st: Estimating a dynamic limited dependent variable model (xtabond)

[Alex Safavi <asafavi@gmail.com>]

Dear Statalisters,

I have a question relating to a dynamic limited dependent variable
model which I wish to estimate. In my study, I'm exploring the
determinants of the firm's export decision, and have balanced panel
data following 650 firms over a 9-year period. I use Stata SE 10.

The equation which I wish to estimate is:
Y_i,t = Y_i,t-1 + Y_i,t-2 + Y_i,t-3 + X_i,t + e_i,t

where:
Y_i,t   - is a dummy variable for whether firm i exported in period t or not
Y_i,t-1   - is the export status of firm i in period t-1
X_i,t    - is a vector representing various plant characteristics,
including average wages, capital stocks, industry dummies and location
dummies (note: the industry & location dummies are constant across
time).

To deal with the violation of the "strict exogeneity assumption", I
was considering using the Arellano-Bond estimator. I've written the
command:

xtabond dexp lnw_1 lnw_2 lnk_1 lnk_2 lnage dcorp  dtextile dchemical
dfood, endogenous(dexp_1 dexp_2 dexp_3 dexp_1dfood dexp_2dfood
dexp_3dfood dexp_1dtextile dexp_2dtextile dexp_3dtextile
dexp_1dchemical) robust

My questions are the following:
(1) Is the syntax of the above command correct?
(2) Is there anything that I should add to the above command to tell
Stata that my dependent variable is binary?
(3) I wanted to control for individual firm heterogeneity (given that
there are unobservable factors such as product quality that are likely
correlated with the explanatory variables). What's the best way of
doing this? Should I just include firm dummies?
(4) The time-constant variables (dcorp  dtextile dchemical dfood) are
being dropped. Is there anything that I can do about this?


Many thanks for your consideration.

Kind regards,
Alex
Undergraduate student

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/