管理工程学报2026，Vol.40Issue(3)：110-124,15.DOI:10.13587/j.cnki.jieem.2026.03.009
基于投资者在线共同搜索的股票网络和联动效应研究
Stock networks and the comovement effect based on investors' online co-searches
摘要
Abstract
Investors' digital footprints,including their online co-searches of different stocks,have attracted increasing attention in information systems(IS)research because they can reflect investors' behavioral patterns and provide insights into financial market outcomes.Previous studies have primarily examined the impact of investor attention to individual stocks,while group-level phenomena,such as investors' co-attention to stocks,have been insufficiently explored.Moreover,little is known about whether and how co-attention to a set of stocks influences stock market outcomes.This paper examines investors' co-attention relationships by establishing co-search-based networks of stocks."Search clusters"represent investors' close attention to a set of stocks,which can reveal investors' correlated preferences or demands for stocks.This study concerns how co-attention to stocks,as manifested by search clusters,relates to stock comovement(or stock price synchronicity),and attempts to answer the following research questions:1)Do stocks within the same search clusters have significant return comovement?2)How does the diversity of search clusters alter the magnitude of the comovement effect?3)What is the relationship between the level of comovement and co-attention intensity on the search clusters?and 4)How does the comovement pattern change when search clusters evolve dynamically? To address the above questions,this study conducted an empirical analysis of the entire Chinese stock market.Investors' online co-searches of stocks during multiple time periods,including July-December 2020 and April-September 2021,were collected from a popular Chinese financial web portal to establish co-search-based networks of stocks.In the network analysis phase,a cohesive subgroup analysis was conducted to identify the cohesive sub-networks of the original co-search-based networks and extract them as search clusters.These search clusters represent cohesive subgroups within which are reachable stocks,indicating simultaneous investor co-attention.Panel-data models were employed,applying both the market and the four-factor model to test the hypotheses and estimate the effect size of within-cluster comovement. Our findings suggest that the returns of stocks within the same search clusters move in the same direction simultaneously,demonstrating the presence of within-cluster return comovement.Furthermore,the level of comovement is significantly and negatively moderated by the degree of return dispersion of the search clusters,while it is significantly and positively associated with the intensity of investors' co-attention toward search clusters.This indicates that cluster diversity and co-attention intensity can significantly alter the magnitude of return comovement,revealing the contingency conditions of the comovement effect.In addition,the dynamic changes of search clusters and return comovement were investigated by focusing on three different scenarios:stocks joining new search clusters,stocks detaching from existing search clusters,and stocks transferring to different search clusters.The results show changes in comovement as stocks enter or detach from search clusters or switch from one cluster to another,which reveals the dynamic patterns of the comovement effect.The instrumental variable method was used to address the potential endogeneity issue,while numerous robustness checks were conducted to confirm the consistency of our findings. This research contributes to the IS and finance literature by highlighting the economic impact of investors' digital footprints on financial markets.By focusing on investors' co-attention to stocks and untangling its impact on stock comovement,this study provides a nuanced understanding of the effect of comovement induced by co-attention,which also extends attention theory documented in previous research.In terms of practical applications,the research findings highlight the value of infomediaries and provide important guidance for investors in portfolio management and investment decision-making.关键词
共同搜索/联动效应/社会网络分析/凝聚子群Key words
Co-search/Comovement effect/Social network analysis/Cohesive subgroup分类
管理科学引用本文复制引用
上官武悦,陈熹,刘彦初,梁仲文..基于投资者在线共同搜索的股票网络和联动效应研究[J].管理工程学报,2026,40(3):110-124,15.基金项目
国家自然科学基金项目(72202193、71991474) （72202193、71991474）
中央高校基本科研业务费专项资金(ZK1164) The National Natural Science Foundation of China(72202193,71991474) （ZK1164）
The Fundamental Research Funds for the Central Universities(ZK1164) （ZK1164）
