The impact of green finance reform and innovation pilot zones on agricultural green development: a causal inference analysis using double machine learning

Green finance is a key mechanism for alleviating agricultural financing constraints and achieving a green transition in agriculture. Drawing on provincial panel data for 2011–2022, this paper measures agricultural green total factor productivity (AGTFP) using the super-efficiency SBM approach with the GML index, incorporating data elements among the inputs. It then examines the impact and mechanisms of Green Finance Reform and Innovation Pilot Zones (GFRIPZs) on agricultural green development, using double machine learning (DML) and a generalized random forest. The results show the following. (1) The establishment of GFRIPZs considerably fosters agricultural green development, and this finding remains robust under various tests; (2) GFRIPZs benefit agricultural green development most in the eastern and central regions, and also in provinces with lower agricultural green development and financial deepening and intermediate marketization, and the policy improves AGTFP mainly through managerial catch-up; (3) These GFRIPZs advance agricultural green development primarily by fostering green technology innovation in agriculture, reducing resource misallocation, and enhancing large-scale agricultural operations. This paper suggests refining the policies linked to these pilot zones, promoting green technology innovation and large-scale production in agriculture, reducing resource misallocation and tailoring measures to local conditions.