Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: panel data

From   Fred Dzanku <>
Subject   st: panel data
Date   Tue, 31 May 2011 14:03:25 +0100

Dear All, I need some help in choosing between some panel data
estimators. I have a production function in which choice of input
levels may be correlated with individual effects in which case a
Hausman-Taylor type estimator may be appropriate (is that right?). My
confusion however is that an initial model under the fixed effects
assumption fails to reject the null hypothesis that all fixed effects
are equal to zero. If there are no fixed effects why should one
estimate a Hausman-Taylor type model which assumes a correlation
between unobserved individual-level random effect and inputs? Indeed,
the xthtaylor estimator also reports sigma_u=0 and therefore rho=0.
Does this estimator them make any sense?

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index