Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Bernardo Schettini <firstname.lastname@example.org>
Wed, 25 May 2011 13:54:12 -0300
Good morning. I am using Inessa Love's panel-VAR codes pvar.ado, along
with sgmm.ado and Helm.ado to do system GMM and forward
mean-defference the variables. I can run the programs and they seem to
work just fine. The problem is I am not experienced in reading or
writing programming language and hence I am having a hard time doing
some basics, and I can't figure out if estimation is exactly the way I
expect. To summarize, I have 4 main questions and I would be very
grateful if anyone have the time to help me.
1 - How do I include exogenous variables in the model? The syntax is
pvar varlist [if exp], [lag(p) options], but all variables in the
valist are treated as endogenous. How am I supposed to include time
dummies or other exogenous variables?
2 - It is adviced to time-demean variables before doing the Helmert
procedure (forward mean-differencing), but why is that? If I opt to
time-demean beforehand, then in the pvar command should I use the
original (untransformed) variables or time-demeaned variables?
3 - Does sgmm actually do system GMM as in Arellano and Bover (1995)
and Blundell and Bond (1998) with additional equations for the
variables in levels using first differences (or Helmert?) as
4 - If so, can I use xtabond2 in Stata and do hypothesis testing
previous to impulse response with pvar and sgmm? This is important
because, for example, I have to test for Granger-causality and there
is no prior to help me define the number of lags in the panel-VAR.
P.S.:The source for the panel-VAR codes provided by Inessa Love is:
* For searches and help try: