Tomislav,
In the following paper there is a rule of thumb for dynamic
panel data, that follows your intuitive analysis:
Hahn, J. and G. Kuersteiner (2002) "Asymptotically Unbiased
Inference for a Dynamic Panel Model with Fixed Effects when
both n and T are large," Econometrica, 70 (2002), 1639-1657
A draft version is available at
http://people.bu.edu/gkuerste/research/NTpanel.pdf
Section 3 is the case that you are looking for, equation 6.
Basically, the bias-corrected estimator (obtained in this paper)
is b1 = (T+1)*b0 + 1/T, where b0 is MLE (OLS) estimator
or the estimator obtained by "brute force".
Rodrigo.
----- Original Message -----
From: "Tomislav Kovandzic" <[email protected]>
To: <[email protected]>
Sent: Saturday, July 08, 2006 2:22 PM
Subject: st: dynamic panel models with fixed effects
Dear List,
I'm working with state panel data for roughly 30 years and estimating a
fixed-effects model with a lagged dependent variable on the right-hand
side to address autocorrelation. Several colleagues have informed me
that including a lagged dependent variables on the RHS when utilizing a
fixed-effects model is simply a no-no, i.e. such estimations yield
inconsistent estimates. My question is whether the inconsistency dies
away as the number of time-periods goes off to infinity? If so, are
there any rules of thumb as to how many time periods are needed to
mitigate the problems described above. A reference would be much
appreciated. If the estimates are inconsistent regardless of the number
of time-periods, can someone offer an alternative approach to correcting
for autocorrelation when using a fixed-effects model? Would cluster
robust standard errors do the trick? Thanks in advance.
Best,
Tom
*
* 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/
*
* 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/
