| Along with the increase of energy’s financial attribute,energy stock market has become a key part for the energy development.As the main participant in the energy stock market,shareholders have irrational investment behavior,which is the main cause for their information interaction and the stock fluctuation.Common-holding relations is a type of shareholders’ relations derived directly from their investment behavior on listed companies,providing a structural perspective to describe information diffusion tunnels.As a result,in recent years,studies on the shareholders’ common-holding relations become hot issue,which are essential for risk estimation and control.Because of the relative significance of herding effect in China stock market,this paper takes China energy stock market as the case.2006-2015 data of China energy related listed companies and their top 10 circulative shareholders are taken as empirical data.The scientific issue is to answer the basic structural units and the evolution mechanism of co-holding relations among the shareholders in China energy stock market.Complex network is applied to model the co-holding networks,Motifs considering multi-attributes for agents are defined,motif network is contributed,and the dynamic link prediction simulation model is set.We explore basic units,structural relations among units,and the evolution mechanism in the co-holding network respectively.Main works and innovative findings are as follows:(1)Motifs with multi types of shareholders and co-holding relations are proposed for exploring basic units and its structural rule in the shareholders’ co-holding network.Shareholders are divided into three types: companies,funds and individuals,while co-holding relations have strong and weak two types.With the combination of these features,improved motifs are used to explore shareholders’ basic common-holding structure,their involvement preference in basic units,units’ significance,and units’ evolution characteristics.We find that,there are regular and preferential unit structures among shareholders.Shareholders tend to keep steady information interaction tunnels,but the need for differences of is promotional for the formation of common-holding relations.Because of clear and investment target,companies and funds have limit but strong common-holding relations with each other.On the contrary,individuals keep much denser tunnels with other shareholders due to their limit information sources.(2)Motif network derived from shareholders’ co-holding network is established for exploring structural relations among basic units to reveal their roles in the common-holding networks.Motif network takes motifs as nodes,and whether they have same information diffusion tunnels as edges.The comparison of degree distribution with other typical networks reveals its Poisson distribution characteristics.Clustering coefficient and the closeness coefficient approves motifs’ roles in networks.We find that,the differences among shareholders promote the density of information diffusion tunnels.Local pattern with multi companies bridges the information spreading.As a result,regulators need to pay special attention to these patterns when any risk occur in energy stock market.Cutting these pattern is good for stopping information spreading.(3)Dynamic link prediction simulation model with motif information is proposed for exploring the evolution mechanism of co-holding networks.The simulation model is set based on the characteristics of identified motifs and their evolution trend.It is able to describe the dynamic evolving process of networks.The feasibility of the simulation model is proved,especially its ability of estimating network scale.Based on the simulation model,we further make projections for the shareholders’ co-holding relations in the China energy stock market.It shows that the spreading paths among shareholders would be longer,so the interactions would be weaker.Above all,this paper provides suggestions about the stock market regulation and the risk control for the China energy financial market,including the selection of regulation target,key patterns for the information diffusion,and the estimation for the future co-holding patterns.Theoretically,the work identifies motifs with shareholders’ financial characteristics,and contributes motif network and the dynamic link prediction simulation model.Empirically,this research answers the basic structural units and the evolution mechanism of co-holding links among shareholders in China energy stock market.The distribution and evolution trend of information tunnels can be estimated,and the selection of market regulating target and the pattern for risk control can be suggested. |
