par Hallin, Marc
Référence Journal of Time Series Analysis, 15, 6, page (607-611)
Publication Publié, 1994-11
Article révisé par les pairs
Résumé : A classical result of Chernoff and Savage (Asymptotic normality and efficiency of certain nonparametric tests. Ann. Math. Statist. 29 (1958), 972–94) on lower bounds for the local asymptotic power of normal-score rank tests, and the resulting Pitman non-admissibility of classical Gaussian testing procedures in general linear models with independent observations, are shown to hold also for linear autoregressive moving-average time series models. The Pitman non-admissibility of usual correlogram-based methods follows.