wishes
by Ignacio Díaz-Emparanza
I would to ask for two new features:
-Complete the CUSUM test (and graphic) with the CUSUMSQ (CUSUM of Squares)
test. See Chapter 7 in Greene's Book or page 155 in Harvey's "The
Econometrics Analysis of Time Series" (The table for c_0 is here),
-Provide a command (similar to "corr") to calculate a cross-correlogram
between two series.
--
Ignacio Díaz-Emparanza
Dpto. de Economía Aplicada III (Econometría y Estadística)
UPV-EHU
18 years
Bug
by Ignacio Díaz-Emparanza
I think there is a bug in Gretl spreadsheet. If I follow these steps Gretl
exits:
(tested on Windows XP and MandrivaLinux 2006 with official gretl.1.6.0)
1- open a new dataset (5 obs of an annual time series, for example)
2- In the spreadsheet I write the values of the "Y" variable, and then "apply"
3- I click on variable/add and write the values for X2, and then "apply"
4- I click on variable/add again to write the values of X3, and then cliccking
on "apply" Gretl exits. In Linux there is no warning message, in Windows
there is a warning saying that the system has detected a problem.
If I write the values of the variables and click on "apply" only at the end,
ther is no problem.
--
Ignacio Díaz-Emparanza
Dpto. de Economía Aplicada III (Econometría y Estadística)
UPV-EHU
18 years
Re: [Gretl-users] accessing $vcv etc.
by Riccardo (Jack) Lucchetti
On Fri, October 13, 2006 15:22, Allin Cottrell wrote:
>> I can see the point, although personally I use (or would like to
>> use) gretl differently.
>
> I can see a case for having the behaviour that you expected (i.e.
> estimating a model via the gui updates the "last model" as seen
> from the console). That wouldn't be difficult to do.
>
Hmm. I don't know. I'm not able to say precisely why, but I'm not comfortable
with the idea that the result of a script may be influenced by something
outside the script itself. Sven, could you give us a little more detail as to
what your plan is?
(cc:ing this to the devel list, because IMO this discussion really belongs there)
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Facoltà di Economia "G. Fuà"
Ancona
18 years, 1 month
Re: [Gretl-users] function problems
by Sven Schreiber
Allin Cottrell schrieb:
> On Fri, 13 Oct 2006, Sven Schreiber wrote:
>
>> alphatest2 (myorder, myrank, myvar, myconsttrend, myboolseasonals)
>
> This should now work with CVS.
>
> Allin
>
Yes thanks, it does.
I started to write a function to test the variables in the VECM for weak
exogeneity, as I "promised" a while ago. Basically what's in Johansen's
book on p. 124-128 or so. It's really just porting an algorithm I have
already in Numerical Python, so it's actually not too much work. That's
why I asked for $s00 etc.
The problem is, I've just been reminded by my code that apart from $sxx
one also needs the R0 and R1 residuals after "concentrating" the system.
Now I'm not sure whether I really should ask you to expose those as
retrievable variables as well, I have the feeling this is getting a bit
out of hand. OTOH, I would really like to avoid to recode the lagging,
differencing, and concentrating of the system/variables, because I'm so
glad I have that set up in my python code once and for all; it really
isn't fun.
Any ideas how to solve this dilemma?
Thanks,
Sven
18 years, 1 month
function problems
by Sven Schreiber
Hi,
I'm currently trying to use the cool function-package features of gretl.
In the process I'm having some problems (with the cvs version from
yesterday).
Illustration:
1. start gretl
2. load example data2-3.gdt
3. execute the following function definition (without the email
linebreaks, of course):
function alphatest2 (int order, int rank, list endo, int consttrend,
bool seasonals)
vecm order rank endo
end function
4. Then execute the following script:
genr myorder = 2
genr myrank = 1
list myvar = unemp infl
genr myconsttrend = 3
genr myboolseasonals = 0
alphatest2 (myorder, myrank, myvar, myconsttrend, myboolseasonals)
... which sends gretl in an infinite loop it seems, so it has to be killed.
Experimentation showed that a function with the command:
vecm order 1 endo
works well.
And something totally unrelated: In the command reference index (text
format), when I click on genr, I get matrix.
thanks,
Sven
18 years, 1 month
Re: VECM estimation: gretl vs JMulTi
by Sven Schreiber
Riccardo (Jack) Lucchetti schrieb:
> On Thu, October 12, 2006 12:43, Sven Schreiber wrote:
>> Well I just tried because I wanted to use the $s00 johansen stuff.
>> ./configure and make work fine; however, I get the following error from
>> running checkinstall ...
>
> Sorry I don't have time to look into this properly at the moment, but I have 1
> quick suggestion: try running configure with the --enable-build-doc switch, as
> in
>
> ./configure --enable-build-doc && make
>
> Not sure if it's relevant, but it may help.
>
Thanks, that did the trick!
Now trying to access $s00...
-sven
18 years, 1 month
Re: VECM estimation: gretl vs JMulTi
by Sven Schreiber
Riccardo (Jack) Lucchetti schrieb:
> Good to know you switched. I hope you join the "build-from-CVS" group soon.
>
>
Well I just tried because I wanted to use the $s00 johansen stuff.
./configure and make work fine; however, I get the following error from
running checkinstall (instead of make install, as I probably don't need
to tell you):
...
> true_fopen == 0 for fopen64("/proc/mounts", "r")
> true_fopen == 0 for fopen64("/proc/mounts", "r")
> cp -fp ../share/functions/gretlfunc.dtd /usr/local/share/gretl/functions
> true_fopen == 0 for fopen64("/proc/mounts", "r")
> ../tools/mkinstalldirs /usr/local/share/gretl
> ../tools/mkinstalldirs /usr/local/share/gretl/doc
> ../tools/mkinstalldirs /usr/local/share/gretl/helpfigs
> cp -fp ../share/gretlgui.hlp /usr/local/share/gretl/gretlgui.hlp
> true_fopen == 0 for fopen64("/proc/mounts", "r")
> cp: Aufruf von stat für „../share/gretlgui.hlp“ nicht möglich: No such
file or d irectory
> make[1]: *** [install_help] Fehler 1
...
I built gretl 1.6.0 successfully from the tarball with the same steps (I
think), so what's wrong now?
In configure, "build documentation" is set to "no", is that a problem?
And initially, I hadn't removed the 1.6.0 version yet when trying to
checkinstall the first time. Was that Ok?
This is all on ubuntu 6 (dapper).
any hints appreciated,
Sven
18 years, 1 month
Re: VECM estimation: gretl vs JMulTi
by john w
But there is still another uncommon thing in JMulTi....Johansen procedure
doesn't allow restrictions on beta! You can test them but you can't impose
them.
The good thing is that when you use gretl and JMulTi together you get a
killind data machine.
Fortunatelly, gretl has a great possibilitis to import and export datas, so
this two softwares can be used as one.
This is why I'm opting for vecm estimation in gretl with allowed imposing
restrictions on beta.
In general, JMulTi forum is a little bit "dead" in past few months. So, it
is not so easy to reach the developers.
Riccardo Jack Lucchetti schrieb:
>On Thu, August 31, 2006 12:07, Sven Schreiber wrote:
>
>>Ok I can reproduce that. As I suspected, the difference seems to relate to
>>deterministic terms, namely the treatment of seasonals.
>>
>>It very much looks like the seasonals in Jmulti are not centered, which
>makes > a difference if you have a restricted constant, as you know.
>>So this is a problem of Jmulti which is not nice at all, and thank you for
>>pointing it out. I'll contact Markus (the jmulti man) about it, but I
>>think
>>he's on vacation right now.
>
>Well spotted.
>
fyi, starting with version 4.13 Jmulti now also uses centered seasonals
in vecm estimation (it always had done so in the cointegration test
module). Thank me for annoying Markus until he gave up ;-)
-sven
_________________________________________________________________
Windows Live Messenger has arrived. Click here to download it for free!
http://imagine-msn.com/messenger/launch80/?locale=en-gb
18 years, 1 month
Re: VECM estimation: gretl vs JMulTi
by Sven Schreiber
Riccardo (Jack) Lucchetti schrieb:
> On Tue, October 10, 2006 09:57, Sven Schreiber wrote:
>> fyi, starting with version 4.13 Jmulti now also uses centered seasonals
>> in vecm estimation (it always had done so in the cointegration test
>> module). Thank me for annoying Markus until he gave up ;-)
>>
>
> That's good to know. Unfortunately, I'm stuck with an older version (4.04)
> because 4.12 complains about the java version I have on my machine (Sun Java
> from debian testing). I must find the time to tell the JMulTi people about
> this.
>
>
I think the key is that the official min requirement is JRE 5 (alias
1.5, as opposed to <= 1.4.x), which works ok on Ubuntu (did I say
already I switched?), but don't know if it's in debian due to license
issues. If your Java is 1.4, I guess the Jmulti people (==Markus) don't
really care.
-sven
18 years, 1 month
Re: [Gretl-users] VECM estimation
by Riccardo Jack Lucchetti
Sven, I'm cc-ing this to the developer list because I think it's the best
place for this discussion
On Wed, August 30, 2006 21:31, Sven Schreiber wrote:
>> the matrices S00 & co. are quite easy to build by combining lists, loop, genr
>> and matrix functions; all in all, you need moment matrices of residuals. On
>
> Sure, but dealing with sample lengths, lagged differences, exogenous
variables etc. is not fun and it's already implemented in a nice dialog, so
why not use it?
Well, because IMO it doesn't "feel" like a good idea to have too many
fetchable variables post-estimation. It's a balancing act here: a few things
would be simple to do in a script, but you're likely to use them a lot (think
$uhat), some other things would be more cumbersome (example: $jbeta) so we
might as well make them fetchable once estimation is over for the user's
convenience, even if you're unlikely to use them except in few special
situations. In my mind S00 & co. belong to neither group, but I'm open to be
convinced of the contrary.
>> Note that we can already do this via a LR test.
>
> Where is that documented?
Well, it seems it isn't. :-(
But basically, this is what "restrict" does after "vecm"; the corresponding
GUI element is "Test/Linear restriction" from the VECM output window.
It looks like the manual needs to be beefed up a little. Actually, a chapter
on hypothesis testing wouldn't hurt.
>
>> While I tested this, I realised that I couldn't reproduce Johansen's
results
>> with JMulTi; either I'm missing something, or JMulTi's VECM estimation goes
through a different algorithm from ours. We do VECMs as per Johansen's
book;
>> so does PcGive, and I believe there should be no differences between our
results and what PcGive yields.
>
> I have all those programs, do you want me to test/confirm something?
Please try to reproduce the Denmark example you find in Johansen's book. The
data are available as denmark.gdt among gretl example datasets. JMulTi seems
to do strange things with VECMs. I exported the denmark dataset to JMulTi
format via the new ad-hoc facility: then, I could reproduce the same results
we have in gretl for almost everything (descriptives, VARs, etc.) but VECMs
proved impossible. I understand you know the main developer of JMulTi
personally: it may be a good idea to ask his opinion if/when we're certain
that results are indeed different.
If you could do the same with MicroFit too, that'd be great.
Riccardo "Jack" Lucchetti
Dipartimento di Economia
Facoltà di Economia "G. Fuà"
Ancona
18 years, 1 month
