The Volatility of the Stock Market and Financial Cycle: GARCH Family Models

Abstract

The paper examines the association between financial market volatility and actual economic incidents. We specifically analyze the statistical characteristics of the stock price series and its association with the financial cycle. Using 20 years of Vietnamese main stock VNIndex daily data from 2 August 2000 to 31 December 2020, we select the most adequate Generalized Autoregressive Conditional Heteroskedasticity (GARCH) family models and corresponding distribution rules. The paper initially assesses several types of GARCH models’ criteria, namely the log-likelihood, AIC and BIC, in choosing the best model to illustrate the financial cycle. We further use three different distribution rules, namely the normal distribution rule, the Student-t statistic distribution, and the Generalized Error Distribution (GED), in selecting the best GARCH model. The results show that Exponential Generalized Autoregressive Conditional Heteroscedastic (EGARCH) with student-t statistic distribution seems the best suited to demonstrate the stock price and its return volatility. It also suits the marginal distribution of the financial cycle. Our study further validates the lead time and volatility between the selected model results and the significant financial events using the turning point and Bull-Bear application (Lunde and Timmermann, 2004). Although the recommended model has shown no evidence as an effective forecast tool for the financial cycle in long run, this study paves the way for extensive research in the future.