EventsThe 1st International Online Conference on Risk and Financial Management
Published
This submission belongs to the session S2. AI in Economics and Finance of the event The 1st International Online Conference on Risk and Financial Management
Published date
13 Jun, 2025
Academic Editor
author-avatarSvetlozar Rachev
Citation
Keyao WANG, AI-Driven Policy Effects on Stock Market Anomalies: Evidence from China's Digital Finance Era, in Proceedings of The 1st International Online Conference on Risk and Financial Management, 17 June–18 June 2025, MDPI: Basel, Switzerland
Share
Email
Facebook
Twitter
LinkedIn

AI-Driven Policy Effects on Stock Market Anomalies: Evidence from China's Digital Finance Era

Keyao WANG 1
1. The Hong Kong University of Science and Technology, CITIC Bank, Beijing, 100023, China, China
Abstract

This study investigates the linkage between policy events and abrupt stock market fluctuations in China during 2024, analyzing how regulatory agencies—including the People’s Bank of China (PBOC), China Securities Regulatory Commission (CSRC), and China Banking Regulatory Commission (CBRC)—formulate pre-emptive policies to mitigate sudden market volatility, prevent asset bubbles, and curb systemic financial risks. Integrating behavioral finance theory, monetary economics, emergency event theory, and monetary policy frameworks, we redefine "emergency events" within China’s institutional context and conduct a micro-level analysis using event study methodology supplemented by a Principal Component BP Neural Network (PC-BPNN) algorithm. Focusing on the Shanghai Composite Index (SCI) as a market proxy, we address three core questions: (1) whether China’s monetary policy exerts macro-level intervention effects on stock markets; (2) whether PC-BPNN outperforms existing models in predicting stock prices and deriving normal returns; and (3) whether monetary policy retained significant influence amid frequent 2024 market emergencies.

Methodological innovations include redefining stock market emergencies, applying PC-BPNN for price prediction, and evaluating policy efficacy through event studies. Our key findings reveal the following:

(1)Monetary Policy Effectiveness: China’s monetary tools demonstrate measurable macro-level market intervention capabilities, validating their role in market regulation.

(2)PC-BPNN Superiority: The PC-BPNN model achieves higher accuracy in price forecasting compared to traditional methods, establishing its utility for subsequent research.

(3)Policy Attenuation Mechanism: Frequent abnormal market declines in 2015 nullified monetary policy’s significance (p > 0.05), exposing a self-reinforcing vicious cycle: investor pessimism and distrust reduced policy responsiveness, exacerbating sell-offs and liquidity drain. Concurrently, emergency events amplified negative sentiment, weakening policy transmission and undermining regulatory control, a dynamic that intensified bubble risks while rendering stabilization measures ineffective.

Keywords
Algorithmic governance
policy credibility
digital financial stability
multimodal AI
China’s capital markets
Navigating international stock markets using nonlinear quantitative investing methods
Binance USD Delisting and Stablecoin Repercussions: A Local Projections Approach