Rethinking energy allocation: Can green finance be the solution? — Evidence from machine learning method

In the context of intensifying energy system reform, distortions in energy resource allocation have emerged as a substantial constraint on green development. Considering the limited effectiveness of traditional policy instruments in correcting structural misallocation, it is uncertain whether Green Finance Reform and Innovation Pilot Zones (GFP), which are institutional innovations aimed at promoting the green transition, can effectively reshape resource allocation. Employing a double machine learning difference-in-differences (DML-DID) framework, this paper investigates the effect of GFP on energy resource misallocation (ERM) across 30 Chinese regions from 2011 to 2022. The empirical results demonstrate that GFP exerts a significant mitigating effect on ERM. The robustness of this conclusion is established through multiple tests, including the Goodman-Bacon decomposition, the Callaway and Sant'Anna estimator, and synthetic DID. Mechanism analysis, from the perspectives of social structure and agent behavior, reveals that GFP helps alleviate ERM by promoting industrial upgrading, enhancing financial agglomeration, advancing technological innovation, and increasing environmental investment. Heterogeneity analysis indicates that the reduction in ERM is more pronounced in regions characterized by high resource dependence and low levels of marketization. This study contributes to the emerging framework linking green finance and resource allocation by providing causal evidence on how GFP addresses structural inefficiencies in resource allocation and offers empirical support for regionally tailored policy interventions.