Family-to-family child migration network of informal adoption in China

The paper addresses the under-researched phenomenon of informal (unregistered) child adoption in China, where adoption has long been used to cope with family crises and resource constraints. In the context of strict birth control, particularly the one-child policy introduced in 1978, some families privately sent additional children away, often without legal procedures, leaving children without legal identity and access to services. Informal adoption overlaps conceptually with crimes like child abandonment and trafficking, complicating legal judgments and research. The study clarifies informal adoption as cases where birth families do not seek profit and adopting families do not complete legal procedures, aligning with the best interests of the child. The research aims to explore the temporal-spatial patterns and migration networks of informal adoption (1924–2018) using large-scale public welfare website data, to understand how social and policy forces shaped these patterns, highlight gender dynamics linked to son preference, and provide insights relevant to protecting children’s rights and explaining China’s historically high sex ratio at birth.

Prior scholarship on adoption in China largely examined legal frameworks and procedure development (e.g., Bai 1998; Wang 2000; Feng 2008; Zhang et al. 2010; Yang 2015) or conducted localized surveys of adoption practices (Li 2000; Yang 2004; College 2006). Broader demographic and legal analyses used census data (Johansson and Nygren 1991; Sten 1995), court judgments (Cai and Xin 2019; Huang and Weng 2019), and public welfare website data (Li et al. 2017a, 2017b, 2018, 2019; Wang et al. 2018) to study irregular child migration and trafficking. Informal adoption has often been conflated with illegal adoption and child trafficking, obscuring differences in motives and processes. Notably, trafficking studies identify demand-driven gender preferences (favoring boys) and distinct source/target geographies. This paper builds on these strands by explicitly distinguishing informal adoption from trafficking, leveraging public welfare website data to examine long-term spatiotemporal patterns and social network structures of informal adoption at provincial and city levels.

Data source and scope: The authors crawled records from the Baby Coming Back Home website (http://www.baobeihuijia.com), where users can freely register information about seeking family connections. Using the keywords “Bao Yang” (informally adopted in) and “Song Yang” (informally adopted out), they collected 16,041 informal adoption entries dated from 1924 to November 26, 2018 (00:00). After excluding entries labeled “legal adoption,” “suspected trafficking crime,” “missing,” and “abandonment,” 15,685 cases remained. The study area covered all of China; base map administrative boundaries were from the 2015 Resources and Environment Data Cloud Platform (http://www.resdc.cn/). Children were defined as under 18 years of age. Analytical tools: ArcGIS was used to visualize spatial patterns and migration paths. UCINET and Gephi were employed for social network analysis at the province and city levels. Measures: - Gini-Hirschman coefficient (geographic concentration): Children were treated as a special commodity to measure concentration of out-adoption and in-adoption locations over time. GO_t = 100 * sqrt(Σ_j (X_jt / X_t)^2), GI_t = 100 * sqrt(Σ_j (I_jt / I_t)^2), where m is the number of locations, X_jt is the number sent from area j at time t, and I_jt is the number sent to area j at time t. Higher values indicate more concentrated distributions. - Social network analysis: Provinces (and later cities) were modeled as nodes; directed edges represented migration paths weighted by case counts. Centrality and betweenness metrics were used to evaluate node importance and control over information flow; edge betweenness captured key connecting paths. Normalized indices were used to compare in/out quantities and degrees across provinces with differing populations. Visualization and thresholds: Spatial distributions were mapped by province and city; ratios of in- to out-adoptions identified inflow vs. outflow provinces. City-level networks used the Yifan Hu force-directed layout to identify clusters and hubs. Path dependence was considered present when a connection occurred more than twice, indicating a stable information channel.

- Temporal dynamics and policy/disaster effects: Severe famine periods and strict birth control campaigns markedly increased informal adoptions. During strict one-child policy enforcement, son preference intensified. In the post-2013 era (two-child policy), overall informal adoptions declined and gender imbalance narrowed; the mean sex ratio rebounded, with one reported minimum sex ratio of 28.94 during earlier periods of strong son preference and selective adoption. - Age and gender: Informally adopted children were predominantly under 1 year old, especially after the onset of birth control policies; gender bias reflects supply-side behavior (birth families giving away girls). The post-2013 period showed continuing predominance of infants and a narrowing male–female gap (mean sex ratio 67.35 reported), suggesting weakening son preference. - Spatial concentration: Informal adoptions were more prevalent in provinces and zones with higher population densities; spatial concentration increased in the most recent period (reported Gini-Hirschman coefficients for out- and in-adoptions of 23.36 and 52.92, respectively, in the small-fluctuation period post-2013). - Provincial roles (normalized indices): • Key source provinces (highest normalized out-quantity): Sichuan (2.00), Jiangsu (1.75), Henan (1.66). Other notable sources include Guangdong, Anhui, Hubei, Shandong. • Key target provinces (highest normalized in-quantity): Henan (1.91), Hebei (1.79), Shandong (1.55), Guangdong (1.18). Henan functioned as both a major source and target. • Outflow imbalance (in/out ratio): Strongest outflow provinces included Hainan (0.22), Yunnan (0.25), Guizhou (0.26), Sichuan (0.26), Xinjiang (0.28) (lower ratios indicate net outflow). Strong inflow (ratio > 1.03): Hebei, Shandong, Tianjin, Beijing, Henan, Fujian. • Overall spatial pattern: An “inverse T” shape characterized by sending from central regions toward eastern or western directions and intake by northern/southern coastal regions. - Migration distances and shares: 58.4% of informal adoptions occurred within the same province; 37.47% occurred within the same city. Interprovincial paths concentrated in central and eastern China, often short-distance flows from poorer south-central/southwestern provinces to more developed eastern provinces. High-incidence interprovincial paths included Shaanxi → Henan, Shanxi → Hebei, Jiangsu → Shandong, Shanxi → Henan; also Sichuan → Hebei and Sichuan → Henan. - City-level distributions: Major out-adoption cities included Chongqing (638 cases; 4.15% of out-adopted children), followed by Shanghai (2.30%), Wuhan (1.67%), Xi’an (1.63%), Chengdu (1.50%), Hefei (1.48%). Major in-adoption cities included Putian (4.44%), Chongqing (3.12%), Shanghai (3.05%), Shijiazhuang (2.70%), Zhengzhou (2.27%), Xuzhou (2.20%), Beijing (2.13%). Provincial capitals and large municipalities showed strong involvement, consistent with roles as economic and transport hubs. A notable intra-provincial path was Fuzhou (Fujian) → Putian (Fujian) (2.87% of total), far exceeding other paths such as Suzhou (Anhui) → Xuzhou (Jiangsu) (0.52%). - Network structure and hubs: • Node betweenness (top cities): Chongqing (0.1326), Shanghai (0.1168), Beijing (0.0725), Wuhan (0.0621), Luoyang (0.0530), Xi’an (0.0522), Suzhou–Jiangsu (0.0498), Xuchang (0.0468), Nanjing (0.0463), Handan (0.0424), among others; central and eastern hubs and the Sichuan–Chongqing agglomeration dominated. • Edge betweenness (top paths): Luoyang → Chongqing (0.0524), Beijing → Xuchang (0.0480), Handan → Suzhou–Jiangsu (0.0393), Shanghai → Xi’an (0.0384), Guangzhou → Zhanjiang (0.0322), etc. Overall edge control was low; the top 20 paths (2.93%) accounted for only 2.61% of total path control. • Clusters: The city network (density ~0.001) exhibited four tight-knit clusters, including an inland–coastal group (Xi’an–Zhengzhou–Shanghai hub), a Southwest–North China group (Chongqing and Shijiazhuang as targets, Dazhou as connector), a Putian-centered Fujian cluster, and a Beijing-centered mini-cluster. - Comparison with trafficking: Informal adoption is supply-driven (birth families giving away girls, often infants), contrasting with child trafficking’s demand for boys and older ages (3–5 years). Geographic patterns overlap partly (e.g., Sichuan, Jiangsu, Henan as sources; Henan, Hebei, Shandong as targets), highlighting policy-relevant hotspots.

The findings clarify the distinct nature of informal adoption relative to child trafficking and reveal how major social forces—famine, war-time hardship, and stringent birth control—shaped the timing, geography, and gender composition of informal adoptions. The predominance of infants and the female bias among informally adopted children indicate supply-side decisions by birth families under policy and economic pressures. Spatial analyses demonstrate an “inverse T” distribution with central provinces as sources and coastal/eastern provinces as targets, mirroring broader patterns of economic development and migration. Network analyses identify key urban hubs (Chongqing, Shanghai, Beijing) and strong intra-provincial corridors (e.g., Fuzhou→Putian), suggesting that major cities serve as both information conduits and destinations. These insights address the study’s goals by providing empirical evidence on when and where informal adoption concentrates, how it propagates through social networks, and how policy interventions (e.g., two-child policy, gender equality advocacy) correlate with declines in cases and reduced gender imbalance. The results furnish actionable guidance for targeted monitoring and child protection efforts in identified hotspots and along high-incidence paths, and they help interpret anomalies in historical sex ratios at birth.

This study operationalizes a clear, non-criminal definition of informal adoption and, using large-scale public welfare website data (1924–2018), maps its spatiotemporal evolution and migration networks across China. It identifies policy and disaster periods that amplified informal adoptions, documents strong supply-side gender bias (favoring giving away girls), reveals concentrated spatial patterns and key provincial/city hubs, and delineates prominent corridors and clusters. These contributions provide a baseline for authorities to target interventions, improve information platforms for family reunification, and distinguish informal adoption from trafficking in policy and research. Future work should integrate multi-source data and mixed methods to improve coverage and accuracy; quantify policy impacts (e.g., one-child and two-child policies) on informal adoption dynamics; conduct comparative analyses across informal adoption, trafficking, and abandonment; and pursue case studies tracing life trajectories and decisions to return to birth families, aiming to design better protections for informally adopted children.

The dataset is self-reported and drawn from a public website, introducing potential reporting errors and selection bias. Many families may be unwilling to disclose adoption information online, limiting coverage. Informal adoption postings can be inconsistently labeled or conflated with other categories, reducing retrieval precision. These factors may affect the completeness and generalizability of findings, though the study provides an initial, large-scale exploratory baseline.