Gretl-Hansl "IDE" for Sublime editor
by Artur Tarassow
Dear all,
some weeks ago I've started to switch to the sublime editor
(https://www.sublimetext.com/) for writing gretl code.
Even though I like the simplicity of the gretl editor and its features
such as syntax highlighting and auto-itendation, it lacks some features
of modern IDEs such as "goto-anything", "goto-definition", custom
keybindings, fancy themes, git-implementation, snippets etc. which make
life much easier when working on larger projects. Don't get me wrong,
the gretl editor is great but was _never_ supposed to become a proper
software-development IDE but rather has another focus which is totally fine.
So I started to write the "Hansl-Gretl-Language" package for sublime
which includes the following features:
- 3 gretl build-systems (client mode, batch mode, and REPL mode) for
executing hansl code by means of sublime (plots are also working!)
- syntax-highlighting
- completion of gretl commands, accessors and keywords
- some snippet examples for speeding up coding
The project still has the following (known) issues:
- no auto-itendation (still have to figure out how this works)
- issues with some corner-cases which are not syntax-highlighted (regex
can become so hard!)
The package can be downloaded through sublimes package control system,
and can be found here:
https://packagecontrol.io/packages/Hansl-Gretl-Language
If somebody wants to participate on this project, check out the code on
my github repository:
https://github.com/atecon/Hansl-Gretl-Language
For those interesting in the sublime editor, check out "OdatNurd"'s
brilliant tutorials on youtube:
https://www.youtube.com/user/nurdz
Enjoy the package,
Artur
1 year, 11 months
BACE Estimation difficulties
by Brian Revell
I thought i would try this option estimating a simple model (always the
best way to gain experience and confidence with the routine).
Independent variable, 1 indep variable . No lag on Y. . Zero selected for
out of sample forecast Constant -can be omitted. Best 4 models selected
-again as the default.
This was the output:-
Dependent viariable was removed from the X list.* It is not there -there is
only a single indep variable. *
Error message from BACE_GUI():
Average model size should be greater then 0 and lower than K !!! *Average
model size specified is the default 1*
*Some guidance wlecomed please*
*Brian*
2 years, 10 months
Gretl wiki
by Federico Fiorani
Dear all,
I would like to know your opinion about opening a wiki for Gretl.
I heard that a test had been done but the wiki was closed shortly
afterwards.
In my opinion, a wiki could be very useful, for example, to supplement
what cannot be in the documentation or manual, such as examples of
scripts, or to explore topics that may not be entirely clear at first
glance and It would also be useful for those who are new to the software.
Regards
Federico
2 years, 12 months
Re: small Johansen test - minimum sample size
by Sven Schreiber
Am 30.10.2022 um 08:53 schrieb Lars Ahnland:
> According to Lütkepohl and Krätzig (2004, Applied Time Series
> Econometrics, p. 89) it is viable to relax the assumption that all
> variables should be I(1) in levels so that the concept of
> cointegration can is extended by including any stationary linear
> combination (providing variables are I(0) in first differences).
>
Exactly, and what we're saying is that in general a linear combination
can also mean to just pick one element by having a unit vector as the
cointegration vector in this broader sense. This then would not mean
that really two variables are cointegrated in the narrower sense.
> ------------------------------------------------------------------------
> *Från:* Cottrell, Allin <cottrell(a)wfu.edu>
>
> No, what Sven is pointing out is a common "gotcha" with the Johansen
> test: one of the series in the llst proposed for testing is I(0). This
> in itself does _not_imply that there's any cointegration going on.
To be clear: I don't _know_ that one series is I(0). Nor does Lars
_know_ that it is I(1). The evidence is just not super clear, given the
small sample size. So the test result could be a reflection of the
capital share being I(0), or there could alternatively be genuine
cointegration where the debt ratio enters with a small coefficient.
(Check the estimates of the cointegration vector under rank 1.)
So, coming back to the original question which is only a little
gretl-related in the sense of clarifying if the "johansensmall" package
by Andreas Noack Jensen and myself can be used with 30 observations or
so: Yes, I think it's possible and the results are not "wrong". Of
course, the bootstrap approach never promises to achieve _exactly_ the
nominal test size (for example, 5% rejection under the null). We just
hope that it corrects the size distortion in small samples _to some
extent_. At the same time you will lose test power. So it's still the
case that the uncertainty surrounding the test result gets bigger when
you have less data, and you definitely have to interpret the results
with caution.
cheers
sven
3 years
Re: small Johansen test - minimum sample size
by Cottrell, Allin
On Sat, Oct 29, 2022 at 3:27 PM Lars Ahnland <lars.ahnland(a)outlook.com> wrote:
>
> Also, what do you mean by only "one component is stationary". The point is that residuals are stationary with cointegration?
No, what Sven is pointing out is a common "gotcha" with the Johansen
test: one of the series in the llst proposed for testing is I(0). This
in itself does _not_imply that there's any cointegration going on.
Allin Cottrell
3 years
Re: small Johansen test - minimum sample size
by Sven Schreiber
Am 28.10.2022 um 16:42 schrieb Lars Ahnland:
> Hi,
>
> Sorry, I used the two variables Debt_ratio and Capital_share, with a
> restriced contant model. The lag length, according to a majority of
> information criteria, is 2.
>
OK, I tried your first subsample (BTW, I found it quite nice how gretl
directly imported the series from the xlsx file without a hiccup...),
and I also get a p-value below 5%. Remember, however, that this just
means that the test finds one stationary component ( = "direction") in
the system. By eyeballing that series I would bet that the capital share
is an I(0) variable, at least in the perspective of this subsample. So
yes, I'd say that the result is reliable, but it may not mean what you
were expecting.
cheers
sven
3 years
Re: small Johansen test - minimum sample size
by Sven Schreiber
Am 28.10.2022 um 15:11 schrieb Lars Ahnland:
> Hi again, I discovered that I had confused the significance level. It
> should be 5% significance in all cases, both bartlett and
> boot-strapped p-values (1896-1929 and 1982-2008).
Well, this already makes more sense. Apart from that, you forgot to give
your exact specification; at the very least which variables you are
including in the test. For example, I see that there is a column "t" in
your datafile, which represents a time trend. Such a variable should of
course not be treated as a standard endogenous one. If you did, then the
answer is that the result is invalid.
cheers
sven
3 years
Re: small Johansen test - minimum sample size
by Lars Ahnland
Hi again, I discovered that I had confused the significance level. It should be 5% significance in all cases, both bartlett and boot-strapped p-values (1896-1929 and 1982-2008).
Best,
Lars
________________________________
Från: Lars Ahnland <lars.ahnland(a)outlook.com>
Skickat: den 28 oktober 2022 13:10
Till: gretl-users(a)gretlml.univpm.it <gretl-users(a)gretlml.univpm.it>
Ämne: [Gretl-users] Re: small Johansen test - minimum sample size
Hi Sven,
Thanks for the reply! Here are the excel files (datafile "6 + China 1896-2008"; and results). The datafile is a weigthed average of two variables for a number of countries. It is the sheet 1 I have used with Gretl. As you can see, I have used sub-sets of the time series, referring to specifiuc macroeconomic policy regimes, in 1896-1929, 1946-1973 and 1982-2008.
Best regards,
Lars
________________________________
Från: Sven Schreiber <sven.schreiber(a)fu-berlin.de>
Skickat: den 27 oktober 2022 16:02
Till: gretl-users(a)gretlml.univpm.it <gretl-users(a)gretlml.univpm.it>
Ämne: [Gretl-users] Re: small Johansen test - minimum sample size
Am 27.10.2022 um 12:44 schrieb Lars Ahnland:
Hi,
I am using the package johansensmall, for johansen cointegration test with small sample sizes. I have a dataset with only 29 samples. Now, I get a significant reslut at the 5% level with the Bartlett correction, and at 1% level with the boot-strapped p-values. I report both values. Would you say these results are valid, considering the very small sample size?
Hi Lars,
indeed I'd say it's unusual that the bootstrap version rejects the null at a stricter level than the Bartlett correction. For such a small sample size I would typically expect that the bootstrap test is unable to reject. If you could share this particular dataset and your test specification (deterministics, lags), maybe we can have a better idea of what's going on.
cheers
sven
3 years
Re: small Johansen test - minimum sample size
by Sven Schreiber
Am 27.10.2022 um 12:44 schrieb Lars Ahnland:
> Hi,
>
> I am using the package johansensmall, for johansen cointegration test
> with small sample sizes. I have a dataset with only 29 samples. Now, I
> get a significant reslut at the 5% level with the Bartlett correction,
> and at 1% level with the boot-strapped p-values. I report both values.
> Would you say these results are valid, considering the very small
> sample size?
>
Hi Lars,
indeed I'd say it's unusual that the bootstrap version rejects the null
at a stricter level than the Bartlett correction. For such a small
sample size I would typically expect that the bootstrap test is unable
to reject. If you could share this particular dataset and your test
specification (deterministics, lags), maybe we can have a better idea of
what's going on.
cheers
sven
3 years
small Johansen test - minimum sample size
by Lars Ahnland
Hi,
I am using the package johansensmall, for johansen cointegration test with small sample sizes. I have a dataset with only 29 samples. Now, I get a significant reslut at the 5% level with the Bartlett correction, and at 1% level with the boot-strapped p-values. I report both values. Would you say these results are valid, considering the very small sample size?
Best,
Lars
3 years
