zhaojumping解释的很好!当您使用指令xtwest时,我觉得楼主应当好好看Westerlund在Stata Journal的文章,以及其刊登于2007年的文章。
关于zhaojumping的理解,应当是基于文章中的:
Two tests are designed to test the alternative hypothesis that the panel is cointegrated as a whole, while the other two test the alternative that at least one unit is cointegrated.
这四个检定统计量,Gt Ga指的是Group mean test,后两个都有P,即Panel test
我比较好奇的是,您的Ga与Pa都无法拒绝原假设,这也是您困扰的原因吧!
其实,Westerlund在它的实证结果演练里,就曾点出这种可能的状况与不完美,
他于The Stata Journal的 (2008) 8, Number 2, pp. 240 【页】中指出
In small datasets (such as in this application with T = 32), the results may be
sensitive to the specific choice of parameters such as lag and lead lengths and the kernel
width. When we restrict the short-run dynamics and use a shorter kernel window, the
Ga statistic no longer rejects the H0 of no cointegration
【我想说的是,这里应当提供楼主一种思考,是不是您的实证样本,时间期太少? 至少作者32期都发生那样的状况了!】
【当然,进一步在power的问题探讨上,这可能要去看2007年Westerlund的paper吧!】