GM-GARCH MODEL: APPLICATION ON THE CROATIAN STOCK MARKET

This research deals with an empirical comparison of the forecasting abilities of standard GARCH (Generalized AutoRegressive Conditional Heteroskedasticity) and GM-GARCH (Grey Model) models. Although the standard GARCH models are probably the most popular approach in stock risk modelling and forecasting, there have been many different extensions over the years of those models. One possible direction is the Grey Systems Theory, which focuses on uncertain and grey data. Thus, it is very suitable for stock market applications. Based on daily data for the period 4 January 2017 - 5 May 2020 for the stock market index CROBEX, GARCH and GM-GARCH models are estimated and compared based on out-of-sample forecast capabilities. The results indicate that the GM based model is superior compared to its counterpart. That is why it is advisable to incorporate the models from Grey Systems Theory into the investment strategies which focus on hedging or minimising the portfolio risk.