Abstract
The management and prevention of government debt risk is a global topic. In China, due to problems such as implicit debt and uneven regional fiscal performance, it is particularly necessary to explore how to effectively measure and prevent local government debt risks. In this article, we comprehensively consider the debt status and fiscal performance to design a local government debt risk assessment system. According to the debt risk index (DRI), we define the debt risk levels of local governments and find that debt risk has a rapidly increasing pattern and distinct regional characteristics. In addition, we further design a machine learning-based early warning system to predict the risk level of local government debt in the future. We extensively collect explanatory variables based on the previous literature and illustrate variables with high feature importance. Finally, our local government debt risk early warning system achieves an overall accuracy of 92% on the testing set and has a better performance by comparing it to the general debt risk indicator.
Recommended Citation
Zhang, Shuai and Zheng, Haichao, "Local Government Debt Risk Assessment And Early Warning System Based On Machine Learning" (2022). WHICEB 2022 Proceedings. 6.
https://aisel.aisnet.org/whiceb2022/6