[R-SIG-Finance] Changing (seasonal) conditional distribution
Francis X. Diebold
fdiebold at sas.upenn.edu
Tue May 20 18:50:08 CEST 2014
Just add seasonals as "exogenous" variables in the GARCH equation. Use of
Fourier terms can achieve significant economy relative to a full set of
dummies. See http://www.ssc.upenn.edu/~fdiebold/papers/paper53/reprint.pdf.
--
Francis X. Diebold
Paul F. and Warren S. Miller Professor of Economics
Professor of Finance and Statistics
University of Pennsylvania
www.ssc.upenn.edu/~fdiebold/
@FrancisDiebold
No Hesitations: www.fxdiebold.blogspot.com
-----Original Message-----
From: r-sig-finance-bounces at r-project.org
[mailto:r-sig-finance-bounces at r-project.org] On Behalf Of
r-sig-finance-request at r-project.org
Sent: Tuesday, May 20, 2014 6:00 AM
To: r-sig-finance at r-project.org
Subject: R-SIG-Finance Digest, Vol 120, Issue 19
Send R-SIG-Finance mailing list submissions to
r-sig-finance at r-project.org
To subscribe or unsubscribe via the World Wide Web, visit
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
or, via email, send a message with subject or body 'help' to
r-sig-finance-request at r-project.org
You can reach the person managing the list at
r-sig-finance-owner at r-project.org
When replying, please edit your Subject line so it is more specific than
"Re: Contents of R-SIG-Finance digest..."
Today's Topics:
1. Changing (seasonal) conditional distribution in a fGarch
model (PoddyOne)
2. Re: Changing (seasonal) conditional distribution in a fGarch
model (alexios ghalanos)
3. Re: Changing (seasonal) conditional distribution in a fGarch
model (PoddyOne)
4. Re: Changing (seasonal) conditional distribution in a fGarch
model (PoddyOne)
5. Re: Changing (seasonal) conditional distribution in a fGarch
model (Alexios Ghalanos)
----------------------------------------------------------------------
Message: 1
Date: Mon, 19 May 2014 15:10:52 -0700 (PDT)
From: PoddyOne <padarn at gmail.com>
To: r-sig-finance at r-project.org
Subject: [R-SIG-Finance] Changing (seasonal) conditional distribution
in a fGarch model
Message-ID: <1400537452329-4690850.post at n4.nabble.com>
Content-Type: text/plain; charset=us-ascii
Hi there,
I'm working with a time series that has a clear seasonal component in it's
conditional distribution. Not only does the variance increase at certain
parts of the seasonal cycle, but also the distribution becomes skewed.
I would like to fit an fGarch model to this time series, but with a
conditional distribution which depends on a dummy variable. I was wondering
if anyone knew if this was possible in fGarch or any of the similar
packages?
/(As an aside, I recognise that this may not be the easiest way to go about
modelling this. For a little more detail: This is a time series with a
strong diurnal cycle in it. The series has been 'detrended' and is
stationary at least under the usual metrics. However, still clearly during
the midnight hours, the residuals are strongly positively skewed. I have
tried transforming the series, but the large disparity in both the skew and
scale see to prohibit this working nicely.
Any other suggestions are welcome)/
--
View this message in context:
http://r.789695.n4.nabble.com/Changing-seasonal-conditional-distribution-in-
a-fGarch-model-tp4690850.html
Sent from the Rmetrics mailing list archive at Nabble.com.
------------------------------
Message: 2
Date: Mon, 19 May 2014 23:34:31 +0100
From: alexios ghalanos <alexios at 4dscape.com>
To: PoddyOne <padarn at gmail.com>
Cc: r-sig-finance at r-project.org
Subject: Re: [R-SIG-Finance] Changing (seasonal) conditional
distribution in a fGarch model
Message-ID: <4A68E0AD-6B2F-4B31-9D6E-7487F38C8DCA at 4dscape.com>
Content-Type: text/plain; charset=windows-1252
Hi,
Try reading this:
http://www.unstarched.net/2013/03/20/high-frequency-garch-the-multiplicative
-component-garch-mcsgarch-model/
You?ll have to switch to using rugarch, but hopefully it should not be too
much of a sacrifice...
Alexios
On 19 May 2014, at 23:10, PoddyOne <padarn at gmail.com> wrote:
> Hi there,
>
> I'm working with a time series that has a clear seasonal component in
> it's conditional distribution. Not only does the variance increase at
> certain parts of the seasonal cycle, but also the distribution becomes
skewed.
>
> I would like to fit an fGarch model to this time series, but with a
> conditional distribution which depends on a dummy variable. I was
> wondering if anyone knew if this was possible in fGarch or any of the
> similar packages?
>
> /(As an aside, I recognise that this may not be the easiest way to go
> about modelling this. For a little more detail: This is a time series
> with a strong diurnal cycle in it. The series has been 'detrended' and
> is stationary at least under the usual metrics. However, still clearly
> during the midnight hours, the residuals are strongly positively
> skewed. I have tried transforming the series, but the large disparity
> in both the skew and scale see to prohibit this working nicely.
>
> Any other suggestions are welcome)/
>
>
>
> --
> View this message in context:
> http://r.789695.n4.nabble.com/Changing-seasonal-conditional-distributi
> on-in-a-fGarch-model-tp4690850.html
> Sent from the Rmetrics mailing list archive at Nabble.com.
>
> _______________________________________________
> R-SIG-Finance at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions
should go.
>
------------------------------
Message: 3
Date: Mon, 19 May 2014 17:17:15 -0700 (PDT)
From: PoddyOne <padarn at gmail.com>
To: r-sig-finance at r-project.org
Subject: Re: [R-SIG-Finance] Changing (seasonal) conditional
distribution in a fGarch model
Message-ID: <1400545035633-4690853.post at n4.nabble.com>
Content-Type: text/plain; charset=us-ascii
Great! Thanks for that, this looks like pretty much exactly what I am
looking for.
--
View this message in context:
http://r.789695.n4.nabble.com/Changing-seasonal-conditional-distribution-in-
a-fGarch-model-tp4690850p4690853.html
Sent from the Rmetrics mailing list archive at Nabble.com.
------------------------------
Message: 4
Date: Mon, 19 May 2014 18:36:38 -0700 (PDT)
From: PoddyOne <padarn at gmail.com>
To: r-sig-finance at r-project.org
Subject: Re: [R-SIG-Finance] Changing (seasonal) conditional
distribution in a fGarch model
Message-ID: <1400549798601-4690858.post at n4.nabble.com>
Content-Type: text/plain; charset=us-ascii
Hi Alexios
After having (somewhat) digested the link you sent me to the seasonal
variance models in your package, I'm under the impression that the
conditional distribution is still restricted to being normally distributed
(with varying variance)? I was hoping to use a skewed-normal as the
conditional distribution at certain times of day.
Another thing is that at the bottom of the page you say:
" Another possible direction for expansion would be to treat the diurnal
effect separately for each day of the week."
This is an interesting idea, and is something I thought about, but was a
little uneasy about the ideas that popped into my head about how to go about
this. For example, one simple approach might be (at least in my case) to
transform each hour of the diurnal cycle separately to try and match the
conditional distributions as closely as possible. However, this seems like
it would leave to an overly complicated model.
Cheers
--
View this message in context:
http://r.789695.n4.nabble.com/Changing-seasonal-conditional-distribution-in-
a-fGarch-model-tp4690850p4690858.html
Sent from the Rmetrics mailing list archive at Nabble.com.
------------------------------
Message: 5
Date: Tue, 20 May 2014 03:18:52 -0500
From: Alexios Ghalanos <alexios at 4dscape.com>
To: PoddyOne <padarn at gmail.com>
Cc: "r-sig-finance at r-project.org" <r-sig-finance at r-project.org>
Subject: Re: [R-SIG-Finance] Changing (seasonal) conditional
distribution in a fGarch model
Message-ID: <8531EA42-C8AD-4F0A-810B-A551E6BDA916 at 4dscape.com>
Content-Type: text/plain; charset=us-ascii
Hi,
Yes, you have a wide choice of skewed and shaped distributions which are
covered in detail in the vignette.
As to the expansion, I'm a little busy right now to look at this but feel
free to contribute an enhancement if you are able to.
Best,
Alexios
> On 19 May 2014, at 20:36, PoddyOne <padarn at gmail.com> wrote:
>
> Hi Alexios
>
> After having (somewhat) digested the link you sent me to the seasonal
> variance models in your package, I'm under the impression that the
> conditional distribution is still restricted to being normally
> distributed (with varying variance)? I was hoping to use a
> skewed-normal as the conditional distribution at certain times of day.
>
> Another thing is that at the bottom of the page you say:
>
> " Another possible direction for expansion would be to treat the
> diurnal effect separately for each day of the week."
>
> This is an interesting idea, and is something I thought about, but was
> a little uneasy about the ideas that popped into my head about how to
> go about this. For example, one simple approach might be (at least in
> my case) to transform each hour of the diurnal cycle separately to try
> and match the conditional distributions as closely as possible.
> However, this seems like it would leave to an overly complicated model.
>
> Cheers
>
>
>
> --
> View this message in context:
> http://r.789695.n4.nabble.com/Changing-seasonal-conditional-distributi
> on-in-a-fGarch-model-tp4690850p4690858.html
> Sent from the Rmetrics mailing list archive at Nabble.com.
>
> _______________________________________________
> R-SIG-Finance at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only. If you want to post, subscribe first.
> -- Also note that this is not the r-help list where general R questions
should go.
>
------------------------------
_______________________________________________
R-SIG-Finance mailing list
R-SIG-Finance at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
End of R-SIG-Finance Digest, Vol 120, Issue 19
More information about the R-SIG-Finance
mailing list