Anupam asked:
###
Now, I am specifying the command as:
xi: xtabond y l(0/1).x1 l(0/1).x2 i.time, diffvars(iz1 iz2 iz3) lags(1)
robust small artests(1);
In above specification, y is the depvar, x1 and x2 are covariates with their
lags. And, z's in diffvars() are strictly exogenous
variables which are already first-differenced (i.e. iz's). My questions are:
does xtabond perform the first differencing for all the variables
specified in the command line? Do I need to first-difference the vars in
diffvars() first and then specify them in the command line? Do I also need
to first difference the y and x's and then specify them in the command line?
or I do not need to do any first-differencing before specifying them in the
command line - 'xtabond' would do that and estimate?
###
I have not the manual for xtabond here, but if I remember correctly, xtabond
performs the differencing of all the variables before the estimation (so,
you don't need to do it). The only variables that xtabond assumes to be
already in first-difference are those included in diffvars. In general, you
do not need to include variables in diffvars, unless you want to prevent
that xtabond differences some variables (for instance time-invariant
dummies).
I hope this is of some help.
Mario
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/