首页 | 本学科首页   官方微博 | 高级检索  
     检索      


Variance ratio tests of the seasonal unit root hypothesis
Institution:1. School of Mathematical Sciences, Zhejiang University, Hangzhou 310027, PR China;2. College of Mathematics, Physics and Information Engineering, Jiaxing University, Jiaxing 314001, PR China;1. Institut für Mathematische Stochastik, TU Braunschweig, Pockelsstrasse 14, D-38106 Braunschweig, Germany;2. Department of Mathematics and Statistics, University of Cyprus, P.O.Box 20537, CY-1678 Nicosia, Cyprus
Abstract:In this paper we develop tests of the seasonal (quarterly) unit root null hypothesis which reject in favour of stationarity for small values of certain variance ratio statistics, similar to those used by Canova and Hansen (J. Bus. Econom. Statist. 13 (1995) 237) in a different testing context. We demonstrate that our proposed statistics have pivotal limiting distributions under both the null and near seasonally integrated alternatives even when we allow for the possibility of both weak dependence and periodically heteroscedastic behaviour in the driving shocks. This is in contrast to the popular regression-based lag-augmented seasonal unit root tests of Hylleberg et al. (J. Econometrics 44 (1990) 215). A simulation study into the finite sample size and power properties of our proposed tests suggests that they display far superior size properties and, overall, broadly comparable power properties to the corresponding tests of Hylleberg et al. (J. Econometrics 44 (1990) 215), implemented with data-based lag augmentation. The results for the variance ratio tests at the seasonal harmonic frequency are particularly encouraging.
Keywords:
本文献已被 ScienceDirect 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号