Hello. Is the Phillips-Perron test for a unit root available in Gretl?
Or is using R still the best option? (as described in the Tue Dec 15,
[Gretl-users] "Perron test for unit roots in GRETL") Thanks -- jim
-- in part, from the Dec 15 exchange:
Charles Koss schrieb:
> Ricardo, I mean the Phillips-Perron Test for Unit Roots. thanks.
Not quite the perfect answer, but if you have R installed, you can do:
*file menu -> script files -> new script -> R script
*and type (only the inner two lines ;-):
testedvar <- gretldata[,"yourvarname"]
>>> Is there an implementation of the Perron test for unit roots in
>>> on the menu, I see the following tests adf, adf-gls, and KPSS. The
>>> Perron test is not in the menu; are there reasons for not using it?
Indeed, your script works fine whatever the time series frequency may be.
That points out another question: why is the same script "2005:01")> working in monthly and quaterly data but failing in annual data ? As long as we deal with time series it should result to the same behavior regardless of the frequency.
> Message du 26/07/10 09:57
> De : "Ignacio Diaz-Emparanza"
> A : "Gretl list"
> Copie à :
> Objet : Re: [Gretl-users] genr dummy on annual data
> artur bala escribió:
> > Hello,
> > On an annual dataset spanning from 1970 to 2010, I want to create a
> > dummy valid from 2005 and on.
> > I tried the following command line :
> > * genr D = (t > "2005")*
> > but the newly created variable D has zeros only. Same result through the
> > GUI.
> > Note that this script does work fine for a monthly dataset (let's say
> > *genr D = (t > "2005:01")*).
> > best,
> > artur
> for annual data, "t" and "obs" refers to the 1-entry index. Thi is
> exactly the same as if you put a single number, genr D= t>2005, with
> quarterly or monthly data. This is indicating the observation at the
> 2005th place in the series. For the behaviour that you want, you should use
> genr D = (t > obsnum(2005))
> (see discussion in gretl-dev list some months ago, initiated at the
> http://lists.wfu.edu/pipermail/gretl-devel/2010-February/002461.html )
> Ignacio Diaz-Emparanza
> DEPARTAMENTO DE ECONOMÍA APLICADA III (ECONOMETRÍA Y ESTADÍSTICA)
> UPV/EHU Avda. Lehendakari Aguirre, 83 | 48015 BILBAO
> T.: +34 946013732 | F.: +34 946013754
> Gretl-users mailing list
Une messagerie gratuite, garantie à vie et des services en plus, ça vous tente ?
Je crée ma boîte mail www.laposte.net
I will be out of the office starting 07/26/2010 and will not return until
Hi All, I will be out of the office from July 26 through July 28, returning
to the office July 29, 2010. Please, if you have an immediate question
regarding Cost Model or Construction Economics, please email or contact
Brad A Njus, 510 625 4595, with your questions.
On an annual dataset spanning from 1970 to 2010, I want to create a
dummy valid from 2005 and on.
I tried the following command line :
* genr D = (t > "2005")*
but the newly created variable D has zeros only. Same result through the
Note that this script does work fine for a monthly dataset (let's say
*genr D = (t > "2005:01")*).
I just want to report a bug using the current cvs on ubuntu 10.04.
Loading the attached session file and running the attached script leads
to a crash with the error:
I was playing around and fitting an AR(2) model using the yearly
sunspots data provided by R data(sunspots), actually with more years
than in the R default dataset, and I was curious as to how the roots
of the AR coefficients are computed in gretl?
When I run
arima 2 0 0 ; SUNACTIVITY --conditional
I get a constant and phi_1 = 1.39181 and phi_2 = -.690287
IIUC, the roots to check for stability should be the solution to the
characteristic polynomial, which is in this case, X^2 - 1.39181 * X +
.690287. Using MATLAB's (or NumPy's) roots function, I find that the
roots of this equation are both
But gretl says they are both 1.0081-.6575j, which is not a root of the
above. Am I misunderstanding something? If so, could someone point
to a reference?
PS. I can provide the data if needed.
Use a script and run the following example:
vecm 2 2 X
Normally the script should continue without errors until the last
restriction is tested. But adding the "-full" option for the second
restriction results in no proper testing for the third case any more.
Every time one has to reestimate the model again before one can test the
third restriction. Is this programmed that way on purpose?
Hello gretl community,
I am just working on a VECM and try to impose long-run restrictions on
beta. In order to test for the over-identifying restrictions I would
like to apply a bootstrap LR test to overcome the small sample bias issue.
I am wondered whether anybody of you knows an available free software
which can handle this issue? Unfortunately, I did not find anything.
Thanks in advance!
I found that if I use a shorter subsample, for example, running the
smpl 1988/01/04 1989/12/29
gf1 <- gnuplot djclose --time-series --output=display --with-lines
By examining the gnuplot command, it seems ok as after editing the gnuplot
set xtics (....)
After the "set xtics" command, there are some Chinese characters which look
correct. However, these Chinese characters are not appropriately shown in
the graph windows. I realize that this might be caused by the translation of
Do you have any suggestions?
How are information criteria (e.g. AIC) calculated for VAR and VECM? For
univariate models, the formula is given on page 151 in the
documentation, but what about multivariate models? I can’t find any
documentation in the manual.
Cheers, Torben Schmith
Center for Ocean and Ice
Danish Meteorological Institute
DK-2100 Copenhagen Ø
Phone direct : +45 39 15 74 44
Phone switch : +45 39 15 75 00
Fax : +45 39 15 73 00
E-mail : ts(a)dmi.dk
Homepage : http://ocean.dmi.dk/staff/ts/ts.html