A GARCH Framework Analysis of COVID-19 Impacts on SMEs Using Chinese GEM Index

ABSTRACT

Stock market return analysis and forecasting are an important topic in econometric finance research. Since the traditional ARIMA models do not consider the variation of volatility, their prediction accuracy is not satisfactory to represent highly volatile periods of any stock market. The GARCH model family resolves the heteroskedasticity of a time series, and hence, it performs better in periods of high volatility. This paper explores the impact of the COVID-19 epidemic on Chinese small- and medium-sized enterprises (SMEs) using a GARCH model for Business as usual (BAU) simulation. We use the Chinese Growth Enterprise Market (GEM) stock index to represent the economic situation of SMEs during the COVID-19 period. Then, we extract, analyze, and predict changes in GEM stock volatility, explore the impact on and recovery status of SMEs, and predict their future trends. For BAU simulation, we first preprocess the GEM stock index between 2018 and 2020 and determine the order of autocorrelation and lags of the data to build the mean model. An ARCH effect test on the residual term of the mean equation was found to be significant and help to decide the order of the GARCH framework. Using the model, a BAU simulation was created and compared statistically with the actual GEM index during 2020. The comparison successfully demonstrated that the GEM index has increased volatility during the pandemic, which is in line with our prior hypothesis.