Re: error in summary output for mgarch model

[Patrick Burns <pburns@pburns.seanet.com>]

I don't have a comment on the errors that you are
getting, but I do on trying to fit the model in the first
place.

300 observations is probably too few to get a meaningful
estimate from the model.  It certainly is if the data are
daily.   I don't have all that much experience with weekly
and monthly data -- it might not be terrible in these cases.


Patrick Burns
patrick@burns-stat.com
+44 (0)20 8525 0696
http://www.burns-stat.com
(home of S Poetry and "A Guide for the Unwilling S User")

Abhijit Sarkar wrote:

> Hi!
>
> I am pretty new, not even a month into SPlus/SPlusFinMetrics. I would
> appreciate any help from experienced users in this group.
>
> I am trying to estimate an MGARCH-in-Mean model with three time series,
> each of length 303. The mean equation includes three exogeneous variables.
> The conditional volatility equation also includes three exogeneous
> variables. The conditional volatility term in mean equation is in the form
> logvar. I am using BEKK(1,1).
>
> The command I am using works fine...ie no error message comes and
> convergence is reached...but when I call summary(...) problem occurs. The
> details are given below:
> ----
>  
>
> > mg.logvar.g<-mgarch(mg~1+ar(2)+logvar.in.mean,
> >    
> x=cbind(niftyfin,ytmfin,exchfin), z=cbind(sigma.niftyfin.lag1,
> sigma.ytmfin.lag1, sigma.exchfin.lag1),series.start=2, x.start=2,
> z.start=2, ~bekk(1,1))
>
> Convergence R-Square = 0 is less than tolerance = 0.0001
> Convergence reached.
>
>  
>
> > summary(mg.logvar.g)
> >    
> Problem in cov[1:NP, 1:NP]: Array subscript (52) out of bounds, should be
> at most 51
> Use traceback() to see the call stack
>
>  
>
> > traceback()
> >    
> 8: eval(action, sys.parent())
> 7: doErrorAction("Problem in cov[1:NP, 1:NP]: Array subscript (52) out of
> bounds, should be at most 51",
> 6: cov[1:NP, 1:NP]
> 5: vcov.mgarch(x, method = method, partial = F)
> 4: summary.mgarch(mg.logvar.g)
> 3: summary(mg.logvar.g)
> 2: eval(expression(summary(mg.logvar.g)))
> 1:
> Message: Problem in cov[1:NP, 1:NP]: Array subscript (52) out of bounds,
> should be at most 51
> ---
>
> Any assistance in pointing out what may be the problem/mistake will be
> highly appreciated.
>
> Yours thankfully,
>
> Abhijit
> IITBombay, India
>
>
> --------------------------------------------------------------------
> This message was distributed by s-news@lists.biostat.wustl.edu.  To
> unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with
> the BODY of the message:  unsubscribe s-news
>
>
>
>