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: Cross sectional dependence

From   Rodrigo Refoios Camejo <>
Subject   st: Cross sectional dependence
Date   Thu, 16 Jun 2011 19:37:13 +0100

Hi all,

I have a panel with the prices of 25 drugs over 31 years (T=31 N=25).
The panel is (non-randomly) unbalanced, because it starts with very
small N (e.g. in 1980 only 4 molecules were available) and molecules
enter the market gradually and and tend to stay in the market. I am
running a model and was trying to check for cross sectional dependence
using the normal test available but none seems to work:
- xttest2 reports error: too few common observations across panel. no
observations r(2000)
- xtcsd, pesaran error: the panel is highly unbalanced. Not enough
common observations across panel to perform Pesaran's test.
insuficient observations r(2001)
- xtcd too many variables specified r(103)

I'm using xtreg a1 b1 b2 b3, fe robust

If none works, should I just assume that there is cross sectional
dependence and use the DRiscol-Kraay SE instead of the Newey-West SE?
Also, the commands available for these two assume FE pooled OLS. Is
there any way to have a RE with those corrections in place?

Many thanks in advance for the help!


*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index