10:43 a.m.
Hi
Just one question: is there any script for gretl that calculate DOLS (Stock
& Watson 1993)?
Thank you
Claudio
Obrigado / Thanks for your time and attention.
Claudio D. Shikida
http://www.cdshikida.net and http://works.bepress.com/claudio_shikida/
Esta mensagem pode conter informação confidencial e/ou privilegiada. Se
você não for o destinatário ou a pessoa autorizada a receber esta mensagem,
não poderá usar, copiar ou divulgar as informações nela contidas ou tomar
qualquer ação baseada nessas informações. Se você recebeu esta mensagem por
engano, por favor avise imediatamente o remetente, respondendo o presente
e-mail e apague-o em seguida.
This message may contain confidential and/or privileged information. If you
are not the addressee or authorized to receive this for the addressee, you
must not use, copy, disclose or take any action based on this message or
any information herein. If you have received this message in error, please
advise the sender immediately by reply e-mail and delete this message.
3:09 p.m.
Am 12.11.2013 11:43, schrieb Claudio Shikida (敷田治誠 クラウジオ):
Hi
Just one question: is there any script for gretl that calculate DOLS
(Stock & Watson 1993)?
In principle it's easy to do, since x(+1) gives you the lead (and thus
x(+1)-x the lead diff, and so on), which you can use with the ols
command; if necessary, create the regressors in a loop. Nothing special
that you need apart from that, I'd say.
However, it's true that while there's 'lags()' to generate list of lags
quickly, this doesn't seem to cover leads, and IMHO this would be a
useful feature.
Also, because it's impossible to enter leads via GUI into the standard
OLS-dialog, perhaps it's worth thinking about a separate DOLS entry in
the time series submenu?
cheers,
sven
3:13 p.m.
Thanks, Sven.
I just got it in R (well, but it would be great to do it in Gretl too, at
least for comparison of algorithms) but the features of lag-lead and
sandwich for the t-statistics could be two good reasons to create a time
series submenu.
Best Wishes,
Claudio
Obrigado / Thanks for your time and attention.
Claudio D. Shikida
http://www.cdshikida.net and http://works.bepress.com/claudio_shikida/
Esta mensagem pode conter informação confidencial e/ou privilegiada. Se
você não for o destinatário ou a pessoa autorizada a receber esta mensagem,
não poderá usar, copiar ou divulgar as informações nela contidas ou tomar
qualquer ação baseada nessas informações. Se você recebeu esta mensagem por
engano, por favor avise imediatamente o remetente, respondendo o presente
e-mail e apague-o em seguida.
This message may contain confidential and/or privileged information. If you
are not the addressee or authorized to receive this for the addressee, you
must not use, copy, disclose or take any action based on this message or
any information herein. If you have received this message in error, please
advise the sender immediately by reply e-mail and delete this message.
On Tue, Nov 12, 2013 at 1:09 PM, Sven Schreiber <svetosch(a)gmx.net> wrote:
Am 12.11.2013 11:43, schrieb Claudio Shikida (敷田治誠 クラウジオ):
> Hi
>
> Just one question: is there any script for gretl that calculate DOLS
> (Stock & Watson 1993)?
>
In principle it's easy to do, since x(+1) gives you the lead (and thus
x(+1)-x the lead diff, and so on), which you can use with the ols
command; if necessary, create the regressors in a loop. Nothing special
that you need apart from that, I'd say.
However, it's true that while there's 'lags()' to generate list of lags
quickly, this doesn't seem to cover leads, and IMHO this would be a
useful feature.
Also, because it's impossible to enter leads via GUI into the standard
OLS-dialog, perhaps it's worth thinking about a separate DOLS entry in
the time series submenu?
cheers,
sven
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
3:33 p.m.
Am 12.11.2013 16:13, schrieb Claudio Shikida (敷田治誠 クラウジオ):
Thanks, Sven.
I just got it in R (well, but it would be great to do it in Gretl too,
at least for comparison of algorithms) but the features of lag-lead and
sandwich for the t-statistics could be two good reasons to create a time
series submenu.
BTW, so far I was only thinking about the point estimates, and there
it's just a matter of how to create the needed regressors. For the
adjustment of the t-stats of course a package that does it all-in-one
could be useful. This would actually be an excellent use case of
user-contributed function packages, now that those can also be plugged
into the menus.
I can commit to producing this in the Christmas break, if someone
reminds me ;-) Or if you cannot wait, do it yourself...
cheers,
sven
4:40 p.m.
On Tue, 12 Nov 2013, Sven Schreiber wrote:
Am 12.11.2013 16:13, schrieb Claudio Shikida (敷田治誠 クラウジオ):
> Thanks, Sven.
>
> I just got it in R (well, but it would be great to do it in Gretl too,
> at least for comparison of algorithms) but the features of lag-lead and
> sandwich for the t-statistics could be two good reasons to create a time
> series submenu.
>
>
BTW, so far I was only thinking about the point estimates, and there
it's just a matter of how to create the needed regressors. For the
adjustment of the t-stats of course a package that does it all-in-one
could be useful. This would actually be an excellent use case of
user-contributed function packages, now that those can also be plugged
into the menus.
I can commit to producing this in the Christmas break, if someone
reminds me ;-) Or if you cannot wait, do it yourself...
Maybe this could help:
<hansl>
set echo off
set messages off
function matrix dols(series y, list X, scalar order)
scalar k = nelem(X)
list extra = null
list dX = diff(X)
loop foreach i dX --quiet
loop for j=1..order --quiet
extra += $i(-j)
extra += $i(+j)
end loop
end loop
ols y X dX extra --robust --quiet
b = {1} | -$coeff[1:k]
bs = $coeff[1:k] ~ $stderr[1:k]
vn = varname(X)
printf "\n-------------- DOLS ESTIMATION --------------\n"
modprint bs vn
return b
end function
open hamilton.gdt --quiet
# log of US price level
genr p = 100*(log(PZUNEW)-log(PZUNEW[1973:01]))
# log of exchange rate
genr s = -100*(log(EXRITL)-log(EXRITL[1973:01]))
# log of Italian price level
genr pf = 100*(log(PC6IT)-log(PC6IT[1973:01]))
# Set the sample period used by Hamilton
smpl 1974:2 ;
# Estimate VECM: lag order 12, cointegration rank 1
Hamilton <- vecm 12 1 p s pf --quiet
# Estimate cointegration relationship via DOLS
list X = s pf
beta_dols = dols(p, X, 12)
print beta_dols
</hansl>
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
4:48 p.m.
Am 12.11.2013 17:40, schrieb Riccardo (Jack) Lucchetti:
On Tue, 12 Nov 2013, Sven Schreiber wrote:
>
> I can commit to producing this in the Christmas break, if someone
> reminds me ;-) Or if you cannot wait, do it yourself...
Maybe this could help:
...
loop for j=1..order --quiet
aha, gotcha -- the official syntax is w/o "for"
ols y X dX extra --robust --quiet
right, of course HAC standard errors come built-in in gretl, no extra
hassle...
-sven
4:55 p.m.
loop for j=1..order --quiet
aha, gotcha -- the official syntax is w/o "for"
;-)
I prefer to use the 'for variant'.
_______________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
5:08 p.m.
Hi,
Wow! I will test it with the same data I've used in R and, if you can, I
can report it here, in this list.
Thanks!
Claudio
Obrigado / Thanks for your time and attention.
Claudio D. Shikida
http://www.cdshikida.net and http://works.bepress.com/claudio_shikida/
Esta mensagem pode conter informação confidencial e/ou privilegiada. Se
você não for o destinatário ou a pessoa autorizada a receber esta mensagem,
não poderá usar, copiar ou divulgar as informações nela contidas ou tomar
qualquer ação baseada nessas informações. Se você recebeu esta mensagem por
engano, por favor avise imediatamente o remetente, respondendo o presente
e-mail e apague-o em seguida.
This message may contain confidential and/or privileged information. If you
are not the addressee or authorized to receive this for the addressee, you
must not use, copy, disclose or take any action based on this message or
any information herein. If you have received this message in error, please
advise the sender immediately by reply e-mail and delete this message.
On Tue, Nov 12, 2013 at 2:57 PM, Sven Schreiber <svetosch(a)gmx.net> wrote:
Am 12.11.2013 17:55, schrieb Pindar:
>
> loop for j=1..order --quiet
>
>> aha, gotcha -- the official syntax is w/o "for"
> ;-)
>
> I prefer to use the 'for variant'.
>
It may stop working at some point, however. (Not me who decides this.)
-s
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
5:34 p.m.
On Tue, 12 Nov 2013, Sven Schreiber wrote:
> loop for j=1..order --quiet
aha, gotcha -- the official syntax is w/o "for"
You're right, it's one of those habits I have to kick.
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
3817
days inactive
3817
days old
9 comments
4 participants
participants (4)
-
Claudio Shikida (敷田治誠 クラウジオ) -
Pindar -
Riccardo (Jack) Lucchetti -
Sven Schreiber