基金网络能够提高投资绩效吗?

Can Fund Networks Improve Investment Performance?

基金网络在金融市场的信息流动中发挥着重要作用。本文利用基金共同持股关系构建了一个有效的基金网络数据集,以中国资本市场股票型基金2005-2018年季度数据为研究样本,考察基金网络是否以及如何影响投资绩效。结果表明,基金在基金网络中越处于网络中心地位,基金的投资绩效越高。使用基金家族网络作为工具变量缓解内生性偏误后,基金网络仍然对投资绩效具有显著的正向影响。进一步地,本文考察了基金网络影响投资绩效的渠道,结果表明,基金网络主要通过提高基金的选股技能、资产配置技能和管理技能影响投资绩效。最后,本文考察了基金网络对基金市场份额的影响,研究发现基金网络会显著提高基金的市场份额,对基金在市场上的占有率有积极的正向影响。

Institutional investors managing public funds are rapidly becoming a crucial part of financial markets. Accordingly, there is widespread interest in how these fund managers make investment decisions. Recent studies show that due to the complementarity of information structures, social networks play an important role in fund managers’ asset allocation and diversification decisions. We focus on how an important social network-enabled interaction between funds-the fund holdings network-affects the investment performance of funds.There are different theoretical perspectives on how a fund’s co-ownership network affects investment performance. On the one hand, such networks produce information diffusion effects, and fund managers at the center of a fund network will receive signals earlier and obtain more valuable information than fund managers at the periphery of the network. Furthermore, the information held by fund managers at the center of the network will be more accurate, which will have a positive impact on the investment performance of their funds. On the other hand, a fund network can induce the "free riding on friends" effect. Compared with fund managers at the periphery of the network, fund managers at the center of the network support more "free riders". Under this condition, the fund network may have a negative impact on their fund performance.We use data on China’s capital market stock funds from January 2005 to September 2018 to examine the impact of fund networks on investment performance. The results show that the centrality of a fund’s position in a fund network has a significant positive impact on its investment performance. The results have obvious economic significance: we find that an increase of one standard deviation in the degree, closeness, or eigenvector measures of a fund’s position in the network increases investment performance by an average of 54.43%, 51.15%, and 50.93% respectively.We also examine the channels through which network position affects performance. Specially, we examine fund stock selection skills, asset allocation skills, and fund management skills. First, institutional investors located at the center of a network receive rich, accurate, and timely information about stock pricing. This information can promote fund managers’ analyses and mastery of stock fundamentals, thereby improving their stock selection skills. Second, the transmission of information between funds is an important factor that affects managers’ asset allocation strategy. Fund managers located at the center of a network have more accurate information and can use their information to allocate assets independently and effectively, that is, they have improved asset allocation skills. Third, social networks can promote the dissemination of information, thereby improving managers’ fund management skills.Finally, we examine the influence of fund networks on fund shares, which we define as a fund’s market share. A fund network provides channels for the dissemination and exchange of information. Fund managers at the center of a network receive faster and more valuable information, which allows them to implement high-quality portfolio management and product differentiation strategies. High-quality portfolio management and product differentiation are effective strategies for gaining market share. This study shows that fund networks have a significant positive impact on funds’ market share.This research makes two major contributions. First, we contribute to the literature on the interactions between institutional investors’ shareholdings. Most studies have assumed that institutional investors are homogeneous, and there is a lack of research on the heterogeneity of institutional investors and their interactions. We expand this field by constructing a model of a dynamic co-ownership fund network. The fund network allows for the exchange of information between co-ownership funds, reflects the interactions between institutional investors, and distinguishes institutional investors’ holdings on the basis of their network location. Using a series of centrality measurement methods, we examine the effects of network location on a fund’s ability to obtain information and take actions.Second, we contribute to the research on the impact of social relationships on investment performance by examining the impact of social networks on performance from the perspective of dynamic institutional investor interactions. We build a model of fund networks to illustrate dynamic institutional investor shareholding interaction, and examine its impact on fund investment performance. We find that in a fund network based on co-ownership, positive information diffusion effects dominate, which leads to better investment performance for funds with more relationships. In addition, the social interaction between funds in a network improve the managers’ stock selection skills, asset allocation skills, and management skills, which ultimately improve the funds’ investment performance.