基于广义非线性工具变量法的综列单位根检验  

Panel unit root test based on generalized nonlinear instrument variables estimation

作　　者：杨继生[1] 王少平[1] 

机构地区：[1]华中科技大学经济学院,武汉430074

出　　处：《系统工程理论与实践》2009年第4期111-118,共8页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(70571026);教育部人文社会科学研究青年项目(08JC790043);华中科技大学人文社科青年基金(2007006)

摘　　要：当截面个体之间显著相关时,非线性工具变量法(NIV)综列单位根检验存在严重的分布扭曲.该文基于似无关回归形式的可行广义最小二乘法和非线性工具变量估计方法,提出了广义非线性工具变量法(GNIV)综列单位根检验,以修正NIV检验的分布扭曲.在存在综列单位根的原假设下,GNIV检验统计量的极限分布为标准正态分布,而在备选假设下则趋向于负无穷大.仿真实验结果显示,在截面个体之间显著相关时,GNIV检验的有限样本性质显著优于Chang的NIV检验和Pesaran的CIPS检验.This paper proposes a unit root test for panel data with cross sectional dependence. The proposed test generalizes the nonlinear IV unit root test proposed by Chang, and hence can correct the size distortion of Chang＇s test when cross sectional dependence is moderate to high. The main idea is to eliminate the cross sectional dependence through covariance matrix weighting and then apply Chang＇s test to the weighted data. Under some sufficient conditions, the proposed test statistic has a limiting standard normal distribution under the null hypothesis and converges to negative infinity under the alternative hypothesis. The finite sample performance of the proposed test is evaluated through a small scale simulation study. The simulation results show that the proposed test compares favorably to the tests proposed by Chang and Pesaran in the presence of cross sectional dependence.

关 键 词：综列单位根检验 非线性工具变量 可行广义最小二乘法 截面相关 

分 类 号：F064.1[经济管理—政治经济学]