Stock habitats and information flow: How do different co-attention behaviors in online communities shape market reactions?
研究了在线投资社区中投资者和内容贡献者的共同关注行为如何形成股票集群（栖息地），发现投资者共同关注行为比内容贡献者更能预测股票收益，并揭示了信息在不同栖息地间的流动方向。
Investors increasingly use online investment communities to acquire financial market information before making trading decisions to reduce the cost of information acquisition and get more abundant content. Due to limited attention, investors tend to focus their trading only on a subset of assets that align with their personal investment preferences. Thus, the attention behavior of investors in the communities can reflect their focus trends and indicate future stock movements. Unlike previous research that mainly focused on investor common search and viewing behaviors, we constructed stock clusters based on different common attention behaviors data (i.e., common follow behavior by investors and common mention behavior by content contributors) and compared their predictive capabilities on stock returns. After controlling for some deterministic factors, we verified the existence of comovement among stocks within the clusters (i.e., stock habitats) and found that investors' common attention behaviors can better predict stock returns compared to content contributors. To explore the mechanism, we found a possible direction of information flow between different stock habitats and revealed the leading role of content contributors in online investment communities. This study enriches the literature on stock habitats and information diffusion in online investment communities and provides practical decision support on portfolio management for investors. Moreover, online platform managers can also use our conclusions to provide better decision-making assistance for market participants.