[R-SIG-Finance] cointegrated series with Data Missing

[Eric Zivot]

With time series data, filling in missing data can be tricky. State space
models are particulary well suited for this problem because you can specify
explicitly the law of motion and the Kalman filter will automatically
fill-in the missing values in an optimal way.
ez 

-----Original Message-----
From: r-sig-finance-bounces at stat.math.ethz.ch
[mailto:r-sig-finance-bounces at stat.math.ethz.ch] On Behalf Of Brian G.
Peterson
Sent: Thursday, August 14, 2008 9:09 AM
To: mvstatistics at yahoo.com.br; R-SIG-Finance
Subject: Re: [R-SIG-Finance] cointegrated series with Data Missing

I don't believe that the missing data methodologies for cointegrated time
series are really any different than for other (irregular) time series that
you want to compare to each other.

As I said in an earlier post, xts and zoo contain multiple methods for
merging the date indexes from multiple series and then additional methods to
apply to NA handling, like last observation carried forward and
interpolated.

Regards,

   - Brian


Marcus Vinícius Soares wrote:
> Dear John,
>  
> I´m trying to test for cointegration. I thought that if I use an MA as
estimation with last data before each missing, perhaps one or two weeks, it
could do the trick. But I can find some week sazonality, since market in a
Friday may be more or less active than in a Monday. The fact is that I can't
test cointegration with missing, and I wouldn't just delete the unpaired
data.
> Sugestions?
>  
> I searched in the list of historical mails of this list and didn't find
anything that fits on this case.
> 
> Rgds,
> Marcus
> 
> --- Em ter, 12/8/08, John Frain <frainj at tcd.ie> escreveu:
> 
> De: John Frain <frainj at tcd.ie>
> Assunto: Re: [R-SIG-Finance] cointegrated series with Data Missing
> Para: mvstatistics at yahoo.com.br
> Data: Terça-feira, 12 de Agosto de 2008, 11:00
> 
> It is not clear what you aim to do with the series.  Are you
> 
> 1) Trying to test for cointegration
> 
> 2) Estimating a cointegrating relationship
> 
> 3) Estimating missing values.  If so how are you going to use the 
> estimates.  Perhaps a) estimate a VECM, b) fill in missing values, c) 
> reestimate and d) cycle to convergence.
> 
> Best Regards
> 
> John
> 
> 
> 
> 
> 
> 2008/8/12 Marcus Vinícius Soares <mvstatistics at yahoo.com.br>:
>> Hi there!
>>
>> First, I´m from Brazil, so sorry for any English mistake.
>> I have two cointegrated series with Data Missing in both. They are 
>> two
> finance series, one from USA and another from Brazil, so in holidays 
> here I have data for USA serie and same for holidays in USA.
>> I think usual Data Missing thecnics do not apply, since they are time
> series, and are also cointegrated.
>> Can someone help me?
>>
>> Thanks very much,
>>
>> Marcus Vinicius
>>
>>
>>
>>      Novos endereços, o Yahoo! que você conhece. Crie um email novo 
>> com
> a sua cara @ymail.com ou @rocketmail.com.
>> http://br.new.mail.yahoo.com/addresses
>>        [[alternative HTML version deleted]]
>>
>>
>> _______________________________________________
>> R-SIG-Finance at stat.math.ethz.ch mailing list 
>> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
>> -- Subscriber-posting only.
>> -- If you want to post, subscribe first.
>>
> 
> 
> 
> 
> ----------------------------------------------------------------------
> --
> 
> _______________________________________________
> R-SIG-Finance at stat.math.ethz.ch mailing list 
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only.
> -- If you want to post, subscribe first.

_______________________________________________
R-SIG-Finance at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only.
-- If you want to post, subscribe first.