![Statalist](/includes/images/head_statalist.jpg) |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: xtabond2 and instruments (HELP)
Dear all,
I'm a newbie with Stata and I'm trying to estimate a dynamic panel data
model using an unbalanced data set with more than 350 firms over the
1995 -2000 period. Each firm has a minimum of three consecutive years
of data. This is the dynamic model:
y_it = y_it-1 , x1_it , x2_it , x3_it, alpha_i alpha_t uit
where alpha_1 and alpha_t represent firm-specific effects and
time-effects,
All this given, I have two different questions:
1) In my specification, all variables should be treated as endogenous
and time dummies are included among the independent variables. In
particular, y_it-2, x1_it-2 , x2_it-2 , x3_it-2 should be used as
instruments. I'm asking whether the following command string is correct
in this case. Using the One Step Robust GMM System estimator
(xtabond2), I should write:
xi:xtabond2 y l.y x1 x2 x3 i.year, robust gmm(y x1 x2 x3, lag(2 2))
Is ii right or incomplete?
----------------------------
2) In the (applied) econometrics article from which I have borrowed
such procedure, I also read this important note:
"I [the author] have investigated whether the explanatory variables are
predetermined or strictly exogenous with respect to the error term. To
do this, I started using instruments dated t-2 for each regressor.
Later, I added the instrument dated t-1 to analyze the potential bias
arising from the correlation between x_it-1 and the first-differenced
error term, delta_uit . To investigate the possibility of strict
exogeneity we also included the current value, x_it in the instrument
set. This investigation leads me to conclude that the explanatory
variables are neither predetermined nor strictly exogenous. I,
therefore, use instruments dated t-2 in our estimation."
** With respect to the model above, what concrete steps I should do to
follow this sentence? When the author say that "we added the instrument
dated t-1 to analyze the potential bias arising from the correlation
between x_it-1 and the first-differenced error term, delta_uit ", it is
unclear to me how (and where) he actually analyze the "potential bias"
between x_it-1 and the first-differenced error term. Is there an
automatically generated statistics to see? Or something else? And how I
should modify (step by step) the gmm style options? Cookbook
suggestions would be very useful, too.
Can anybody offer any assistance to solve these problems?
Looking forward to any response, thank you in advance.
Yours faithfully,
Robert Brienza,
Bocconi U. (Milan, Italy)
________________________________________________________________________
Email and AIM finally together. You've gotta check out free AOL Mail! -
http://mail.aol.com
*
* 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/