Liu Xueyong, Jiang Cheng
School of Management and Engineering, Capital University of Economics and Business, Beijing 100070, China.
Chaos. 2020 Mar;30(3):033120. doi: 10.1063/1.5131066.
The objective of this study is to examine the multi-scale feature of volatility spillover in the energy stock market systematically. To achieve this objective, a framework is proposed. First, the wavelet theory is used to divide the original data to subsequences to analyze the multi-scale features, and then the Generalized Autoregressive Conditional Heteroskedasticity model with Baba, Engle, Kraft, and Kroner specification (GARCH-BEKK) and the complex network theory are used to construct the spillover networks. Finally, the stock prices in the energy sector of China from 2014 to 2016 are used to conduct experiments. The main contribution of this paper is that we find various features of volatility spillover transmission in different time scales among energy stock prices. The results indicate that the volatility spillover effects are more fragmented in the short term, while the volatility changes will be only transmitted by a small number of important stock prices in the long term. In addition, we captured the key paths of volatility transmission by using the smallest directed tree of network under different timescales.
本研究的目的是系统地考察能源股票市场波动溢出的多尺度特征。为实现这一目标，提出了一个框架。首先，利用小波理论将原始数据划分为子序列以分析多尺度特征，然后使用具有 Baba、Engle、Kraft 和 Kroner 规范的广义自回归条件异方差模型（GARCH-BEKK）和复杂网络理论构建溢出网络。最后，使用2014年至2016年中国能源行业的股票价格进行实验。本文的主要贡献在于，我们发现了能源股票价格在不同时间尺度上波动溢出传递的各种特征。结果表明，波动溢出效应在短期内更为分散，而长期来看，波动变化仅通过少数重要股票价格进行传递。此外，我们通过使用不同时间尺度下网络的最小有向树捕获了波动传递的关键路径。
