The impact of labor mobility with fellow townsmen on the wages of rural migrants: evidence from China

The paper examines whether moving with fellow townsmen (people from the same county) affects the wages of Chinese rural migrants. Against the backdrop of rapid urbanization and reforms to the Hukou system, hundreds of millions of rural residents migrate for better jobs. In China’s vernacular society, kinship and geographic ties structure social networks that may reduce job search costs and risks. Existing research offers mixed findings on ethnic or hometown aggregation: some studies highlight positive wage effects via information sharing and better job matching, while others suggest potential negative effects due to segregation or exploitation. The authors note limited causal evidence on how moving with fellow townsmen affects wages, potential selection bias in prior studies, and a lack of analysis of heterogeneous impacts across occupations, enterprise ownership, and generations. This study uses CMDS 2017 data and robust methods (OLS and PSM) to identify the wage effect and its mechanisms, contributing novel evidence within the Chinese context.

The literature spans assimilation theories and social networks. Straight-line assimilation posits inevitable convergence to natives over time, whereas segmented assimilation emphasizes context-dependent, selective adaptation shaped by ethnic attributes and structural constraints. Empirical work shows both benefits and drawbacks of ethnic/hometown aggregation: networks can disseminate job information, improve job matching, provide social support, and enhance earnings; yet clustering may impede integration or correlate with lower earnings in some contexts. Human capital theory (Mincer; Becker) frames wage determinants as schooling and experience, but labor market segmentation and bargaining power also matter. For China, hometown-based networks resemble ethnic enclaves, facilitating recruitment and promotion by employers, reciprocity among migrants, and potentially greater bargaining leverage. Prior work in China suggests hometown clustering can raise earnings, especially where informal networks mitigate information asymmetry, though negative findings exist as well. The authors develop three hypotheses: (1) moving with fellow townsmen increases wages; (2) it does so by improving job information access; and (3) it strengthens wage bargaining power, further boosting wages.

Data: The study uses the 2017 China Migrants Dynamic Survey (CMDS) from the National Health Commission (NHCPRC), a nationally representative cross-sectional survey of migrants aged 16+ living outside their home county for more than one month. Sampling used stratified, multi-stage PPS; the total national survey is ~170,000. The authors obtained data via agreement with NHCPRC. The analytic sample focuses on rural migrants (agricultural/rural-transfer/resident household registration), aged 16–60, with flow time >6 months, employed (excluding employers, self-employed, and others) and non-negative wages. To avoid multiple-flow confounding, the sample retains cases where “this movement time” coincides with “first time to leave household registration.” After excluding missing key variables, N=16,699 remained (655 moved with fellow townsmen only; 16,044 moved without fellow townsmen-only). Variables: Dependent variable is monthly wage/net income (log-transformed). Key independent variable indicates whether, at first move, the migrant moved only with fellow townsmen (1), versus any other configuration (including moving alone or with both fellow townsmen and family/classmates/relatives) coded 0, to isolate pure hometown co-mobility. Controls include individual characteristics (years of schooling; work experience and its square; gender; marital status; CCP membership) and job characteristics (industry: primary, secondary, producer services, consumer services [reference]; occupation: white-collar, business services, consumer services, blue-collar, non-fixed [reference]; ownership: state-owned, private, foreign-capital, other [reference]; inflow region: eastern, central, western, northeastern [reference]). Model: An extended Mincer wage equation regresses log wages on the hometown co-mobility indicator, schooling, experience (and squared), and the control vectors. Estimation: (1) Descriptive statistics and mean-difference tests; (2) OLS regressions with progressively richer controls; (3) Propensity Score Matching (PSM) to mitigate selection on observables with neighbor (1:4), radius (0.01), and kernel matching; ATT is reported. Robustness: (a) Excluding migrants whose parents had migration experience to address intergenerational network confounding; (b) restricting to first-move windows close to the 2017 survey (2012–2016; and 2016 only) to reduce timing mismatch. Mechanisms: Mediation-type analyses assess (i) information search via a “network of fellow townsmen” measure based on who migrants mostly contact locally, and (ii) wage bargaining via participation in hometown association activities, using probit marginal effects for mediators and OLS for wages, with standard errors clustered at province level.

Descriptive: Mean log wages are higher for those moving with fellow townsmen (8.164) than solitary movers (8.012), with a significant difference at 1%. Migrants moving with fellow townsmen have fewer years of schooling but more work experience and are more likely male and blue-collar; both groups predominantly work in private enterprises, with the eastern region as the main inflow area. OLS: The hometown co-mobility coefficient remains positive and significant across specifications. With full controls (industry, occupation, ownership, region), the coefficient is 0.075 (SE=0.017), significant at 1%, implying a wage premium for those who moved only with fellow townsmen at first move. Schooling and experience have expected positive effects (with a negative quadratic for experience). PSM: ATT estimates are positive and significant, corroborating OLS: neighbor matching ATT ≈ 0.070 (t≈3.36), radius ≈ 0.077 (t≈4.13), and kernel ≈ 0.102 (t≈5.57), indicating robust wage premiums in the range of about 7–10% in log points. Robustness: After excluding migrants whose parents had migration experience, the co-mobility effect remains positive and significant (e.g., coefficients ≈ 0.067***; ≈ 0.090** for 2012–2016 first movers; and ≈ 0.156*** for 2016 first movers), addressing parental network transmission and timing concerns. Mechanisms: Information search—moving with fellow townsmen increases the likelihood of strong local fellow-townsmen networks (probit marginal effect ≈ 0.159, SE≈0.021), and both the network variable and a hometown association indicator raise wages (association coefficient ≈ 0.022*, SE≈0.009); the direct effect of co-mobility on wages remains positive (≈ 0.071**, SE≈0.017). Wage bargaining—co-mobility raises wage bargaining ability (proxied by participation in hometown associations; marginal effect ≈ 0.117**, SE≈0.018), and bargaining ability increases wages (≈ 0.052***, SE≈0.011); the direct co-mobility effect on wages remains positive (≈ 0.068**, SE≈0.017). Heterogeneity: By occupation—significant positive effects in business services (≈ 0.358**), consumer services (≈ 0.112***), and blue-collar (≈ 0.043**), but not in white-collar jobs. By enterprise type—no significant effect in foreign-capital firms; positive effects in private firms (≈ 0.075**) and suggestive in state-owned enterprises (≈ 0.129, not consistently significant). By generation—stronger positive effects for older-generation migrants (≈ 0.107**); effects for the new generation are weaker and not statistically significant. Family co-movement: Moving with spouse, parents, or children shows no significant wage gains; qualitative interpretation suggests family caregiving leads to preference for more stable, lower-paying jobs, effectively trading off wages for stability.

The findings directly address the research question, showing that moving with fellow townsmen at first migration yields higher wages for rural migrants in China. The results are robust to controls, selection on observables (PSM), parental migration experience exclusion, and timing restrictions. Mechanistically, hometown co-mobility enhances access to job information and strengthens wage bargaining power through fellow-townsmen networks and participation in hometown associations, which mitigates information asymmetry and improves negotiation leverage in segmented labor markets. The heterogeneous impacts indicate the benefits are concentrated in occupations and enterprise types where informal recruitment and network introductions are prevalent (consumer services, blue-collar, private and state-owned firms), while formalized recruitment (white-collar, foreign-capital firms) shows limited network wage effects. Intergenerational differences suggest older migrants rely more on traditional hometown networks, magnifying gains, whereas newer generations’ urban integration and human capital diminish marginal network benefits. Policy implications include building formal employment service platforms, expanding accessible job information channels (intermediaries, dedicated job fairs, labor cooperation bases), and strengthening labor rights protection systems (dispute mediation/arbitration, union/social welfare support) to reduce dependence on informal networks and ensure equitable outcomes. The patterns likely generalize to other developing contexts and to regions influenced by Confucian culture where kinship/geographic ties structure migrant networks.

Using nationally representative CMDS 2017 data, the study shows that rural migrants who moved only with fellow townsmen at their first move earned significantly higher wages than comparable migrants who did not. This effect persists across OLS and PSM approaches and after robustness checks excluding parental migration influence and aligning move-to-survey timing. Two mechanisms underpin the premium: enhanced information access via fellow-townsmen networks and improved wage bargaining power (proxied by participation in hometown associations). The wage gains are concentrated in producer/consumer services and blue-collar jobs, and in private and state-owned enterprises; effects are stronger for the older generation, and absent for white-collar and foreign-capital contexts. Future research should examine second and subsequent moves, employ panel data to study dynamics, and further probe how institutional contexts mediate network effects on wages.

The study analyzes only the first migration episode, not subsequent moves; the 2017 cross-sectional design limits timeliness and causal dynamics; lack of panel data restricts assessment of the evolving impact of hometown co-mobility over time. Additionally, the CMDS first-move co-mobility measure cannot capture all later co-movement patterns, and unobserved factors beyond observables matched in PSM may remain.