摘  要
近年来，在国家“十四五”规划及《金融科技发展规划（2022-2025年）》等政策推动下，金融科技与金融机构数字化转型已成为深化金融供给侧改革的核心战略。商业银行依托人工智能、区块链等技术加速业务模式革新从传统线下服务向数字普惠金融延伸，通过精准风控模型和场景化金融产品扩大了服务覆盖面。然而，这一转型在提升服务效率与普惠性的同时，亦催生了长尾客户信用风险传导加剧、银行竞争加剧等新型挑战。在此背景下，本文聚焦数字普惠金融对商业银行信用风险承担的影响机制，通过实证分析探讨其带来的效应，旨在为平衡金融创新与风险防控提供理论依据，助力银行业在数字化转型中实现稳健发展。
本文基于2011-2022年我国上市商业银行数据，结合长尾理论、信息不对称理论、金融创新理论、金融中介理论、数据经济学理论，实证检验发现数字普惠金融对商业银行信用风险承担具有非线性关系，具体呈现先促进后抑制的“倒U型”关系：初期技术渗透加剧银行间竞争与长尾客户逆向选择，导致风险上行；后期数据积累驱动智能风控迭代与客户画像精细化，风险逐步缓解。在经过工具变量法缓解内生性问题，以及经过改变样本区间和替换核心变量等方法进行稳健性检验后，结论依旧可靠。机制检验表明，数字普惠金融通过净利差与流动性创造双重路径影响商业银行信用风险承担。股权结构异质性方面显示，国有大型商业银行信用风险承担对数字普惠金融发展敏感性更强，具有规模异质性；区域异质性方面显示，东部地区商业银行数字普惠金融发展与商业银行信用风险承担显著为负，而中西部地区商业银行信用风险承担受数字普惠金融发展影响不显著，具有区域异质性。维度异质性方面，数字普惠金融覆盖广度较使用深度及数字化程度在缓解商业银行信用风险承担方面更加显著。
基于以上实证结果，本文提出针对性政策建议。第一，把握动态拐点，实施精准风控；第二，疏通机制堵点，强化多维协同；第三，分类精准施策，激发转型动能；综之，本文的研究能够为推动我国银行业主体健康发展，防范化解金融风险提供现实依据，为强化金融稳定保障体系提供实践参考。

关键词：数字普惠金融 信用风险承担 非线性关系 

Abstract
In recent years, propelled by national policies such as the "14th Five-Year Plan" and the "Fintech Development Plan (2022-2025)", financial technology (Fintech) and the digital transformation of financial institutions have emerged as core strategies for deepening supply-side reform in the financial sector. Commercial banks, leveraging technologies like artificial intelligence and blockchain, are accelerating business model innovation, extending services from traditional offline models towards digital inclusive finance. This shift, facilitated by sophisticated risk control models and scenario-based financial products, has significantly expanded service coverage. However, while enhancing operational efficiency and financial inclusion, this transformation concurrently introduces novel challenges, including intensified credit risk transmission among long-tail customers and heightened interbank competition.
Against this backdrop, this study investigates the impact mechanism of digital inclusive finance (DIF) on commercial banks' credit risk-taking. Through empirical analysis, it examines the resultant effects, aiming to provide a theoretical foundation for balancing financial innovation with risk prevention and control, thereby supporting the banking sector in achieving stable development during its digital transition.
Utilizing panel data from listed commercial banks in China spanning 2011 to 2022, and grounded in theories including the Long Tail theory, Information Asymmetry theory, Financial Innovation theory, Financial Intermediation theory, and Data Economics theory, the empirical findings reveal a non-linear, inverted U-shaped relationship between DIF and banks' credit risk-taking. Specifically, DIF initially exacerbates risk due to intensified competition and adverse selection among long-tail customers during the early stages of technological adoption; subsequently, as data accumulates, enabling iterative improvements in intelligent risk control and refined customer profiling, risk gradually subsides. This conclusion remains robust after addressing potential endogeneity concerns using instrumental variable techniques and undergoing rigorous robustness checks, including altering sample periods and substituting core variables.
Mechanism tests demonstrate that DIF influences commercial banks' credit risk-taking through dual channels: net interest margin (NIM) and liquidity creation (LC). Heterogeneity analysis further indicates: (1) Regarding ownership structure, state-owned large commercial banks exhibit greater sensitivity of credit risk-taking to DIF development, reflecting scale heterogeneity. (2) Regionally, DIF development in eastern China shows a significant negative correlation with bank credit risk-taking, whereas its impact on banks in central and western regions is statistically insignificant, indicating regional heterogeneity. (3) Dimensionally, the coverage breadth of DIF demonstrates a more pronounced effect in mitigating bank credit risk-taking compared to its usage depth and level of digitization.
Based on the empirical findings, this study proposes the following targeted policy recommendations:First, identify dynamic inflection points to implement precision risk control measures.Second, address institutional bottlenecks to strengthen multi-stakeholder collaboration.Third, adopt categorized precision strategies to stimulate transformational momentum.In conclusion, this research provides empirical grounding for promoting the sound development of China's banking sector and mitigating financial risks, while offering practical insights for enhancing the financial stability safeguarding framework.

Keywords：Digital inclusive finance; Credit risk taking;Non-performing loan ratio;