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

st: Re: constant terms in -xtivreg-?


From   "Mark Schaffer" <M.E.Schaffer@hw.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   st: Re: constant terms in -xtivreg-?
Date   Wed, 16 Nov 2005 18:04:48 -0000

Philippe,

Subject:        	constant terms in -xtivreg-?
Date sent:      	Wed, 16 Nov 2005 17:00:25 +0100
From:           	"philippe van kerm" <philippe.vankerm@ceps.lu>
To:             	<m.e.schaffer@hw.ac.uk>

> Dear Mark,
> 
> I'm playing about with your -xtivreg2- command. It looks very useful!
> 
> But I have come across a, seemingly silly, question: why is there, by
> default, a constant term when using first-differences?

It was basically because that way, -xtivreg2- mimics Stata's official 
-xtivreg-.

> I have only
> tried the command with no instrumental variables, that is, just as a
> way to compute the first-difference model. In this case, I have the
> *impression* that adding a constant biases the coefficients (though I
> might be completely wrong!). I fact I observe large differences
> between the FE and the FD coefficients in my test data, but the
> coefficients become much closer when adding the -noconstant- with the
> -fd- model. I'm pretty sure I miss something obvious, but can't figure
> out what?
> 
> Is it simply that, to estimate the simple FD model, -noconstant- is
> *required*?

Exactly!  The constant term in the FD model is just a simple linear 
trend.  If you had a variable t in the raw data, then the FD of t is 
just t-(t-1) = 1.  -nocons- implies no trend.

> (In this case, maybe a line in the help file would be
> useful?) Does it also hold if you specify instruments?

Yes (and yes).

> A related question is whether you know of a trick in the FD model to
> recover an estimate for the constant term (other than taking the mean
> of the residuals in the back-transformed model); that is stg akin to
> adding the grand mean back in the "within transformation" as is
> implemented in the FE model with -xtreg-?

This doesn't ring a bell.  Maybe someone else on Statalist knows the 
answer to this one...?

--Mark

> 
> Hope you are doing well,
> 
> 
> Best wishes
> 
> 
> Philippe
> 
> 
> 
> 
> 
> <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>>
> Dr. Philippe Van Kerm
> IRISS - Integrated Research Infrastructure in 
>   the Socio-Economic Sciences
> CEPS/INSTEAD
> BP48, L-4501 Differdange, Luxembourg
> >> http://www.ceps.lu/iriss <<
> >> http://econpapers.repec.org/RAS/pva19.htm <<
> <<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>><<<>>> 
> 
> **********************************************************************
> This email and any files transmitted with it are confidential and
> intended solely for the use of the individual or entity to whom they
> are addressed. If you have received this email in error please notify
> the system manager.
> 
> This footnote also confirms that this email message has been swept
> **********************************************************************
> 
> 

Prof. Mark E. Schaffer
Director
Centre for Economic Reform and Transformation
Department of Economics
School of Management & Languages
Heriot-Watt University, Edinburgh EH14 4AS  UK
44-131-451-3494 direct
44-131-451-3296 fax
44-131-451-4202 CERT administrator
http://www.sml.hw.ac.uk/cert

*
*   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