Amid growing pressures on China’s real economy, the government has adopted financial concession policies to encourage banks to reduce fees, extend preferential loans, and support the real sector. Using panel data from 115 Chinese commercial banks between 2016 and 2022, this paper investigates the impact of such policies, with loan loss provisioning (LLP) serving as the main indicator of banks’ concessionary responses. It introduces a novel classification framework to distinguish three types of behavior: (1) Type I pseudo-concessions—excessive LLP to suppress profits; (2) Type II pseudo-concessions—reduced LLP to inflate earnings; (3) substantive concessions—LLP reductions used to ease capital constraints and expand credit. Using content-based analysis of annual reports and a generalized DID approach, this paper empirically tests whether and how banks adjust LLP in response to policy signals.

The results show that financial concession policies significantly lower LLP, alleviate capital constraints, and reduce the procyclicality of credit supply, thereby enhancing financing for the real economy. No evidence has been found for either type of pseudo-concessions, suggesting that banks undertake substantive profit-concessions. Further subgroup regression results indicate that substantive profit-concession behavior is more pronounced among banks located in regions with greater fiscal pressure, operating in more competitive markets, or exhibiting lower risk tolerance. This suggests that banks are indeed actively responding to policy guidance, driven by multiple external environmental factors.

This paper contributes to the literature in three ways: First, it transcends the cognitive boundaries of the traditional theory of bank avoidance behavior, revealing that under the guidance of flexible policies, banks may proactively align with national policy objectives rather than merely seeking to circumvent regulatory requirements. Second, it constructs a more interpretable and operational framework to identify profit-concession behavior and systematically distinguishes three types of profit-concession behavior, expanding the methodological framework of micro-level banking research. Third, it employs the frequency of policy-related terms in banks’ annual reports as a proxy for their level of policy exposure, which better aligns with the characteristics of soft policy environments and offers stronger explanatory power and greater relevance for policy implications.