Green Finance Policy Impact Analysis Based on Integrated Knowledge Graph

Due to the inherent fragmentation of policy, market, ESG and carbon data, tracking genuine sustainable impact remains challenging. This paper addresses prevailing issues in green finance policy impact assessments—specifically, unaccounted data heterogeneity, omitted-variable bias, and fragmented multi-source information—by applying an analytical approach based on an integrated Knowledge Graph (KG). The method utilizes a multidimensional KG constructed from policy, market, ESG (Environmental, Social, and Governance), and carbon data to mitigate fragmentation in green finance assessment. Natural Language Processing (NLP) is employed to structure policy knowledge, and a graph embedding algorithm is designed to quantify dynamic policy impacts on the scale of green credit, corporate emission reduction, and regional carbon intensity. Validation via a Difference-in-Differences (DID) model using data from 2016 to 2023 in China’s green finance pilot zones demonstrates a significant positive policy effect: the green credit share increases by an average of 2.15 percentage points, and reductions in carbon intensity are confirmed. The KG method achieves 89.7% accuracy, representing a 22.5-percentage-point improvement over traditional regression methods. The study concludes that integrating KG technology substantially enhances visualization and causal inference capabilities in green finance policy impact analysis, enabling policymakers to trace cause-and-effect relationships—from the implementation of a new green credit scheme to measurable sustainability improvements—thus providing robust decision support for optimizing the policy tool mix.