Dynamic energy stock selection based on shareholders’ coholding network

In stock market investment, the first step is to select stocks. To do so, stock traders have depended on two major tools: fundamental analysis and technical analysis, while ignoring the impact of large shareholder behavior on stocks. The major shareholders in the Chinese stock market show a serious herd effect; they can control or influence the stock price of the listed companies by increasing or reducing their holdings, which may affect the stability of the market. Thus, it is significant to study shareholders’ holding behavior. In this paper, we constructed a shareholder coholding network of listed energy companies in the Chinese stock market based on quarterly shareholder data from 2012 to 2018. We selected stocks by choosing central and peripheral topological indicators in the training period, namely, degree (D), betweenness centrality (BC), closeness centrality (CC) and eigenvector centrality (EC). We also used other traditional trading strategies to select stocks. Then, we used the selected stocks to build the portfolio and measure the performance in the test period. We identified the market conditions in different test periods and evaluated the performance of the portfolios in different market conditions. The results show that the topological indicators of a shareholder coholding network can be used to guide actual investment, and the suggestions for investment in different market environments are presented in the conclusion.