Statalist


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

st: ivregress with cointegrated variables


From   "Podesta', Federico" <Federico.Podesta@amm.unitn.it>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: ivregress with cointegrated variables
Date   Thu, 9 Jul 2009 16:53:58 +0200

Dear all,

I am using a panel dataset composed by 16 units observed over 21 time points. all variables measured as levels are affected by the non-stationarity problem. Accordingly, I have adopted mean group estimator (via xtpmg Stata module) for testing for co-integration. In particular I have used the following STATA syntax:

xtpmg d.y d.x1 d.x2, ec(ec) lr(l.y l.x1 l.x2) replace 

My results reveal that the variables tend to return to a long-run equilibrium. Nevertheless, x2 appears as a endogenous variable. Consequently. I would like to use IV estimator. However no Stata command allows to estimate IV regression with panel integrated processes. Consequently, I mean to run IV regression in a repeated cross-section framework. In particular I would use the following syntax

bys year: ivregress y x2 (x1= x3)

but I wonder whether it is correct estimate such repeated cross-section models when we have already obtained evidence about cointegration. In other words can I test for cointegration without use IV estimator, on the one hand, and estimate IV repeated cross-section models, on the other hand?
   
Thanks a lot in adavance for any your hel 

Best 
Federico 

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index