Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[no subject]



----------------------------------------------------------------

On Tue, 6 Oct 2009, Hewan Belay wrote:

Dear all,

I was working through the description of xtdpd in Stata's book manual "Longitudinal Data/Panel Data [XT]". Example 5 of xtdpd in both the Stata 10 book (p. 78) and the Stata 11 book (p.80) start by specifying the regression: 

xtdpd L(0/1).n L(0/2).(w k) yr1980-yr1984 year,
div(L(0/1).(w k) yr1980-yr1984 year) dgmmiv(n) hascons

But shouldn't the command correctly read instead:

xtdpd L(0/1).n L(0/2).(w k) yr1980-yr1984 year,
div(L(0/2).(w k) yr1980-yr1984 year) dgmmiv(n) hascons


After all, since w and k and their lags included in the regression are treated as exogenous, the div() command should include also the second lags since these are specified in the regression. If the stata manual however is correct in specifying the example, I have not properly understood the xtdpd command, hence my question. 

Many thanks,
Hewan
-----------------------

Hewan is correct.

Here we are treating w and k as strictly exogenous, so the first differences of those variables can be used as instruments in the difference equation. Our model includes the first and second lags of those variables as regressors as well, so we must include the first and second lagged differences of those variables as instruments for the difference equation. Therefore, the correct syntax is 

   . xtdpd L(0/1).n L(0/2).(w k) yr1980-yr1984 year,
           div(L(0/2).(w k) yr1980-yr1984 year) dgmmiv(n) hascons

We will correct the example in the manual entry.

 -- Brian Poi
 -- [email protected]


      

*
*   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/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index