Analyzing Dynamic Regimes of GARCH Model on Stock Price Volatility

Abstract

Listen

Translate article

As a result of volatility dynamics, investors and other stakeholders in businesses and industries have difficulty making financial decisions. Generalized Autoregressive Conditional Heteroskedasticity (GARCH) models are the most widely applied in the analysis of financial derivatives volatility. Volatility persistence is a common issue when analyzing stock prices, making it cumbersome for GARCH models. The GARCH model is transformed into the Makov switching GARCH model to check for dynamics in volatility persistence. Markov Regime-Switching GARCH (MSGARCH) models permit the conditional mean and variance to change across regimes over time. The Markov switching GARCH models incorporate the regime variables in the parameter space, making it viable for the parameters to be estimated by the maximum likelihood estimation method, unlike the classical GARCH models. Zenith Bank plc’s daily closing stock prices, a top-tier stock on the Nigerian Stock Exchange market, are fitted using the GARCH and MSGARCH models. The comparison between the MSGARCH model and the classical GARCH model was verified using the AIC and BIC metrics as well as the one with the maximum log likelihood estimates. The outcome suggests that MSGARCH model performs better than the single-regime GARCH model and that it yields significantly better out of-sample volatility forecasts. The results will aid the stakeholders to leverage and mitigate risks in their investment on the selected stocks.