The rapid urbanization and industrialization in China have led to a massive influx of rural migrant workers into urban areas, providing crucial labor for the non-agricultural sector. However, this migration also leaves behind an aging rural population and creates challenges for rural revitalization. The Chinese government, under President Xi Jinping's rural revitalization strategy, aims to encourage the return of migrant workers to contribute to rural development. This study focuses on understanding the factors influencing migrant workers' willingness to return home (MWRH). A significant challenge faced by migrant workers is their difficulty integrating into urban society, often encountering social exclusion and discrimination. This social exclusion impacts their sense of identity and may contribute to their decision to return home. Therefore, this study examines the influence of social discrimination on MWRH, considering the mediating role of identity. Understanding this relationship is crucial for developing effective policies to support both rural revitalization and urban-rural integration.

Existing literature on MWRH primarily focuses on two aspects: the "pull" factors from rural areas (favorable policies, improved infrastructure, etc.) and the "push" factors from urban areas (lack of social security, high cost of living, discrimination, etc.). Studies have highlighted the influence of factors like income, education, household registration policies, children's education, and social network exclusion on MWRH. The concept of social exclusion, often used interchangeably with discrimination, stems from theories like Frank Parkin's Social Exclusion theory, emphasizing the role of factors like family background, language, and religion in excluding groups. Prior research on social discrimination against migrant workers in China has explored its causes (institutional factors, income disparity), effects (reduced willingness to settle in cities, weakened cultural identity), and potential solutions. Studies have also explored the importance of urban identity in migrant workers' social integration and its influence on their MWRH. However, the mediating role of identity in the relationship between social exclusion and MWRH remains under-researched, prompting this study to address this gap.

This study utilizes data from the 2017 China Migrants Dynamic Survey (CMDS), a nationally representative survey of migrant workers. The sample includes 47,011 individuals aged 15–59. The dependent variable, MWRH, is a binary variable indicating whether a migrant worker intends to return home. The key explanatory variable is a composite social discrimination index constructed from five dimensions: interpersonal interaction discrimination, labor market discrimination, occupational discrimination, cultural discrimination, and regional discrimination. Each dimension is measured using relevant questions from the CMDS. For example, interpersonal interaction discrimination is measured by asking respondents about the people they interact with most frequently. A higher score on the index indicates a higher level of discrimination experienced. Identity is the mediating variable, operationalized using questions assessing the migrant workers' sense of belonging and acceptance in the city. The study also includes a range of control variables at the individual, family, and mobility levels (gender, ethnicity, age, education, income, marital status, etc.). A Probit model is employed to analyze the relationship between social discrimination, identity, and MWRH, along with a mediating effect test using the Baron and Kenny (1986) method and the Sobel test, further validated using the Bootstrap test. To address potential endogeneity issues, an instrumental variable (IV) Probit model is used with "whether to apply for a residence permit" as the instrument. Robustness checks involve using the Logit model and the entropy evaluation method (EEM) to calculate the social discrimination index.

The results support the hypothesis that social discrimination significantly and positively affects MWRH. The marginal effect of social discrimination is 0.96%, indicating a 0.83% increase in the probability of returning home for each unit increase in the discrimination index. This relationship remains robust even after using instrumental variables to address endogeneity. The study also confirms the mediating role of identity. Social discrimination negatively affects identity, and low identity, in turn, increases the willingness to return home. The mediating effect accounts for 50.77% of the overall effect of social discrimination on MWRH. Heterogeneity analyses reveal that the effect of social discrimination on MWRH is stronger among: (1) rural residents in the western and northeastern regions, (2) older generations, (3) those with lower incomes, and (4) those with limited intra-provincial mobility. These findings are consistent across different model specifications (Probit vs. Logit) and different methods of measuring social discrimination. The robustness of the findings is further supported through additional sensitivity analysis by removing the samples of workers who are 20 years old and younger and/or 60 years and older.

The findings demonstrate the significant role of social exclusion and identity in shaping migrant workers' decisions to return home. The strong positive effect of social discrimination on MWRH highlights the importance of addressing the social integration challenges faced by migrant workers in urban China. The mediating role of identity suggests that interventions aimed at fostering a sense of belonging and acceptance among migrant workers can effectively reduce their desire to return to their hometowns. The heterogeneity observed across different subgroups emphasizes the need for targeted policies to address the specific challenges faced by particular segments of the migrant population. The study's findings offer important implications for policymakers seeking to promote rural revitalization and enhance urban-rural integration.

This study contributes to the literature by demonstrating the significant and robust relationship between social discrimination, identity, and the willingness of migrant workers to return home. The findings highlight the need for policies that foster social inclusion, promote a positive image of migrant workers, and improve their quality of life in urban areas. Future research could investigate specific policy interventions designed to enhance social integration and strengthen the sense of urban identity among migrant workers. Further research could explore the long-term impacts of return migration on both rural and urban areas.

While the study utilizes a large, nationally representative dataset, there are some limitations. The cross-sectional nature of the data limits the ability to establish causal relationships definitively. The reliance on self-reported measures of social discrimination and identity may introduce potential biases. The study focuses primarily on the individual-level factors, while ignoring the potential influence of macroeconomic factors on MWRH. Future studies should explore the effects of these factors on migrant workers' decisions.