如题:在STATA中的XTGLS或XTPCSE,好像均是"处理"PANEL DATA模型的异方差的(即估计具有异方差的模型);问题就是——能用STATA来"检验"PANEL DATA模型的异方差吗?具体是什么命令?用的是什么检验方法?
这好像已经是一个老生常谈的问题了;在本版块中已经有几个完全类似的帖子;但基于以下的"问题背景",再发贴,深盼各路高手指点迷津,不胜感谢!
"问题背景":
在 Baltagi 的 "Econometric Analysis of Panel Data(Third edition)"中,将PANEL DASTA MODEL 中的异方差归为三类:
(1) Mazodier and Trognon(1978): μi are heteroskedastic, but νit is
homoskedastic;
(2) Baltagi(1988): keeping the μi homoskedastic and impose the heteroskedasticity on the νit ;
(3) Randolph(1988): a more general heteroskedastic model where both the μi and the νit are assumed heteroskedastic.
并“文献综述”式的列出了几种主要的检验方法:
(1)Verbon (1980) derived a Lagrange multiplier test
(2)Lejeune(1996), dealt with maximum likelihood estimation and Lagrange multiplier,testing of a general heteroskedastic one-way error components regression mode
(3)Holly and Gardiol (2000) derived a score test for homoskedasticity
(4)Baltagi, Bresson and Pirotte (2005b) derived a joint Lagrange multiplier test