yet publicly released. The CLDS dataset is particularly well-suited for this investigation because it offers a comprehensive view of return migrants within rural hometowns. To ensure data relevance, the analysis is limited to returning rural migrants—individuals with rural hukou residing in rural hometowns with prior cross-county labor migration experience. I also further restrict to those who have answered the questions on reintegration measures.

This last sample selection procedure may raise a concern about external validity. Thus, I compare the distribution of characteristics of all return migrants and those return migrants with reintegration experiences observed to understand the extent that the selected samples can be representative of the entire return migrant samples in the data. The comparison is shown in

Table A1. Sample selection representativeness.

Variables	All Return Migrant Samples	Return Migrant Sample Used	p-Value from Statistical Tests (t-Test, Proportion Test, or Chi-Squared Test)
Demographic Traits
Female	38.40%	37.40%	0.43
Age			1
<25	9.20%	9.90%
26–35	22.90%	23.90%
36–45	23.10%	24.40%
46–55	28.00%	26.90%
55+	16.70%	15.00%
Married	86.60%	86.20%	0.67
Socioeconomic Status
Education			1.00
No Education	5.20%	5.40%
Primary School	30.60%	30.10%
Middle School	49.00%	50.30%
High School	15.20%	14.20%
Employment Status			0.87
Employee	31.80%	29.70%
Employer	1.20%	1.20%
Self-Employed	11.80%	11.50%
Farmer	40.60%	44.30%
Unemployed	14.60%	13.30%
Income Level			1.00
Unemployed	14.60%	13.30%
1st Quartile	21.50%	23.10%
2nd Quartile	29.80%	30.80%
3rd Quartile	13.70%	13.50%
4th Quartile	20.50%	19.20%
Economic Status Satisfaction			0.98
Very Unhappy	9.40%	9.10%
Unhappy	27.30%	28.30%
Neither Happy nor Unhappy	34.40%	33.80%
Happy	22.20%	21.70%
Very Happy	6.80%	7.10%
Migration Experience and Plan
Last Migration Duration	3.33	3.20	0.28
Return Duration	8.46	8.14	0.15
Plan to Settle at Home	85.10%	84.50%	0.50
Institutional Factors
Number of Available Services Supporting Farmers	2.70	2.74	0.39
Non-Agricultural Economy	25.70%	22.50%	0.00
N	2219	1624

in the Appendix A. The distributions are almost identical between all return migrant samples and the selected return migrant samples. Particularly, Table A1 presents variable-specific statistical comparisons—t-tests for continuous measures, one-sample proportion tests for dichotomous variables, and χ² goodness-of-fit tests for multi-category variables. Except for one variable, none of the comparisons between the full and selected return migrant samples reach statistical significance, confirming that the subsample is representative of the overall population.

3.2. Dependent Variables

Social reintegration is defined as a return migrant’s (re)establishment of interpersonal support networks and broader social adaptation within their home village. This is assessed using three variables. The first two variables are about the establishment of social networks, including the number of friends/acquaintances who can provide support and help and the quality of relationships with neighbors. The number of friends/acquaintances is assessed using an ordinal measure including 0, 1–10, 11–20, and 20+, while relationships with neighbors are assessed on a scale from “Very Unfamiliar” to “Very Familiar”. The establishment of social networks is the main measure of social reintegration [19,27]. The third variable is a self-report Likert scale, where participants are asked to indicate the extent to which they agree it is getting harder to adapt to rural life. Responses are recorded on a scale ranging from “Strongly Agree” to “Strongly Disagree”. “Strongly Agree” will suggest experiencing a hard social adaptation process, while “Strongly Disagree” will suggest experiencing a smooth social adaptation process, capturing the social-psychological or social-cultural aspect of social reintegration [23,28].

Cultural reintegration is defined as adherence to local norms and values. It is first evaluated by examining participants’ likelihood of giving gifts during happy events in their village, a practice deeply rooted in the cultural traditions in China, reflecting the culture of renqing (human obligations), guanxi (relationship), yuan (destiny or fate), reciprocity, family orientation and Mianzi (face) [57]. This assessment focuses on their willingness to offer gifts when relatives and friends in the village organize joyful occasions, as well as when others in the village celebrate such events, capturing the adherence to the traditions of the community [24]. Responses are categorized from “Always” to “Hardly”. I recode these responses so that they go from “Hardly” to “Always”, with a higher level indicating higher cultural reintegration. Because these two measures not only capture cultural reintegration but also partly capture social reintegration, I also include a third measure for cultural reintegration—son preference. The question asked whether having a son is better than having a daughter, and the responses range from strongly agree to strongly disagree. I recode these responses so that they go from “Strongly Disagree” to “Strongly Agree”, with a higher level indicating higher cultural reintegration. It is known that son preference tends to be stronger for people in rural areas than in urban areas [58].

Civic reintegration is defined as willingness to engage in community and political life. It is measured by assessing participants’ willingness to participate in community activities, including both local and political activities [24]. Participants are asked if they would be willing to contribute labor or funds if the village organized road construction and if they would be willing to participate in elections of the village cadre. Responses are categorized from “Very Willing” to “Very Unwilling”. I recode these responses so that they go from “Very Unwilling” to “Very Willing”, with a higher level indicating higher civic reintegration.

3.3. Independent Variables

Independent variables consist of four groups of determinants of reintegration identified and discussed previously, including demographic traits, socioeconomic status, migration experience and plans, and institutional factors. Participants’ demographic traits include age, gender, and marital status. Age is categorized into five groups (<25, 26–35, 36–45, 46–55, 55+). Gender is measured as whether the respondents are females, and marital status captures whether the respondents are married.

Socioeconomic status is assessed through education level, employment status, income level, and satisfaction with economic status. Education level is categorized into four groups (no education, primary education, middle school education, high school education or more). I do not include college education as a separate category given that the number of return migrants with a college education or more is too small. Employment status includes categories for employee, employer, self-employed, farming, and unemployed. Income level is measured by income level, categorized into five groups, including unemployed, first quartile, second quartile, third quartile, and fourth quartile. Satisfaction with economic status is from “Very Unhappy” to “Very Happy” and “Very Satisfied” to “Very Unsatisfied”, respectively. The latter one is reversely coded so that it ranges from “Very Unsatisfied” to “Very Satisfied”. Controlling for these socioeconomic indicators is essential because they proxy economic reintegration, enabling a clearer analysis of the determinants and interrelations of social, cultural, and civic dimensions.

Migration experience and plans include the duration of the last migration, the duration of stay at home and its squared, and future settlement plans. The duration of the last migration and the duration of staying at home are continuous measures (in years). The squared term of duration at home is also included, as reintegration may not follow a linear trajectory over time. Settlement plans are measured based on participants’ intentions to settle at home.

Institutional factors in the participants’ home communities include the number of available services supporting agriculture and benefitting farmers. These services include irrigation and drainage, farming services, pest prevention and control, means of production purchase, planting planning, labor migration organization, and agricultural production technology training. Another institutional factor is the presence of non-agricultural economic activities in the community.

3.4. Methods

The analysis employs multilevel ordered logistic regression due to the hierarchical structure of the data, with return migrants nested within communities. This model incorporates a random intercept for the community to account for the potential variability between communities in the reintegration outcomes of return migrants. This approach offers a robust framework to examine the impact of various factors on reintegration while acknowledging the shared community context.

The significance of the variance captured by this model is crucial in the analysis. If the variance between communities is found to be statistically significant, it implies that there are significant differences in reintegration outcomes among different communities. This, in turn, provides valuable insights into the role of the community context in shaping the reintegration experiences of return migrants.

When exploring the determinants of social, cultural, and civic reintegration, these reintegration outcomes are used as the dependent variables, and the four groups of determinants (i.e., demographic traits, socioeconomic status, migration experience and plans, and institutional factors) are used as the independent variables. To assess multicollinearity, I compute variance inflation factors for all covariates (

Table A2. Multicollinearity test of controls.

Variables	VIF
Demographic Traits
Female	1.29
Age (Reference group = ≤ 25)
26–35	3.07
36–45	3.69
46–55	4.05
55+	3.44
Married	1.3
Socioeconomic Status
Education (Reference group = No Education)
Primary School	4.85
Middle School	5.74
High School	3.56
Employment Status (Reference group = Employee)
Employer	1.08
Self-Employed	1.28
Farmer	1.84
Unemployed	1.95
Income Level (Reference group = 1st Quartile)
2nd Quartile	1.81
3rd Quartile	1.64
4th Quartile	1.95
Economic Status Satisfaction (Reference group = Very Unhappy)
Unhappy	2.99
Neither Happy nor Unhappy	3.19
Happy	2.74
Very Happy	1.71
Migration Experience and Plan
Last Migration Duration	1.07
Return Duration	7.44
Return Duration Squared	7.3
Plan to Settle at Home	1.13
Institutional Factors
Number of Available Services Supporting Farmers	1.04
Non-Agricultural Economy	1.08

). The majority of predictors had Variance Inflation Factors (VIFs) below 2, and the highest VIF was 7, indicating no serious collinearity. Then, when exploring how these dimensions of reintegration are interconnected, I add the other two dimensions as the independent variables separately as two additional regression models. For example, for social reintegration, I first explore whether the four groups of determinants explain it and then add two additional models, one with cultural reintegration included and another with civic reintegration included. It is crucial to conduct this analysis because these relationships are not strictly causal; rather, they are interrelated. While one could introduce interaction or moderating terms to capture conditional interdependencies, doing so with multiple measures per dimension would generate a significant amount of coefficients and hinder interpretation.

4. Findings

4.1. Determinants of Social Reintegration

Table 1. Determinants of social reintegration (multilevel ordered logistic regression, odds ratio).

Variables	Social Reintegration
	Social: Number of Friends	Social: Relationship with Neighbors	Social: Feeling About Adaptation
Demographic Traits
Female	0.648 ***	0.551 ***	0.870
	(0.0814)	(0.0639)	(0.0983)
Age (Reference group = ≤25)
26–35	0.905	1.199	0.984
	(0.197)	(0.238)	(0.195)
36–45	0.856	1.700 *	1.062
	(0.206)	(0.370)	(0.231)
46–55	0.824	1.944 **	1.065
	(0.202)	(0.432)	(0.236)
55+	0.692	1.560 +	1.132
	(0.196)	(0.398)	(0.291)
Married	1.436 *	1.295	1.113
	(0.250)	(0.207)	(0.176)
Socioeconomic Status
Education (Reference group = No Education)
Primary School	0.625 +	0.684	0.706
	(0.163)	(0.161)	(0.164)
Middle School	0.908	0.719	0.655 +
	(0.238)	(0.172)	(0.154)
High School	0.988	0.571 *	0.614 +
	(0.292)	(0.153)	(0.163)
Employment Status (Reference group = Employee)
Employer	4.247 **	3.609 *	0.835
	(2.128)	(1.836)	(0.397)
Self-Employed	1.149	0.950	0.617 **
	(0.215)	(0.162)	(0.104)
Farmer	1.421 *	1.176	0.964
	(0.212)	(0.160)	(0.131)
Unemployed	0.905	1.203	1.034
	(0.200)	(0.245)	(0.209)
Income Level (Reference group = 1st Quartile)
2nd Quartile	1.076	1.163	0.893
	(0.167)	(0.165)	(0.126)
3rd Quartile	1.004	1.156	1.004
	(0.198)	(0.209)	(0.182)
4th Quartile	1.224	1.158	1.032
	(0.232)	(0.204)	(0.179)
Economic Status Satisfaction (Reference group = Neither Happynor Unhappy)
Very Unhappy	0.898	1.074	1.283
	(0.183)	(0.200)	(0.241)
Unhappy	0.772 +	1.054	0.789 +
	(0.106)	(0.131)	(0.0976)
Happy	1.482 **	1.187	1.066
	(0.218)	(0.160)	(0.143)
Very Happy	1.345	2.997 ***	1.724 **
	(0.299)	(0.666)	(0.364)
Migration Experience and Plan
Last Migration Duration	0.988	0.982 +	0.996
	(0.0122)	(0.0108)	(0.0108)
Return Duration	0.996	0.996	1.045 **
	(0.0171)	(0.0148)	(0.0156)
Return Duration Squared	1.000	1.000	0.999
	(0.000553)	(0.000457)	(0.000462)
Plan to Settle at Home	0.754 +	0.995	1.371 *
	(0.117)	(0.143)	(0.195)
Institutional Factors
Number of Available Services Supporting Farmers	1.018	1.082 *	0.954
	(0.0400)	(0.0398)	(0.0368)
Non-Agricultural Economy	0.913	0.686 *	1.122
	(0.165)	(0.115)	(0.196)
Cut 1	−2.009 ***	−4.769 ***	−4.034 ***
	(0.401)	(0.465)	(0.399)
Cut 2	1.443 ***	−2.894 ***	−2.238 ***
	(0.400)	(0.383)	(0.368)
Cut 3	2.612 ***	−0.837 *	−1.077 **
	(0.407)	(0.369)	(0.363)
Cut 4	-	1.270 ***	1.105 **
	-	(0.370)	(0.363)
Random Effect Variances
Community	0.392 ***	0.343 ***	0.419 ***
	(0.112)	(0.0886)	(0.106)
Observations	1624	1624	1624
Number of Groups	186	186	186

SE in parentheses. *** p < 0.001, ** p < 0.01, * p < 0.05, ⁺ p < 0.1. The significance of the variance of random effect is based on the Likelihood-ratio (LR) test.

shows the factors that influence social reintegration. The results show that demographic traits significantly affect social reintegration. First, females have a harder time establishing their social networks than males, reflected by having fewer friends/acquaintances who can provide them support and help (OR = 0.648, p < 0.001) and being less familiar with neighbors (OR = 0.551, p < 0.001). This corresponds to roughly 35% to 45% lower odds of network ties. This is consistent with the findings in the international context that females experience a harder reintegration process [19]. However, those who are married tend to have more friends/acquaintances (OR = 1.436, p < 0.05). This represents about a 44% increase in odds. Those in middle age (i.e., 36–55) are more familiar with neighbors than those in young adulthood (i.e., less than 25). The odds ratios are between 1.700 to 1.944, corresponding to 70–94% higher odds of familiarity with neighbors. This may reflect the role of marital status in leading to social networks of higher density and kin closeness [59]. This also suggests that the age effect may not be linear. Those who are in their middle age may find it easier to reintegrate than those in their young adulthood. However, the old ones who are expected to face additional challenges in reintegration do not differ from the young ones [32]. Yet, none of the demographic traits influenced the subjective social reintegration.

In addition, socioeconomic status exhibits a complex relationship with social reintegration. Return migrants with high school or above education are less familiar with their neighbors (OR = 0.571, p < 0.05), representing a 43% reduction in odds. This likely suggests that those with higher education are less likely to interact with their neighbors. The gap in education between rural and urban and the rural-to-urban migration experience may also lead to the gap in perceptions and values, which, in turn, makes it harder for social interactions between highly educated return migrants and rural residents [38,60]. Employment status plays a major role. Compared to employees, employers have over three times the odds of larger networks (OR = 4.247, p < 0.01) and neighbor familiarity (OR = 3.609, p < 0.05), and farmers also enjoy higher network counts (OR = 1.421, p < 0.05). Interestingly, self-employed individuals report lower subjective adaptation (OR = 0.617, p < 0.01). Studies have suggested that self-employment is a major path to economic reintegration [36,44]. Yet, this finding suggests that self-employment could lead to a lower level of subjective social reintegration, suggesting a potential trade-off between the two. These findings reflect how employment status can facilitate or hinder social reintegration. Furthermore, respondents who are satisfied with their economic status show substantially higher social reintegration. Specifically, those reporting “Very Happy” show nearly a 200% higher odds in neighbor familiarity and 70% higher odds in subjective adaptation.

Migration experience and plans demonstrate significant impacts on social reintegration. Over time, return migrants become more socially reintegrated, with each additional year at home increasing the odds of subjective adaptation by about 4% (OR = 1.045, p < 0.01). Additionally, those who plan to settle at home display higher levels of social reintegration (OR = 1.371, p < 0.05). This suggests that embedded characteristics such as higher abilities and aspirations influence social reintegration [35,61]. Institutional factors influence familiarity with neighbors. Return migrants residing in communities with more agricultural service provision are more familiar with neighbors (OR = 1.082, p < 0.05), while return migrants residing in communities with non-agricultural economies are less familiar with neighbors (OR = 0.686, p < 0.05). This suggests that social reintegration tends to be easier in more rural, agriculture-focused settings where community bonds and daily interactions around farming activities are stronger. In addition, a significant random intercept for all social reintegration measures suggests substantial community-level variations in social reintegration outcomes. Both institutional factors and community-level variances suggest the importance of the hometown conditions in influencing social reintegration [16,17,18,21,22,26,28,30,31,32,55].

4.2. Determinants of Cultural Reintegration

Results on cultural reintegration are shown in

Table 2. Determinants of cultural reintegration (multilevel ordered logistic regression, odds ratio).

Variables	Cultural Reintegration
	Cultural: Gifts to Friends	Cultural: Gifts to Others	Cultural: Son Preference
Demographic Traits
Female	0.771 *	0.878	0.717 **
	(0.0933)	(0.102)	(0.0797)
Age (Reference group = ≤25)
26–35	1.542 *	1.155	1.364
	(0.324)	(0.234)	(0.269)
36–45	1.966 **	1.575 *	1.073
	(0.456)	(0.350)	(0.232)
46–55	1.587 +	1.438	1.339
	(0.374)	(0.326)	(0.295)
55+	1.571 +	1.641 +	1.434
	(0.426)	(0.432)	(0.363)
Married	2.019 ***	1.186	0.718 *
	(0.345)	(0.193)	(0.112)
Socioeconomic Status
Education (Reference group = No Education)
Primary School	0.834	0.675 +	0.809
	(0.205)	(0.156)	(0.188)
Middle School	0.897	0.693	0.539 **
	(0.224)	(0.164)	(0.127)
High School	0.930	0.674	0.418 ***
	(0.261)	(0.180)	(0.111)
Employment Status (Reference group = Employee)
Employer	1.125	2.474 +	0.569
	(0.572)	(1.288)	(0.260)
Self-Employed	0.701 +	0.778	1.207
	(0.128)	(0.139)	(0.207)
Farmer	1.100	1.186	1.219
	(0.160)	(0.166)	(0.164)
Unemployed	1.170	1.043	1.055
	(0.251)	(0.216)	(0.209)
Income Level (Reference group = 1st Quartile)
2nd Quartile	1.234	1.177	0.745 *
	(0.185)	(0.170)	(0.104)
3rd Quartile	1.090	1.189	0.874
	(0.209)	(0.218)	(0.155)
4th Quartile	1.218	1.047	1.054
	(0.226)	(0.187)	(0.179)
Economic Status Satisfaction (Reference group = Neither Happy nor Unhappy)
Very Unhappy	0.875	0.777	0.693 +
	(0.168)	(0.148)	(0.131)
Unhappy	1.280 +	0.973	0.982
	(0.171)	(0.122)	(0.120)
Happy	1.005	1.032	0.964
	(0.143)	(0.142)	(0.128)
Very Happy	1.196	1.450 +	0.697 +
	(0.266)	(0.306)	(0.149)
Migration Experience and Plan
Last Migration Duration	1.006	1.010	0.989
	(0.0119)	(0.0114)	(0.0108)
Return Duration	0.998	1.020	1.015
	(0.0159)	(0.0154)	(0.0153)
Return Duration Squared	1.000	0.999	0.999 *
	(0.000491)	(0.000464)	(0.000474)
Plan to Settle at Home	1.044	1.058	0.888
	(0.158)	(0.155)	(0.127)
Institutional Factors
Number of Available Services Supporting Farmers	1.051	0.986	0.950
	(0.0470)	(0.0541)	(0.0338)
Non-Agricultural Economy	1.183	0.731	0.907
	(0.241)	(0.181)	(0.147)
Cut 1	−2.006 ***	−0.962 *	−2.182 ***
	(0.399)	(0.390)	(0.359)
Cut 2	−0.934 *	0.141	−0.0912
	(0.391)	(0.389)	(0.354)
Cut 3	1.461 ***	1.875 ***	1.043 **
	(0.392)	(0.392)	(0.357)
Cut 4	-	-	2.406 ***
	-	-	(0.373)
Random Effect Variances
Community	0.660 ***	1.285 ***	0.305 ***
	(0.145)	(0.224)	(0.0837)
Observations	1624	1624	1624
Number of Groups	186	186	186

SE in parentheses. *** p < 0.001, ** p < 0.01, * p < 0.05, ⁺ p < 0.1. The significance of the variance of random effect is based on the likelihood-ratio (LR) test.

. Demographic traits, gender, age, and marital status all matter for cultural reintegration. Return female migrants are less likely to give gifts during the happy events of relatives and friends (OR = 0.771, p < 0.05), indicating approximately 23% lower odds compared to males. They are also less likely to express son preference (OR = 0.717, p < 0.01), suggesting a significantly weaker adherence to this traditional norm. This again reflects that it is harder for females to reintegrate, as observed in the international context [19]. Return migrants aged between 26 and 35 are more likely to give gifts to friends and relatives (OR = 1.542, p < 0.05), and those aged 36–45 are even more likely to do so (OR = 1.966, p < 0.01). The latter group is also more likely to give gifts to others in the village (OR = 1.575, p < 0.05). These findings suggest that middle-aged returnees have 54% to nearly 97% higher odds of engaging in cultural practices compared to the youngest group. Again, this reflects the non-linear effect of age, with young and old ones facing similar reintegration challenges [32]. Those who are married are more likely to give gifts to friends and relatives (OR = 2.019, p < 0.001), doubling the odds compared to unmarried returnees. However, they are less likely to express son preference (OR = 0.718, p < 0.05), indicating a departure from traditional gender norms among the married. This may have to do with the higher exposure to this cultural tradition. For example, married people are expected to receive gifts from relatives and friends during their wedding. Thus, they are also expected to give gifts due to the reciprocity aspect of gift-giving behavior [57,62,63].

Socioeconomic status shows that those with middle school (OR = 0.539, p < 0.01) and high school education (OR = 0.418, p < 0.001) are less likely to express son preference, indicating 46% to 58% lower odds of holding traditional gender bias than those with no education. In addition, those in the middle income level show lower odds of holding gender bias compared to those in the lowest income group (OR = 0.745, p < 0.05).

Return duration matters, but its effect on son preference is non-linear. Specifically, the influence of return duration declines over time, as indicated by the odds ratio for the squared term (OR = 0.999, p < 0.05). Although institutional factors do not matter, the presence of a random intercept indicates notable community-level variations in cultural reintegration outcomes, again suggesting the importance of hometown conditions [16,17,18,21,22,26,28,30,31,32,55].

4.3. Determinants of Civic Reintegration

Table 3. Determinants of civic reintegration (multilevel ordered logistic regression, odds ratio).

Variables	Civic Reintegration
	Civic: Contribute Labor or Funds	Civic: Participate in Elections
Demographic Traits
Female	0.806 +	0.962
	(0.101)	(0.110)
Age (Reference group = ≤25)
26–35	0.965	1.011
	(0.209)	(0.200)
36–45	1.249	1.583 *
	(0.300)	(0.345)
46–55	1.285	2.005 **
	(0.314)	(0.446)
55+	1.214	2.184 **
	(0.342)	(0.560)
Married	1.625 **	1.278
	(0.281)	(0.202)
Socioeconomic Status
Education (Reference group = No Education)
Primary School	1.433	1.267
	(0.369)	(0.299)
Middle School	1.424	1.381
	(0.373)	(0.331)
High School	1.556	1.533
	(0.457)	(0.413)
Employment Status (Reference group = Employee)
Employer	1.823	1.091
	(0.939)	(0.527)
Self-Employed	1.329	1.275
	(0.246)	(0.221)
Farmer	1.662 ***	1.567 ***
	(0.249)	(0.213)
Unemployed	1.282	1.400 +
	(0.283)	(0.282)
Income Level (Reference group = 1st Quartile)
2nd Quartile	1.296 +	1.281 +
	(0.202)	(0.183)
3rd Quartile	1.037	1.270
	(0.204)	(0.232)
4th Quartile	0.938	0.934
	(0.177)	(0.163)
Economic Status Satisfaction (Reference group = Neither Happy nor Unhappy)
Very Unhappy	2.224 ***	1.540 *
	(0.467)	(0.299)
Unhappy	1.012	1.082
	(0.136)	(0.135)
Happy	1.216	1.042
	(0.178)	(0.139)
Very Happy	2.943 ***	2.341 ***
	(0.741)	(0.537)
Migration Experience and Plan
Last Migration Duration	1.001	0.999
	(0.0122)	(0.0114)
Return Duration	1.015	1.004
	(0.0166)	(0.0149)
Return Duration Squared	1.000	1.000
	(0.000499)	(0.000447)
Plan to Settle at Home	0.851	0.655 **
	(0.132)	(0.0953)
Institutional Factors
Number of Available Services Supporting Farmers	1.083	1.089 *
	(0.0531)	(0.0406)
Non-Agricultural Economy	0.612 *	0.698 *
	(0.135)	(0.117)
Cut 1	−4.983 ***	−3.286 ***
	(0.640)	(0.412)
Cut 2	−2.393 ***	−1.273 ***
	(0.425)	(0.366)
Cut 3	−0.758 +	−0.280
	(0.409)	(0.363)
Cut 4	1.603 ***	1.619 ***
	(0.412)	(0.366)
Random Effect Variances
Community	0.858 ***	0.354 ***
	(0.168)	(0.0952)
Observations	1624	1624
Number of Groups	186	186

SE in parentheses. *** p < 0.001, ** p < 0.01, * p < 0.05, ⁺ p < 0.1. The significance of the variance of random effect is based on the likelihood-ratio (LR) test.

shows factors that affect civic reintegration. Demographic traits unveil distinct patterns. Return migrants aged 36 or above are significantly more likely to participate in elections, with odds ratios of 1.583 for those aged 36–45, 2.005 for those aged 46–55, and 2.184 for those 55 or older. This suggests that older individuals are more civically active, possibly due to elder responsibilities. Married return migrants are more likely to contribute labor or funds (OR = 1.625, p < 0.01), reflecting a 63% increase in odds, likely driven by parenting obligations for children [64,65].

Socioeconomic status shows significant effects. Farmers are significantly more likely to both contribute labor or funds (OR = 1.662, p < 0.001) and participate in elections (OR = 1.567, p < 0.001), reflecting approximately 57–66% higher odds than employees. This essentially suggests that farmers are more likely to engage civically, possibly because they are more attached to rural communities [66]. Those who are most satisfied with their economic status are more inclined to engage in both civic activities. However, those who are least satisfied with their economic status are also more inclined to engage in contributing labor or funds, showing a non-linear pattern. Particularly, those who report being ‘Very Unhappy’ with their economic status have between 54% and 122% higher odds of participating in civic activities, while those who report being ‘Very Happy’ have between 134% and 194% higher odds of civic participation.

Migration experience and plans play important roles. Surprisingly, return migrants who plan to settle at home are less likely to participate in elections (OR = 0.655, p < 0.01), with about 34% lower odds. Institutional factors emerged as significant contributors. Each additional agricultural service is associated with higher odds of election participation (OR = 1.089, p < 0.05). However, residing in a non-agricultural economy is linked to lower civic engagement in both contributing labor or funds (OR = 0.612, p < 0.05) and participating in elections (OR = 0.698, p < 0.05), reflecting reductions in odds of about 30–39%. The random intercept highlights substantial community-level variations in civic reintegration outcomes. Both institutional factors and community-level variations underline the importance of institutional support observed in a wide range of international contexts [16,17,18,21,22,26,28,30,31,32,55].

4.4. Interconnections Among Social, Cultural, and Civic Reintegration

Beyond insights on the determinants of reintegration, the key insights that have emerged from the analysis are the interconnections among various dimensions of reintegration—social, cultural, and civic. The results are shown in

Table 4. Connection between social reintegration and cultural reintegration (multilevel ordered logistic regression, odds ratio).

Variables	Social Reintegration			Cultural Reintegration
	Social: Number of Friends	Social: Relationship with Neighbors	Social: Feeling About Adaptation	Cultural: Gifts to Friends	Cultural: Gifts to Others	Cultural: Son Preference
Cultural Reintegration
Giving Gifts During Happy Events of Relatives and Friends (Reference group = Hardly)
Some	1.699 +	0.865	0.997
	(0.546)	(0.259)	(0.290)
Mostly	1.715 +	1.285	0.909
	(0.479)	(0.336)	(0.227)
Always	2.131 **	1.372	1.210
	(0.602)	(0.362)	(0.308)
Giving Gifts During Happy Events of Others (Reference group = Hardly)
Some	1.251	1.267	0.765 +
	(0.205)	(0.189)	(0.115)
Mostly	1.055	1.079	1.055
	(0.165)	(0.154)	(0.151)
Always	1.127	1.687 **	1.097
	(0.209)	(0.289)	(0.190)
Son Preference (Reference group = Strongly Disagree)
Disagree	0.824	0.689 **	0.499 ***
	(0.112)	(0.0870)	(0.0640)
Neither Agree nor   Disagree	1.032	0.751 +	0.356 ***
	(0.177)	(0.120)	(0.0571)
Agree	0.883	0.549 **	0.424 ***
	(0.186)	(0.106)	(0.0838)
Strongly Agree	0.647	0.808	0.852
	(0.191)	(0.234)	(0.250)
Social Reintegration
Number of Friends/Acquaintances (Reference group = 0)
1 to 10				1.336 *	1.086	0.885
				(0.194)	(0.154)	(0.120)
11 to 20				1.480 +	0.972	0.933
				(0.308)	(0.193)	(0.179)
20 or More				1.800 *	1.537 +	0.893
				(0.444)	(0.361)	(0.201)
Relationships with Neighbors (Reference group = Very Unfamiliar)
Unfamiliar				0.670	1.315	1.353
				(0.441)	(0.909)	(0.856)
Neither Familiar nor   Unfamiliar				1.111	2.094	0.765
				(0.684)	(1.355)	(0.455)
Familiar				1.126	2.144	0.737
				(0.685)	(1.374)	(0.434)
Very Familiar				1.539	2.862	0.637
				(0.939)	(1.837)	(0.376)
Feeling Hard to Adapt to Rural Life (Reference group = Strongly Agree)
Agree				0.811	1.391	2.679 *
				(0.326)	(0.533)	(1.030)
Neither Agree nor   Disagree				0.665	1.392	2.618 **
				(0.257)	(0.511)	(0.971)
Disagree				0.853	1.633	1.800
				(0.317)	(0.575)	(0.644)
Strongly Disagree				1.128	1.689	1.116
				(0.426)	(0.604)	(0.406)
Controls
Demographic Traits	Yes	Yes	Yes	Yes	Yes	Yes
Socioeconomic Status	Yes	Yes	Yes	Yes	Yes	Yes
Migration Experience and Plan	Yes	Yes	Yes	Yes	Yes	Yes
Institutional Factors	Yes	Yes	Yes	Yes	Yes	Yes
Observations	1624	1624	1624	1624	1624	1624
Number of Groups	186	186	186	186	186	186

SE in parentheses. *** p < 0.001, ** p < 0.01, * p < 0.05, ⁺ p < 0.1.

,

Table 5. Connection between social reintegration and civic reintegration (multilevel ordered logistic regression, odds ratio).

Variables	Social Reintegration			Civic Reintegration
	Social: Number of Friends	Social: Relationship with Neighbors	Social: Feeling About Adaptation	Civic: Contribute Labor or Funds	Civic: Participate in Elections
Civic Reintegration
Contribute Labor or Funds (Reference group = Very Unwilling)
Unwilling	0.792	0.935	1.448
	(0.790)	(1.033)	(1.858)
Neither Willing nor   Unwilling	1.041	0.789	1.546
	(1.005)	(0.852)	(1.947)
Willing	1.004	1.052	1.513
	(0.958)	(1.126)	(1.896)
Very Willing	1.239	1.599	2.171
	(1.182)	(1.715)	(2.719)
Participate in Elections of Village Cadre (Reference group = Very Unwilling)
Unwilling	0.539	0.416 +	0.780
	(0.255)	(0.191)	(0.332)
Neither Willing nor   Unwilling	0.324 *	0.302 **	0.868
	(0.151)	(0.136)	(0.362)
Willing	0.474 +	0.326 *	0.827
	(0.213)	(0.142)	(0.332)
Very Willing	0.501	0.560	1.391
	(0.225)	(0.246)	(0.559)
Social Reintegration
Number of Friends/Acquaintances (Reference group = 0)
1 to 10				1.069	0.990
				(0.160)	(0.136)
11 to 20				1.050	1.030
				(0.223)	(0.205)
20 or More				1.889 *	1.274
				(0.497)	(0.304)
Relationships with Neighbors (Reference group = Very Unfamiliar)
Unfamiliar				0.646	0.532
				(0.440)	(0.370)
Neither Familiar Nor   Unfamiliar				0.850	0.588
				(0.545)	(0.388)
Familiar				0.992	0.672
				(0.630)	(0.440)
Very Familiar				1.959	1.142
				(1.250)	(0.751)
Feeling Hard to Adapt to Rural Life (Reference group = Strongly Agree)
Agree				0.760	0.369 *
				(0.348)	(0.154)
Neither Agree nor   Disagree				0.366 *	0.375 *
				(0.162)	(0.151)
Disagree				0.551	0.465 *
				(0.237)	(0.181)
Strongly Disagree				1.158	0.858
				(0.506)	(0.339)
Controls
Demographic Traits	Yes	Yes	Yes	Yes	Yes
Socioeconomic Status	Yes	Yes	Yes	Yes	Yes
Migration Experience and Plan	Yes	Yes	Yes	Yes	Yes
Institutional Factors	Yes	Yes	Yes	Yes	Yes
Observations	1624	1624	1624	1624	1624
Number of Groups	186	186	186	186	186

SE in parentheses. ** p < 0.01, * p < 0.05, ⁺ p < 0.1.

and

Table 6. Connection between cultural reintegration and civic reintegration (multilevel ordered logistic regression, odds ratio).

Variables	Cultural Reintegration			Civic Reintegration
	Cultural: Gifts to Friends	Cultural: Gifts to Others	Cultural: Son Preference	Civic: Contribute Labor or Funds	Civic: Participate in Elections
Cultural Reintegration
GivingGifts During Happy Events of Relatives and Friends (Reference group = Hardly)
Some				1.128	0.948
				(0.343)	(0.270)
Mostly				1.248	1.228
				(0.330)	(0.305)
Always				2.824 ***	2.131 **
				(0.769)	(0.541)
GivingGifts During Happy Events of Others (Reference group = Hardly)
Some				0.856	0.672 **
				(0.137)	(0.0991)
Mostly				1.251	0.949
				(0.196)	(0.135)
Always				1.559 *	1.076
				(0.311)	(0.187)
Son Preference (Reference group = Strongly Disagree)
Disagree				0.515 ***	0.573 ***
				(0.0734)	(0.0741)
Neither Agree nor   Disagree				0.475 ***	0.564 ***
				(0.0839)	(0.0901)
Agree				0.436 ***	0.462 ***
				(0.0932)	(0.0889)
Strongly Agree				0.922	1.287
				(0.308)	(0.400)
Civic Reintegration
Contribute Labor or Funds (Reference group = Very Unwilling)
Unwilling	2.303	0.890	1.622
	(2.298)	(0.991)	(1.795)
Neither Willing nor   Unwilling	1.454	1.722	2.235
	(1.406)	(1.867)	(2.420)
Willing	1.880	1.648	1.901
	(1.797)	(1.770)	(2.042)
Very Willing	4.587	3.031	1.414
	(4.393)	(3.260)	(1.519)
Participate in Elections of Village Cadre (Reference group = Very Unwilling)
Unwilling	0.571	0.568	2.199 +
	(0.288)	(0.265)	(1.023)
Neither Willing nor   Unwilling	0.466	0.465 +	2.831 *
	(0.231)	(0.214)	(1.295)
Willing	0.608	0.613	2.706 *
	(0.293)	(0.273)	(1.200)
Very Willing	0.819	0.577	1.777
	(0.396)	(0.258)	(0.790)
Controls
Demographic Traits	Yes	Yes	Yes	Yes	Yes
Socioeconomic Status	Yes	Yes	Yes	Yes	Yes
Migration Experience and Plan	Yes	Yes	Yes	Yes	Yes
Institutional Factors	Yes	Yes	Yes	Yes	Yes
Observations	1624	1624	1624	1624	1624
Number of Groups	186	186	186	186	186

SE in parentheses. *** p < 0.001, ** p < 0.01, * p < 0.05, ⁺ p < 0.1.

. While these dimensions are distinct, they are not isolated from one another; rather, they exhibit complex interconnections that shed light on the multidimensional nature of the reintegration process that is similar to the case of the integration process in the host society [67,68].

The findings in Table 4 reveal that social reintegration and cultural reintegration are highly connected. Giving gifts during happy events of relatives and friends is positively associated with the number of friends/acquaintances. Compared to those who hardly give gifts, return migrants who “Always” do so have significantly higher odds of having more friends (OR = 2.131, p < 0.01), indicating about 113% greater odds of broader support networks. This suggests that participating in this cultural norm fosters interpersonal ties and expands social networks. Giving gifts during the happy events of others also improves familiarity with neighbors: return migrants who “Always” give gifts in this context are more likely to be familiar with neighbors (OR = 1.687, p < 0.01), or about 69% higher odds than those who hardly do so. Interestingly, the relationship between son preference and social reintegration is non-linear. Returnees with intermediate levels of son preference have lower odds of neighbor familiarity and subjective adaptation, whereas those at the extremes (strongly agree or strongly disagree) do not differ significantly. In general, these patterns suggest that weaker adherence to traditional values may coincide with weaker neighborhood ties and perceived adaptation.

Conversely, social reintegration also promotes cultural reintegration. Return migrants with more friends are more likely to give gifts to relatives and friends. Those with 20 or more friends have approximately 80% higher odds of engaging in this cultural practice. In terms of subjective adaptation, return migrants who feel ‘neutral’ about their adjustment to rural life (i.e., neither agree nor disagree that it is difficult) are significantly more likely to express son preference (OR = 2.618, p < 0.01), more than twice the odds of those who strongly agree. Similarly, those who simply ‘agree’ are also more likely to express son preference (OR = 2.679, p < 0.05). These findings underscore the complexity of how perceived adaptation difficulty interacts with cultural attitudes. Specifically, son preference appears stronger among those who are not at the extremes of adaptation difficulty, pointing to a non-linear relationship between social and cultural reintegration.

Social reintegration and civic reintegration are also highly connected, as shown in Table 5. The results suggest that those more active in civic activities, particularly in participating in elections, tend to have fewer friends or acquaintances who can offer support and are less familiar with their neighbors. However, the pattern is not linear. Only those who are ‘neither willing nor unwilling’ to participate show significantly lower social reintegration, with around 70% lower odds of having more friends or being familiar with their neighbors. This phenomenon may be attributed to their busy engagement in civic endeavors, leaving less time for the establishment of social networks.

Reversely, social reintegration appears to support civic engagement. Return migrants with 20 or more friends are significantly more likely to contribute (OR = 1.889, p < 0.05), indicating 89% greater odds than those with no social ties. These results suggest that when people have stronger support networks, they are more inclined to give back to their communities through civic involvement. However, those who express feeling ‘in-between’—neither fully adapted nor unadapted—are less inclined to participate in both civic activities, with more than 50% lower odds. This suggests that while social support can motivate civic engagement, individual perceptions of social reintegration also play an important role.

Finally, cultural reintegration and civic reintegration are also interconnected, though somewhat more weakly, as shown in Table 6. Those who are in between in terms of willingness to participate in elections have nearly 200% higher odds of expressing stronger son preference. Conversely, those who always give gifts during happy events of relatives and friends are much more likely to participate in both civic activities, with over 100% higher odds. In contrast, individuals with intermediate levels of son preference are less likely to engage in both forms of civic participation, with approximately 43% to 56% lower odds. This asymmetry highlights that the relationship between cultural and civic reintegration is not strictly bidirectional or linear. Rather, the underlying results suggest that civic engagement may reflect integration into community norms (which may include traditional values), while strong adherence to traditional values like son preference may limit broader civic participation.

5. Discussion and Conclusions

Given the considerable challenges tied to internal migration, especially for those returning to rural areas, gaining a deeper understanding of the reintegration process is crucial. This study sheds light on the non-economic dimensions of reintegration—social, cultural, and civic—which have received far less attention in the literature. To the best of my knowledge, this is the first study to empirically examine how these dimensions interact in the context of return rural migration in China.

The analysis focuses on four broad categories of factors shaping reintegration: demographic characteristics, socioeconomic status, migration experience and future intentions, and institutional context. Several key patterns emerge. First, the drivers of reintegration vary across dimensions. For example, institutional barriers are closely linked to social and civic reintegration but appear to have little effect on cultural reintegration. Second, even when the same factor is relevant across dimensions, its impact may differ. A case in point is self-employment, which seems to hinder social and cultural reintegration while supporting civic reintegration. Third, there are notable differences across communities, pointing to the importance of local context in shaping reintegration outcomes.

These findings contribute to theory in several ways. They highlight the multidimensional and non-uniform nature of reintegration: what facilitates reintegration in one domain may have no effect—or even a negative effect—in another. The case of self-employment illustrates this trade-off well. More broadly, the results underscore the importance of looking beyond economic reintegration. Success in the economic domain does not necessarily translate into stronger social ties, cultural adaptation, or civic engagement.

The study also shows that the different dimensions of reintegration are interconnected in both positive and negative ways, and these relationships are frequently non-linear. In some cases, progress in one dimension supports advances in another; in others, improvements are offset by declines elsewhere or follow threshold effects rather than steady increments. These patterns are similar to findings in the integration literature for international migrants [68]. It is thus important to study reintegration using a multidimensional framework. Focusing on one aspect alone risks missing the broader picture.

In terms of policy, the findings highlight several important pathways for targeted intervention. Given differences in determinants and outcomes across social, cultural, and civic reintegration, policies should first recognize the distinct yet interconnected nature of these dimensions. A hierarchical approach is recommended. First, economic and social stability provide the essential foundation for reintegration; thus, interventions ensuring employment opportunities and supportive social networks should be prioritized; by doing so, cultural and civic reintegration can also be promoted indirectly. Second, culturally-sensitive policies—such as supporting local community events, traditions, and inclusive practices—should be developed to foster cultural cohesion and reduce social tensions. However, the focus should not be on cultural values, as their links with other dimensions are not linear. Third, civic engagement initiatives, including programs promoting active participation in community governance, should follow once basic social and cultural foundations have been established. At this stage, return migrants are also more willing to participate, and their engagement can contribute to rural development. Additionally, the significant variation observed across communities underscores the importance of localized, place-based interventions. Tailoring policies to reflect specific community needs and conditions can substantially enhance their effectiveness. Moreover, because reintegration dimensions are interdependent, policymakers should consider potential spillover effects—both positive and negative—and account for non-linear effects between dimensions so that the level of intervention remains balanced and context sensitive. Although this study focuses on internal migration within China, the insights can also inform policies related to return migration internationally.

This study also has several limitations worth highlighting. First, my indicators of cultural reintegration—gift-giving frequency and son-preference—are the closest proxies available in the data, but they cannot capture the full range of local norms, values, and practices. Future work would benefit from more comprehensive measures (e.g., use of local dialects, participation in communal rituals, adherence to family customs) that better reflect the multifaceted nature of cultural reintegration. Second, I do not explore heterogeneity in the reintegration process across different returnee types. Examining how key determinants vary by factors such as returnee occupation would both strengthen the robustness of the findings and offer deeper insight into the contingencies of reintegration. Third, given the cross-sectional nature of the data, the observed links between dimensions are correlational rather than causal; employing longitudinal or panel data would allow future research to untangle the directional and potentially reciprocal effects among social, cultural, and civic reintegration.