The rational allocation of talent resources is significant to regional transformation and upgrading high-quality development. Focusing on urban agglomerations in China, this study examines the spatial patterns and underlying mechanisms of graduate talent mobility using employment data from the Ministry of Education Graduate Employment Quality Reports. We utilized the social network analysis method, stickiness rate, external attractiveness index, and directed migration model. The results reveal the following. (1) Spatial Stratification and Typology: A significant “Matthew Effect” characterizes China’s talent landscape. While the Yangtze River Delta and Pearl River Delta exhibit a “high stickiness–high attractiveness” dual-drive pattern, emerging inland agglomerations like Chengdu–Chongqing rely on high internal stickiness as a critical “stabilizer,” maintaining regional resilience through local stock retention despite limited external pull. (2) Complexity of Driving Mechanisms: Ridge regression indicates that while economic development (GDP per capita) and innovation capacity remain core drivers of external attractiveness, public services and institutional costs exert stronger constraints on mobility. (3) Policy Implications: In contrast, monetary talent policies show limited marginal utility. The study concludes that talent governance in urban agglomerations must shift from homogenous “talent wars” to differentiated sustainable strategies. Advanced regions should foster polycentric networks to mitigate overcrowding, while emerging regions should prioritize “soft infrastructure” to lower social costs, leveraging endogenous stickiness for long-term human capital accumulation and spatial equity.
Cui et al. (Thu,) studied this question.