Abstract:

In this paper the properties of t-Student significance test in time series modeling are discussed. In particular GARCH models are considered. Their special property is the volatility clustering effect. Apart from the classic GARCH process, tests were also car-ried out in the case of processes characterized by asymmetry. The paper contains t-test simulation result for the following processes: EGARCH, TGARCH, GJR-GARCH and APARCH. The motivation for using asymmetric models is the universality of the lever-age effect, i.e. a situation in which volatility tends to increase dramatically following bad news, and to increase moderately (or even to diminish) following good news.