Both systematic and idiosyncratic information dissemination contribute to assets co-movement. To proxy for co-movement based on heterogeneous information diffusion, we construct two-layer network structures in terms of the correlations of systematic and idiosyncratic returns for Fama–French 49 industry portfolios. We further delve into the co-movement structures by studying their dynamics and interactions in terms of topological properties at both the system and individual levels. Our findings reveal the following: (1) Co-movement structures exhibit temporal changes and are closely associated with major risk events. (2) At the system level, there is time-varying mutual Granger causality between co-movement structures, particularly during or after recession periods. (3) At the individual level, we identify statistically significant but economically insignificant interplay between the two co-movement structures. Moreover, the co-movement structure based on systematic information diffusion is mainly influenced by industry-level characteristics, including idiosyncratic momentum, idiosyncratic volatility, and market beta. These results deepen our understanding of asset co-movement structures and bear significance for asset allocation and risk management practices.

中文翻译：

---

理解基于异构信息关联的共同运动

系统性和特殊性的信息传播都有助于资产联动。为了代表基于异质信息扩散的联动，我们根据 Fama-French 49 个行业投资组合的系统回报和特殊回报的相关性构建了两层网络结构。我们通过研究系统和个体层面的拓扑特性的动力学和相互作用，进一步深入研究联动结构。我们的研究结果揭示了以下内容：（1）联动结构表现出时间变化，并且与重大风险事件密切相关。 （2）在系统层面，联动结构之间存在时变的相互格兰杰因果关系，特别是在经济衰退期间或之后。 (3) 在个人层面，我们发现两种协同运动结构之间存在统计上显着但经济上不显着的相互作用。此外，基于系统信息扩散的联动结构主要受行业层面特征的影响，包括异质动量、异质波动和市场贝塔值。这些结果加深了我们对资产联动结构的理解，对资产配置和风险管理实践具有重要意义。