|
|
|
| Does AI Adoption by Commercial Banks Enhance Credit Support for Corporate Green Transition？ |
| ZHONG Qian
|
| School of Finance,Guangdong University of Foreign Studies |
|
|
|
Abstract Amid China's comprehensive green transition, a structural conflict has emerged between the urgent need for high-carbon enterprises to decarbonize and their limited access to stable external financing. As a structural financial policy tool, green finance faces persistent challenges in identifying truly green or transitioning firms, a problem rooted in green information asymmetry. This asymmetry results in a mismatch: the supply of green credit is concentrated in fully green enterprises, while demand is concentrated in high-emission firms undergoing low-carbon transition. Addressing this allocation mismatch is critical to improving both the equity and efficiency of green financial resource distribution and supporting the real economy's decarbonization pathway. Artificial intelligence (AI), as a disruptive general-purpose technology, offers new potential to address this problem. Despite AI's proven utility in areas such as credit risk modeling and robot-advisory services, existing literature provides limited evidence on how AI can enable green finance—particularly under China's policy-driven structural financial system. Prior studies often emphasize the broad effects of fintech on the real economy while overlooking how fintech enables structural financial policies to have differentiated impacts on the real economy.
To fill this gap, this paper starts from the perspective of enterprise loan cost and financing constraint and investigates whether and how AI adoption by commercial banks enhances green credit allocation by improving banks' ability to identify green or transitioning firms—what we refer to as their green recognition capability. This paper develops a theoretical model of banks' optimal loan pricing under green financial policy constraints and green information asymmetry, in which AI adoption influences banks' identification precision and thus their credit allocation behavior. The model's predictions are empirically tested using firm-level matched loan data for all A-share listed companies in China from 2013 to 2022. The results show that each 1% increase in a bank's AI adoption rate is associated with a 5.06 basis point reduction in the average loan interest rate markup, with a more pronounced reduction of 7.28 basis points for green firms, while the effect is statistically insignificant for non-green firms. From the perspective of financing constraints, each 1% increase in AI adoption is associated with a 1.61 basis point decrease in firms' financing constraints measured by the KZ index, driven mainly by a 1.40 basis point reduction for green firms. These findings suggest that AI adoption enables banks to better serve green enterprises without adversely affecting non-green firms, effectively reducing both loan costs and financing frictions for green firms. Crucially, this paper finds that the impact is significantly greater for firms with lower green information transparency, providing empirical support for the mechanism whereby AI strengthens banks' green recognition capability and enhances real economy outcomes. The effect is more pronounced for banks with faster loan growth and in regions with higher environmental spending.
Based on these findings, three policy recommendations are proposed. First, promote the deep integration of AI and green finance by strengthening banks' green identification capabilities through unified green standards, transparent data infrastructure, and supportive regulatory frameworks. Second, accelerate the development of transition finance standards by leveraging AI to identify and support decarbonization in high-carbon sectors, thereby reducing banks' perceived risk premiums and improving credit access for transitioning firms. Third, adopt differentiated, region-and sector-specific policies to pilot AI-powered green finance reforms, enabling scalable policy experimentation and tailored support for diverse transition pathways.
This paper contributes to the literature in three ways. First, it provides theoretical and empirical evidence on the feasibility and necessity of integrating AI with green finance, a topic at the frontier of interdisciplinary research between economics and artificial intelligence. Second, it expands the analysis of structural financial policy heterogeneity, demonstrating how targeted financial instruments, when empowered by technology, can yield differentiated real economy outcomes. From the perspective of artificial intelligence enabling green finance, it provides direct evidence for China's comprehensive green transformation. Third, it offers a novel perspective on addressing the financing gap for non-green firms during green transition, and provides actionable insights into how AI can help overcome green information asymmetries and improve the precision of financial support for decarbonization in high-emission sectors.
|
|
Received: 02 September 2024
Published: 02 May 2025
|
|
|
|
|
|
|
|
