关键词: 行业关联, 方差分解, 信息溢出网络, 系统性重要行业

Abstract: Information spillover network is a channel of risk connectedness among different industries. The direction and strength of information spillover are closely related to the industries’ risk contagion characteristics. Using the generalized variance decomposition to establish industry return spillover networks and industry volatility spillover networks for the industry-level index of Shenyin & Wanguo Industries, this paper analyzes the total connectedness and dynamic evolution of China’s industry information spillovers from static and dynamic perspectives. The results show that the information spillover among Chinese industries is higher and stronger connectedness. However, the information spillover fluctuates and shows uncertainty with time. In the long run, it is highly consistent for return spillover networks and volatility spillover networks to identify and rank the systemically important industries. Nevertheless, in the short run there is a difference between them. Specifically, in the long term, bank and non-bank financial industry are the systemically important (information) accepted industries and the machinery and equipment industry is the systemically important (information) transmitted industries. In the short term, banks, non-bank finance, defense industry, food and beverage and household electrical appliances are systemically important industries, but their specific roles (accepted or transmitted information) are uncertain. Our findings have important implications for government industrial policy-makers and industrial regulators.

Key words: industry connectedness, variance decomposition, information spillover network, systemically important industries