ICentral bank communications are gradually influencing corporate finances and sustainability in recent years, but their impact on ESG performance in emerging markets has not been studied. Python (Google Colab) is used to conduct simulation-based quantitative analysis in this paper to examine how communication signals affect the ESG performance of A-share listed companies. XGBoost, SVM, LightGBM and Random Forest were advanced machine learning models applied to synthetic firm-level panel data containing tone, uncertainty and surprise indicators. The regression model reliably predicted the change in continuous ESG data with a relatively small error, but the classification model was unable to identify infrequent upgrade events due to severe class imbalance. Central Bank signals affect the distribution of ESG outcomes rather than inducing frequent, step-up improvements. Recommend a redistribution plan and set an upper bound for event detection. The research has given some theoretical support for signaling theories, provided practical references for policy design, and also pointed out deficiencies in the simulated data and event definitions.
