On Sep 5, 2009, at 2:33 AM, Laura wrote:
Thanks a lot for your answer. I'm trying to use booststrap to estimate
standard errors, because I plugged in my equation the fitted values
of an
auxiliary regression because the regressor it very likely to be
endogenous.
That makes little sense, as the whole point of A-B estimation (via
xtabond, or xtabond2 from SSC) is to allow any RHS variable to be
instrumented to deal with potential endogeneity. Furtrhermore the
standard -bootstrap- really is not appropriate for time-series data;
you should be using a block bootstrap. But there should be no need for
bootstrapping in this context. See Roodman, "How to do xtabond2", at http://ideas.repec.org
Kit Baum
*
* 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/