This study examines whether central bank communication is associated with corporate financing constraints in China, with particular attention to whether policy communication is reflected in firm-level financing frictions rather than only in broader credit expansion or aggregate expectations.
Using quarterly communication texts issued by the People's Bank of China and data on Chinese A-share listed firms from the first quarter of 2007 to the third quarter of 2025, this study constructs a central bank communication sentiment index and estimates its association with corporate financing constraints within a partially linear double machine learning (DML) framework. The empirical analysis recognizes that the communication measure varies at the year-quarter level and therefore supplements the baseline estimates with year-quarter clustered and firm-year-quarter two-way clustered inference.
The baseline DML estimates show that more positive central bank communication is negatively associated with corporate financing constraints. This relation remains negative under stricter clustered inference. Marginal-effect plots further show that the interaction specifications should be interpreted through conditional marginal effects over the observed ranges of the moderators rather than through the coefficient on Comm alone.
This study advances the literature by focusing on financing constraints as a firm-level corporate finance friction and by showing when communication signals become credible and actionable in China's bank-based financial system. The contribution is not that DML is new in itself, but that it helps discipline estimation under high-dimensional controls while the substantive contribution lies in the financing-friction outcome, communication credibility and institutional boundary conditions.
