作 者: ;
机构地区: 广州大学
出 处: 《统计研究》 2011年第1期 99-105,共7页
摘 要: 研究表明相互独立的平稳阈值自回归(TAR)模型之间的回归存在伪回归,且伪回归的产生与样本容量和随机干扰项的分布无关。通过一系列的MC模拟,不仅证实了理论结论,而且模拟结果还表明当持久性相同时,两机制TAR回归模型比三机制TAR回归模型具有更大的拒绝率,原因在于两机制TAR下,OLS法估计得到的标准误具有更厚的左尾。此外在模拟中也发现当随机干扰项服从TAR模型时,Newey-West(1987)的一致异方差估计法是不适用的。 This paper studies that spurious regression exits when a pair of independent threshold regressive models is regressed according to standard OLS method,and it is independent of the sample size and various distribution of the error term.Through a set of Monte-Carlo simulations,this paper finds some evidence that there exits larger reject ratio under two-regime TAR models than under three-regime TAR models when two TAR regression models have the same persistence.It attributes to the estimator of standard error under two-regime TAR models,and the estimator of standard error has larger left-trail probability in two-regime TAR than in the three-regime TAR.As addition,when error term is TAR process,the Newey-West HAC consistent covariance estimate is not suitable.
关 键 词: 伪回归 阈值自回归模型 一致方差-协方差估计 MONTE-CARLO模拟
分 类 号: [C812]
领 域: []