We are working on a project that requires the functions performed by
two of Stata routines, IVREG and XTGLS. (Unfortunately, there is no
single Stata program that can do this, so we are trying to create our
own). XTGLS performs the bulk of what we need, we would like to adjust
its ado file so that it can allow for IV estimation. This can be done
if we can get the XTGLS to replace the OLS residuals that it calculates
in the first step of its two step estimation WITH the IVREG residuals
that we can input to it.
Do you think, you can help us identify what lines of code in XTGLS ado
file needs to be changed so it replaces its calculated OLS residuals
with the IV residuals that we provide it with. Or can you help us to
implement XTGLS with IV. We believe this will require someone with
excellent knowledge of Stata programming and the XTGLS program.
It would be useful to know just what xtgls features you need to use in
an ivreg context. Baum, Schaffer and Stillman have been working on an
enhanced ivreg2 that computes HAC standard errors, GMM, etc. and works
on panel data.
