机构投资者抱团与股价崩盘风险

Institutional Investor Clique and Stock Price Crash Risk

自2008年全球股市崩盘乃至2015年下半年中国发生股灾以来,股价崩盘机制及其防范已经成为学术界研究的热点。本文从社会网络算法出发,研究网络中机构投资者抱团行为如何影响股价崩盘风险。本文运用Louvain算法从机构投资者网络中提取出机构投资者团体,研究发现机构投资者抱团持股比例及持股比重与股价崩盘风险呈显著正相关关系。相对于其他企业,在非国有、代理成本较高以及处在投资者保护水平较高地区的企业中,机构投资者抱团对公司股价崩盘风险的加剧作用更为显著。为控制内生性问题,本文采用2004年保险资金被批准直接入市的政策变化作为外部冲击构造工具变量和自然实验等进行稳健性检验,研究结论依然成立。进一步通过信息、公司治理、交易等多个角度去检验潜在影响机制。与投资者独立同质的假设不同,本文创新性地探索了中国机构投资者在网络中的交互行为,拓展和深化了机构投资者对公司治理、股价崩盘风险影响的研究,并对网络关系和金融研究的交叉融合提供了一定的理论与实证依据。

Since the global stock market crash triggered by the global financial crisis in 2008 and the stock market crash in China in the second half of 2015,the mechanism of stock price crash and its prevention have become hot topics in academic research.From social network algorithm,this paper studies how the collective behavior of institutional investors in the network affects the stock price crash risk.This paper adopts Louvain algorithm to extract institutional investors clique measure from institutional investors’network to investigate its impact on stock price crash risk.We find that there is a significant positive correlation between the proportion of shares held by institutional investors clique and stock price crash risk afterwards.Compared with other enterprises,institutional investors clique plays a more significant role in exacerbating the risk of stock price crash in non-state-owned enterprises,those with higher agent costs and those with higher levels of investor protection.In order to ease the endogenous concern,this paper introduces an external shock of the policy change which first allows insurance funds to invest in stock market.We construct an instrumental variable and a difference-in-difference test to show the robustness of our results.Different from the assumption that investors are independent and homogeneous,this paper innovatively explores the interaction behavior of Chinese institutional investors in the network,and deepens the research about its impact on corporate governance and stock price crash risk.Our findings lend both theoretical and empirical support to future cross-discipline research between network and finance.