Evaluating influential nodes for the Chinese energy stocks based on jump volatility spillover network

Abstract

We employ a complex network approach to dig out the influential Chinese energy stocks in this paper. We first use the 5-min high-frequency data of the Chinese energy stocks over the period of 2013–2018 to build a static jump volatility spillover network. Then a novel method of entropy weight TOPSIS (Technique for Order Preference by Similarities to Ideal Solution) is proposed to evaluate the influential nodes. Furthermore, we construct dynamic networks with the help of time-varying Granger causality test. Empirical analyses show that: (1) Combining static network and the proposed entropy weight TOPSIS scores, we find that China Petroleum Engineering & Construction Corp, Zhengzhou Coal Industry & Electric Power Co.,Ltd., Shenzhen Guangju Energy Co.,Ltd., China Coal Energy Company Limited and Shaanxi Provincial Natural Gas Co.,Ltd. are influential energy stocks. (2) The advantage of entropy weight TOPSIS lies in the fact that it has the highest correlation coefficient between its score and jump volatility is the highest, comparing with the traditional methods such as equal weight, TOPSIS, analytic hierarchy process and principal component analysis. (3) Particularly, by making full use of dynamic network analysis, a very interesting finding in this paper indicates that the network density also provides an “early warning” for possible upcoming crises. (4) In addition, a very interesting fact in point is that most of the stocks with larger market capitalization usually have high-level influence during Chinese stock market crash; such smallcapitalization energy stocks with high scores are however particularly crucial for investors and regulatory authorities to grasp the risk characteristic. The results can provide us some light for finding out those influential energy stocks whose volatilities may cause many other stocks in the energy industry to rise and fall.