Well the bootstrap approach won't give you any robustness against the
model structure misspecification... it is just not cut out for that
purpose, and it is not a magic wand that will give you the answer no
matter what. (It will give you AN answer, but you have AN answer
already from your -xtivreg-, anyway). If you are pretty sure your
endogeneous variable is Poisson, then you might want to model your
main dependent variable and your endogeneous variable jointly using
say -gllamm- or writing your own -ml d0- likelihood... while -gllamm-
is making the first iteration...
On 4/15/06, Jake Kendall <[email protected]> wrote:
> Hi All,
>
> I am considering the following regression but I am unsure of the validity of
> using bootstrapped errors. I would appreciate any comments anyone has.
>
> I am running an instrumental variables panel regression using random effects
> [ xtivreg ..., re]. I believe that the variable to be instrumented is drawn
> from a Poisson distribution so I would like to find a way to correct for
> this. I had thought of using bootstrapped errors as a way to be robust to
> the possible misspecification. Is this valid or is there some way to test if
> it is valid?
>
>
> thanks in advance
> *
> * 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/
>
--
Stas Kolenikov
http://stas.kolenikov.name
*
* 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/
