Spatiotemporal Forecasting of Urban Financial Risk in the Yangtze River Delta
Weiran Lin, Haijing Yu, Caihong Zhu
Keywords:
urban financial risk, dynamic composite measurement, spatiotemporal modeling, graph neural networks, Yangtze River Delta
Abstract:
Urban financial stability is essential for sustainable regional development and the resilience of urban economic systems. However, existing studies provide limited city-level measurement of urban financial risk and rarely account for spatial interactions among cities. This study addresses these gaps by developing an integrated framework for the measurement and forecasting of urban financial risk in the Yangtze River Delta. A multidimensional indicator system is constructed, covering macroeconomic conditions, banking system risk, real estate market risk, and government finance and debt risk, with environmental factors and implicit local government debt risk indicators incorporated to capture emerging sources of financial vulnerability. Using panel data for 27 cities from 2008 to 2023, a dynamic composite evaluation method with an acceleration effect is applied to measure urban financial risk and analyze its spatiotemporal evolution, revealing significant spatial heterogeneity and clustering across cities. To further enhance predictive performance, a spatiotemporal forecasting framework is developed by integrating Graph Convolutional Networks (GCNs) with temporal deep learning models, where a trainable adjacency matrix is employed to capture inter-urban dependence. Empirical results show that the proposed models outperform benchmark time-series models, with the Serial Long Short-Term Memory–GCN model achieving the best performance in terms of R² and RMSE. Overall, the findings demonstrate that incorporating spatial dependence into temporal modeling significantly improves forecasting accuracy, providing methodological contributions to urban financial risk analysis and supporting the development of more effective early-warning systems.
Fulltext download:
 Spatiotemporal Forecasting of Urban Financial Risk in the Yangtze River Delta [PDF file] [Filesize: 1.17 MB]
 10.7441/joc.2026.03.01
Lin, W., Yu, H., & Zhu, C. (2026). Spatiotemporal Forecasting of Urban Financial Risk in the Yangtze River Delta. Journal of Competitiveness, 16(3), 1–36. https://doi.org/10.7441/joc.2026.03.01
|
