A1080
Title: Finite moments testing in a general class of nonlinear time series models
Authors: Jean-Michel Zakoian - CREST (France) [presenting]
Christian Francq - University of Lille and CREST (France)
Abstract: The problem of testing the finiteness of moments for a class of semi-parametric time series encompassing many commonly used specifications is investigated. The existence of positive-power moments of the strictly stationary solution is characterized by the moment-generating function (MGF) of the model, which depends on the parameter driving the dynamics and on the distribution of the innovations. The asymptotic distribution of the empirical MGF is established, from which tests of moments are deduced. Alternative tests relying on the estimation of the maximal moment exponent (MME) are studied. Power comparisons based on local alternatives and the Bahadur approach are proposed. An illustration is provided on real financial data and it is shown that semi-parametric estimation of the MME offers an interesting alternative to Hill's non-parametric estimator of the tail index.