Probabilistic properties of periodic GARCH prosses (Q1009536): Difference between revisions

This paper deals with the probabilistic structure of a GARCH-type stochastic difference equation with periodically time-varying parameters. First the model is written as a stochastic recurrence equation, with independent and periodically distributed coefficients. A necessary and sufficient condition for the PGARCH equation to possess a non-anticipative strictly periodically solution in terms of the Lyapunov exponent for i.p.d. matrices is given. A corollary gives a necessary condition for strict periodic stationarity. A direct consequence is that the moment of some positive order of the periodic stationary solution is finite. In turn this property is exploited in order to establish strong consistency and asymptotic normality of quasi-maximum likelihood estimators for PGARCH models. Some further results on geometric ergodicity are also proven.