Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: ARMA(1,1) with Multiple Panels


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: ARMA(1,1) with Multiple Panels
Date   Wed, 23 Aug 2006 13:23:47 +0100

If your series is nonstationary, why do you think ARMA(1,1) 
is a plausible model? Am I missing something? 

Nick 
n.j.cox@durham.ac.uk 

Alexander Gelber wrote 
 
> Thank you, this was quite helpful.
> 
> However, I understand that conditional MLE is inappropriate for a
> nonstationary series, so that it would not be a good idea to use the
> "condition" option.  If my series is nonstationary, what alternatives
> would I have for estimating an ARMA(1,1) with multiple panels?
> 
> Best,
> 
> Alex
> 
> On Tue, 22 Aug 2006, Vince Wiggins, StataCorp wrote:
> 
> > Many thanks for this extremely helpful suggestion!
> >
> > Regarding the conditional ML estimator for -arima- models, 
> Alexander Gelber
> > <gelber@fas.harvard.edu> goes on to ask,
> >
> > > When you say that the "conditional estimator is conditional on
> > > simple rules for starting the process at the beginning of the
> > > sample," what simple rules or assumptions, in particular, are you
> > > referring to?  I have had trouble finding the answer using Stata
> > > help.
> >
> > Primarily that e_t=0 and u_t=0 in first periods before they 
> can be estimated
> > from the data, see [TS] arima and the associated references 
> for details.

*
*   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