Green Finance Policy and Corporate Investment Efficiency: Evidence from a Quasi-Natural Experiment Using Difference-in-Differences and Panel Data

Abstract

Green finance policies can change firms' financing conditions and investment choices, but their effects are difficult to separate from regional shocks and measurement choices. This study reanalyses a publicly released panel of Chinese listed companies around the 2017 green finance reform and innovation pilot program. The analysis audits exposure coding, fixes firms' provincial membership, excludes provinces entering later pilot rounds, and estimates difference-in-differences models with firm and year fixed effects. The resulting sample contains 18,753 firm-year observations from 2,100 companies in 29 provinces, including five exposed provinces. The main coefficient on the published absolute investment residual is −0.00419. Its province-clustered 95% confidence interval is [−0.01022, 0.00185], and the null-imposed wild-cluster-bootstrap p-value is 0.403. Clustering the same specification by firm instead produces a p-value of 0.031, illustrating the importance of aligning inference with policy variation. Event-study estimates, alternative samples, and leave-one-exposed-province-out checks identify additional uncertainty, including substantial dependence on Guangdong. The evidence is consistent with a reduction in the investment-inefficiency proxy, but does not establish a precisely estimated average provincial effect. The study provides a reproducible framework for evaluating policy exposure, generated outcomes, and small-treated-cluster inference in green finance research.