Can digital-real integration effectively curb urban sprawl? Evidence from China using double/debiased machine learning

Abstract

Urban sprawl (URS) has become a major challenge to China’s sustainable development under the backdrop of rapid urbanization. Digital-real integration (DRI) provides a new pathway for urban spatial governance, yet its concrete impacts and mechanisms remain insufficiently tested. Based on panel data from 280 prefecture-level Chinese cities between 2013 and 2023, this study constructs a multidimensional DRI index and employs a partially linear double/debiased machine learning (DML) framework to estimate the effect of DRI on URS. Multiple robustness checks, including complementary instrumental-variable analyses, yield results consistent with the baseline finding. The findings show that a one-unit increase in DRI is associated with a 0.06-unit reduction in the URS index, and this conclusion holds under various robustness tests. The mechanism analysis provides evidence consistent with three potential transmission pathways: industrial structure upgrading, stronger aggregate city-level correspondence between employment opportunities and resident population, and reduced dependence on highway commuting. Further heterogeneity analysis reveals that the restraining effect of DRI is more pronounced in eastern regions, major national urban agglomerations, and cities at more advanced stages of urban development. This study not only provides empirical evidence on how digital technologies reshape urban spatial structures but also offers policy implications for implementing differentiated digital governance strategies and enhancing the effectiveness of spatial planning.