[quote]luxiahyun 发表于 2010-3-24 14:06

首先,在使用IV(w)的时候需要三个前提:
(i) COR(w, u) = 0 ( ⇐⇒ E(wu) = 0)
(ii) 0 =/= COR(w, x4) (“inclusion restriction” ) x4: the endogenous regressor
(iii) w does not appear in the y equation (“exclusion restriction”)
只要满足这三个条件,那么他就可以成为Instrument variable 而被使用了,管他是几个。(当然,工具变量的个数应该大于等于内生变量个数)
而,根据楼主的提示,这应该是说你的工具变量没有找对,不完全满足以上的三个条件。
其实,在微观计量方面尤其是cross sectional data当中找合适的工具变量又谈何容易啊~
祝楼主好运 [/quote
thank you for your reply!
in mycase, seems not because my IV is not suitable for the model...because i did not even do any tests about IV...this windows showed up when i was just about to start import IV...
and then i used only one explanatory variable...then everything is ok....so i guess it is because of 工具变量的个数应该大于等于内生变量个数...