Sara Gabriela Castellanos Pascacio
>
> I am estimating a model with the xtabond command and find
> the current
> restriction of 800 matsize binding and inadequate for my
> analysis. The
> reason is that at such size I am forced to constrain the
> maximum number of
> lags of the dependent variable that can be used as
> instruments quite below
> benchmark of T_i-p-2 (proposed for the Arellano-Bond
> estimators). My data
> consists of i=202 cross sections and t=87 time periods. For
> a starter, this
> wouldn't seem a too large dataset to estimate a model with
> p=12 lags of the
> dependent variables, two exogenous/predetermined variables, and
> cross-section and time dummies as:
>
> y_it = y_(it-1)a_1 + ... + y_(it-p)a_p + x_(it)b_1 + v_i
> + w_t + e_(it)
>
> Suggestions to overcome this matsize constraint are
> welcome. In fact, I
> switched to STATA 7 from GAUSS where I was using the DPD98.
> At this moment,
> I wonder if I should switch back to GAUSS.
StataSE has a larger limit. In StataSE 8,
you can have a matsize of 11,000. Naturally,
someone will have to pay for the upgrade.
Nick
n.j.cox@durham.ac.uk
*
* 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/