Volatility regimes of selected central European stock returns : a Markov switching GARCH approach

This paper investigates the weekly stock market data of the Hungarian stock index BUX, the Czech stock index PX and the Polish stock index WIG20 spanning from January 7, 2001 to April 18, 2021. The period of more than 20 years enabled to analyse the behaviour of returns and their volatility during both the calm as well as various crises/turmoil periods. Besides the traditional GARCH-type models (GARCH and GJR-GARCH) the two-regime Markov Switching GARCHtype models (MS-GARCH and MS-GJR-GARCH) were estimated in order to examine the volatility switches of the Central European transition stock markets. The t-distribution of error terms was used to capture the dynamics of analysed returns more precisely. The results proved high volatility persistence of individual markets which substantially differed across the both regimes. Furthermore, the GJR-GARCH and MS-GJR-GARCH models clearly confirmed the presence of the leverage effect. Consideration of the MS-GARCH-type models enabled to capture various volatility switches during the analysed period attributable mainly to the global financial crisis 2008-2009, to European debt crisis in 2011 and to the Covid-19 pandemic in 2020. Interesting results were received for the Czech market with the strong leverage effect indicating completely different specification of volatility regimes by the MS-GJR-GARCH model.