The effects of COVID-19 on Chinese stock markets: an EGARCH approach

Abstract

Coronavirus disease 2019 (COVID-19), the disease caused by the novel coronavirus SARS-CoV-2, has greatly affected financial markets, economies and societies worldwide. This study focusses on the Chinese stock markets. Based on Google Trends data during the period from 1 January 2020 to 12 April 2020, and using the exponential generalised autoregressive conditional heteroskedastic (EGARCH) model, this study finds that the higher uncertainty resulting from the COVID-19 pandemic is significantly associated with the drop in China’s composite index, but this impact varies by sectors. Simultaneously, the higher uncertainty due to COVID-19 is significantly associated with greater volatility in stock returns for both the composite index and sector indices.

Keywords:

• Coronavirus
• COVID-19
• EGARCH
• volatility
• Chinese stock markets
• pandemic

Acknowledgements

The author would like to thank the Executive Editor (Professor Qing He) and an anonymous referee for very timely and critical comments. All errors are the author’s sole responsibility.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Among major economies, China is arguably the first to release its economic data for Q1 2020. Data are retrieved from Wind and National Bureau of Statistics. Wind (https://www.wind.com.cn/en/) is the most widely used Chinese economic and financial data and information provider. It serves more than 90% of the financial firms in the Chinese market, and 75% of qualified foreign institutional investors in China.

2 This conclusion is drawn from the Google Trends data based on the keyword ‘coronavirus’ from China, Australia, Canada, New Zealand, the UK and the US. Regarding China, its peak point is in late January, while for the rest of world, the peak point is around mid-March. Results are not reported in this study, but available upon request.

3 Pinyin is the official romanisation system for Standard Chinese in the mainland of China.

4 Empirical evidence also supports this conclusion. In preliminary analysis the coefficient of Baidu Index is insignificant, while the signs and significance levels of other variables are the same as those based on Google Trends data. These results are not reported in this study, but are available upon request.

5 Based on the Akaike info criterion, a combination of independent variables is employed for all sectors but only a few are reported in Tables 2 and 3.

6 Since EGARCH is a conditional variance model, a negative R-squared is possible.

7 Data are sourced from Wind.

8 It may be argued that using US$40 per barrel as the oil price from 9 March 2020 is more appropriate for sectors other than energy. Further regressions show mixed results, but the performance of D_COVID remains unchanged, which may reflect the complicated nature of oil pricing in China. However, this is not the focus of this paper, and should be the topic of future research.

9 Since EGARCH is a conditional variance model, a negative R-squared is possible.