ARCH modelling framework of Engle (1982) and its GARCH generalization of Bollerslev (1986) gave a huge impetus to econometric model building in the field of financial time series with time-varying variance. The main idea of the models was to describe the most typical features of capital markets like volatility clustering, excess kurtosis and fat tails. As empirical evidence shows asymmetry is also a prominent feature of stock market returns volatility. The reaction of risk if stock returns go off the long run trajectory is different in case of positive and negative market news. Thus it is indispensable to employ asymmetric models being a modification of a traditional GARCH. In the paper we used an approach of Engle and Ng (1993) to test for asymmetric effects in stock indices of developed and Central European stock markets.