From emotion to action: How intraday investor sentiment drives market microstructure

This study introduces the concept of “investor sentiment slices” to capture distinct emotional patterns within a single trading day. Using 4.41 million social media posts and natural language processing, we construct high-frequency sentiment indices representing four emotional states: adaptation, anchoring, process, and dormant sentiment. Regression analyses reveal that anchoring and process sentiments drive stock price movements by signaling market expectations, while adaptation and dormant sentiments reflect emotional buffering and accumulation. Importantly, the effects of these sentiment slices vary with market conditions: in pessimistic markets, investor behavior is dominated by emotionally anchored judgments, whereas in recovering markets, process sentiment becomes more influential, supporting rational integration of information. This temporally structured framework connects psychological theory with behavioral finance, showing how intraday sentiment dynamics affect trading and price formation. Our findings offer new insights into the sequential, condition-dependent nature of investor emotions and their role in shaping market behavior throughout the trading day.