Dear STATA users
I choose as specification a fixed effects model. From F tests I observe
that account for individual effects matter. So, I suspect the possible
autocorrelation in the error term. I use the xtregar command in STATA
which correct for ar(1) in fixed effects model. Low BFN values mean that
there is autocorrelation. I also consider the FD estimator which uses
time series variation as well. FD and FE (fixed effects) should give the
same results if the chosen specification is correct. But the results
diverge quite a bit. This may due to the assumptions behind the error
term. Finally, I check the residuals obtained when estimating by first
differences. There is no autocorrelation. My point is what will the
right specification? is the FD model comparable that estimated with FE
with time dummies in levels?
regards
toni
*
* 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/
