Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: question about xtabond command

From   li huang <>
Subject   st: question about xtabond command
Date   Wed, 10 Aug 2005 06:10:31 -0700 (PDT)

>Emily <> asks why if the default
for the number of lags of the dependent variable to be
included in the -xtabond- specification is one,
the estimation output reports a number of observations
that is reduced by an amount equivalent to two lags.
>. tsset country1 year
         panel variable:  country1, 1 to 65
          time variable:  year, 1902 to 1907

>. xtabond lmanux lprivo lpop gtotols lgdp lbmp
lcapital year country1

>Arellano-Bond dynamic panel-data estimation
   Number of obs      =       207

>The Arellano-Bond methodology is based on the
estimation of a first difference equation derived from
the original dynamic panel data model (in levels).
Therefore,the first observation of each panel is
missing due to first-differencing the model,whereas
the second observation is missing due to the inclusion
of one lag of the
dependent variable on the right hand side of the
equation. Thus, given that Emily has 65 panels, the
number of observations available for the -xtabond-
estimation is: 207 (=337-65*2)


Thanks a lot!

I tried adding one option "maxldep(0)" to the command.
Because I don't want to use one lag of dependent
variable on the right hand side of the equation as
instruments. But the second observation is still

(Arellno and Bond only use lagged values of the levels
of the regressors as instruments.)

Could someone tell me how to make the second
observation on each panel not missing?

Many thanks in advance!


Start your day with Yahoo! - make it your home page

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index