A1104
Title: Inference on multiplicative component GARCH without any small-order moment
Authors: Christian Francq - CREST and University Lille III (France)
Baye Matar Kandji - CREST, Institut Polytechnique de Paris (France) [presenting]
Jean-Michel Zakoian - CREST (France)
Abstract: The aim is to investigate the existence of strictly stationary solutions and the asymptotic properties of Quasi-Maximum Likelihood (QML) estimation for a class of multiplicative two-component (short-term and long-run volatilities) GARCH models. We show that the strict stationarity condition is compatible with the infiniteness of any small-order power moment, contrary to the classical GARCH setting. The strong consistency and asymptotic normality of the QML estimator are established under mild conditions. The results are illustrated via Monte Carlo experiments and real financial data.
