A0725
Title: Semi-strong double-autoregressive models: Structure and estimation
Authors: Xuqin Wang - Xiamen University (China) [presenting]
Abstract: The first-order semi-strong double autoregressive model is investigated, where semi-strong means that the errors are not required to be independent over time. A sufficient condition is firstly obtained for a unique, strictly stationary, and ergodic solution of the model without the need to check irreducibility. Consistency and asymptotic normality of the quasi-maximum likelihood estimated parameters are also studied under some mild conditions. In contrast to the existing literature, the innovation variable is not required to be Gaussian or independent over time. Then, extensions to the higher-order semi-strong double autoregressive model are also discussed. Finally, the finite sample performance of the quasi-maximum likelihood estimator is assessed through Monte Carlo simulations.
