Abstract: After more than three decades of development, China’s stock market has rapidly grown into a multi-level stock market that includes the main board market, the SME board market and the GEM board market, ranking the second in the world in terms of market size, second only to that in the United States. The stock market, a socio-economic barometer, was the first hit by the sudden outbreak of Newcastle Pneumonia. The U.S. stock market set an all-time record of four meltdowns in nine days in March 2020, and international stock markets also experienced huge shocks as a result.In China, due to the need to deal with the rapid spread of the epidemic, the country adopted a strict quarantine policy against the epidemic. The uncertainty of supply and demand as well as the lack of awareness of the epidemic exacerbated market panic, and this sentiment was quickly transmitted to the stock market, exacerbating abnormal market volatility, making it particularly important to measure the structural characteristics and risk profile of the stock market.
Previous research on stock market risk has been more based on the Efficient Market Hypothesis (EMH) and the empirical analysis using simple linear analysis tools. As research continues to progress, anomalies in the financial markets emerge and EMH is constantly challenged because it cannot give a reasonable explanation. As a nonlinear and complex dynamic system, a good performance of multiple fractal analysis method based on Fractal Market Theory(FMT) in analyzing nonlinear non-Gaussian distribution series will provide a more effective metric tool to better measure the changes in market structural characteristics. Therefore, this study uses fractal analysis tools to analyze the structural characteristics of the stock market affected by the epidemic metric, giving new evidence on market structural characteristics and risk measures from a fractal perspective.
With intraday 5-minute trading data of the main board, SME and GEM indices: SSE, SZSE, CSI300, SME, and ChiNext are selected from the WIND database. Multifractal detrended fluctuation analysis (MF-DFA) is used to explore the impact of this new coronavirus pneumonia outbreak on our multi-level stock market. Moreover, the advantage of this method over other methods is that it can find long-range correlations of non-stationary time series and avoid misjudgment of correlations. The sliding window tool is also used to provide a detailed portrayal of the dynamic changes in market fractal characteristics and risk levels.
The results of the study indicate that, firstly, the multifractal characteristics of China’s multi-level stock market is significantly enhanced during the epidemic, the complexity and risk intensity of the market significantly increase, and the effectiveness of the market severely weakens; in the later period when the epidemic is effectively controlled, the risk intensity is significantly reduced and the market efficiency is gradually restored. Secondly, during the epidemic period, the market fractal intensity and riskiness of ChiNext are lower and more effective than other sector indices, and the overall market performance is better.In addition, the fractal characteristics of each sector representative index have time-varying multifractality throughout the sample period. Finally, the combination of long memory of index return series with non-normal thick-tailed distribution makes China’s stock market exhibit strong multifractal characteristics during the epidemic period. The findings of this paper reflect, to a certain extent, the reliability of market risk intensity indicators based on multiple fractal analysis tools during the epidemic, and the findings are consistent with the market performance. Therefore, exploring the establishment of a capital market risk measurement and crisis warning mechanism based on FMT will provide investors and regulators with some references and suggestions for crisis warning and risk management.

Key words: COVID-19, multifractals, market risk

摘要： 本文基于多重分形消除趋势波动分析工具(MF-DFA)研究了新冠肺炎疫情对我国股市造成的冲击和影响。研究结果表明:首先,疫情期间,我国多层次股票市场的多重分形结构特征明显增强,市场的复杂程度和风险强度显著提升,市场的有效性也受到严重的削弱;疫情得到有效控制后,风险强度明显下降,市场效率逐步恢复;其次,疫情期间创业板指数较其他板块指数的市场分形强度和风险性低,有效性高,市场整体表现较好, 且在整个样本期内,各个板块中的代表性指数均具有时变的多重分形特征;最后,指数收益序列的长记忆性和非正态厚尾分布的共同作用使我国股票市场在疫情期间呈现出较强的多重分形特征。

关键词: 新冠肺炎疫情, 多重分形, 市场风险