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

st: RE: ivreg2


From   Mark Schaffer <M.E.Schaffer@hw.ac.uk>
To   Fabia Aparecida de Carvalho <fabia.carvalho@bcb.gov.br>
Subject   st: RE: ivreg2
Date   Thu, 24 Feb 2005 21:18:49 +0000 (GMT)

Fabia,

I'm posting this reply to the list, partly in the hope that someone will be
able to help answer the question of why you can't.  Have you registered with
Statalist?

With respect to your ivreg2 questions,

Quoting Fabia Aparecida de Carvalho <fabia.carvalho@bcb.gov.br>:

> Dear Prof. Schaffer, I have been trying to post messages to the
> Statalist but they never go through. Anyway, I think the problem was
> with the version of Stata (mine was 7). I tried the same database
> and the same commands in Stata 8 and it worked.

To be more precise, you were probably using an early version of Stata 7 that
hadn't been updated.  ivreg2 uses the syntax command, and I think this
wasn't in the original Stata 7 release.  The update command fixes these things.

> Could you tell me if ivreg2 allows for serial correlation across
> clusters with the assumption of no heteroscedasticity (i.e., the
> elements in the diagonal matrices of the covar matrix have the same
> variance but the off-diagonal matrices may be different from zero -
> just as in Keane and Runkle 1990)?

I don't have Keane and Runkle at hand, but if I understand the question
correctly then the answer is yes.  If you use the bw() option and you have
panel data, you will get a covar matrix estimate that is robust to serial
correlation.  If you add the gmm option to this, you will also get
coefficient estimates that are efficient in the presence of serial
correlation.  In both case, there is an assumption of homoskedasticity.

To get a covar matrix that is robust to heteroskedasticity as well (or
coeffs that are efficient in the presence of arbitrary heteroskedasticity),
you would add the robust option.

If you add the cluster option, you get robustness in the presence of
arbitrary serial correlation and heteroskedasticity.  The different between
this and the combination of robust and bw() is that the latter makes
assumptions about how quickly the serial correlation dies out that are
required for the asymptotics to work.

Cheers,
Mark

> Sorry for asking this directly to
> you but I have e-mailed the list owner mentioning my problem but I
> haven't received any reply.
>  
> Thank you, Fabia
> 
> 	-----Original Message----- 
> 	From: Mark Schaffer [mailto:M.E.Schaffer@hw.ac.uk] 
> 	Sent: Mon 2/21/2005 2:51 PM 
> 	To: Fabia Aparecida de Carvalho 
> 	Cc: 
> 	Subject: RE: ivreg2
> 	
> 	
> 
> 	Just a note to say that I haven't received the Statalist message
> ...
> 	not sure why.
> 	
> 	--Mark
> 	
> 	Subject:                RE: ivreg2
> 	Date sent:              Mon, 21 Feb 2005 12:36:41 -0300
> 	From:                   "Fabia Aparecida de Carvalho"
> <fabia.carvalho@bcb.gov.br>
> 	To:                     <M.E.Schaffer@hw.ac.uk>
> 	
> 	> thank you very much for the quick response. I've just sent a
> message
> 	> to the list. Best, Fabia
> 	>
> 	>  -----Original Message-----
> 	>  From: Mark Schaffer [mailto:M.E.Schaffer@hw.ac.uk]
> 	>  Sent: Mon 2/21/2005 11:54 AM
> 	>  To: Fabia Aparecida de Carvalho
> 	>  Cc: stillman@motu.org.nz; baum@bc.edu
> 	>  Subject: Re: ivreg2
> 	>
> 	>
> 	>
> 	>  If you give us the examples of what you're doing, we might be
> able to
> 	>  help.  Also, try posting it to Statalist rather than to us
> privately,
> 	>  since others might benefit from the answer - or indeed might
> have the
> 	>  answer themselves.
> 	>
> 	>  --Mark
> 	>
> 	>  Subject:                ivreg2
> 	>  Date sent:              Mon, 21 Feb 2005 11:51:55 -0300
> 	>  From:                   "Fabia Aparecida de Carvalho"
> 	>  <fabia.carvalho@bcb.gov.br> To:                   
> 	>  <m.e.schaffer@hw.ac.uk> Copies to:            
> 	>  <stillman@motu.org.nz>,
> 	>          <baum@bc.edu>
> 	>
> 	>  > Dear Profs. Schaffer, Stillman and Baum,
> 	>  >
> 	>  > I've been trying to run a gmm with newey-west covar matrix on
> a
> 	>  panel, but Stata always tells me the syntax is wrong. When I use
> the
> 	>  ivreg for 2SLS using the same panel, it works. Do you know what
> could
> 	>  possibly be the problem? Thank you, Fabia Carvalho > > Fabia
> Carvalho
> 	>  > Central Bank of Brazil >
> 	>
> 	>  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
> 	>  44-131-451-3485 CERT administrator
> 	>  http://www.sml.hw.ac.uk/cert
> 	>
> 	>
> 	>
> 	>
> 	
> 	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
> 	44-131-451-3485 CERT administrator
> 	http://www.sml.hw.ac.uk/cert
> 	
> 	
> 
> 



Prof. Mark Schaffer
Director, CERT
Department of Economics
School of Management & Languages
Heriot-Watt University, Edinburgh EH14 4AS
tel +44-131-451-3494 / fax +44-131-451-3294
email: m.e.schaffer@hw.ac.uk
web: http://www.sml.hw.ac.uk/ecomes

-------------------------------------------------------------------
DISCLAIMER

This message is subject to http://www.hw.ac.uk/disclaim.htm 

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

*
*   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