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

st: Re: xtabond


From   TEWODAJ MOGUES <tmogues@students.wisc.edu>
To   ara@sam.sdu.dk, Orsetta.CAUSA@oecd.org, Stata _ <statalist@hsphsun2.harvard.edu>
Subject   st: Re: xtabond
Date   Mon, 03 May 2004 18:50:38 -0500

Hi Toni (and Orsetta),

A couple things: When you stipulate   maxldep(3)   and, say, you are using one lag of y as a regressor, then this regressor shows up in the xtabond regression as Dy(i,t-1), which can also be written as y(i,t-1) - y(i,t-2). So the 3 instruments used for this regressor are: 
y(i,t-2),
y(i,t-3),
y(i,t-4).
If, on the other hand, you used in your command  maxldep(2), then the two instruments are: 
y(i,t-2),
y(i,t-3).
I guess the spirit of your question is: "But what if I wanted to use as instruments solely  y(i,t-3)?  Or  y(i,t-3) and  y(i,t-4)?" There are two responses to that: 1) Why would you want to use further away lags and "skip" closer lags? 2) Whatever your reason, you can't with xtabond (I'm pretty sure of that, but if anyone thinks otherwise, please let me know). xtabond just lets you set the furthest away lag, but then uses that lag plus all other closer ones that constitute valid instruments. However, if you insist that you do want to select the instruments for yourself, I strongly recommend you download (free) the command xtabond2. xtabond2 gives you much more freedom in designing the instrument set, and can also do other kool things that xtabond can't. To download, type in STATA
    net search xtabond2
and the instructions how to download it are there.

Good luck!
Tewodaj

Original Message:

 Date: Sun, 02 May 2004 21:48:35 +0200
 From: "Antonio Rodrigues Andres" <ara@sam.sdu.dk 
Subject: (unknown)

Hi everybody
I am using xtabond to estimate a dynamic panel data model. I want to use lagged values of Y as instruments. There is an option named maxldep().
Assume that I want to use only y(i,t-2). If I specify maxldep(3), STATA
uses yt-1 yt-2 yt-3 as instruments of Dy(i,t-1)

Any thoughts
Thanks, Antonio
 Date: Mon, 3 May 2004 12:55:10 +0200
 From: <Orsetta.CAUSA@oecd.org 
Subject: st: RE: xtabond

Hi antonio
If you specify maxldep (3) I think that stata will use yt-2 and yt-3 and yt-4
(it will never use yt-1, it is not a good instrument in the differenced
equation). Perhaps if you want to use yit-2 that menas one lag since yt-1
does not work, so you should try with maxldep(1)
Hope it works
orsetta


~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~
Tewodaj Mogues
Dept. of Agricultural and Applied Economics
University of Wisconsin - Madison
427 Lorch St. #317, Taylor Hall
Madison, WI 53706

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index