guido giaume | 2 May 19:45 2016
Picon

yahoo_get.gfn: ticker(s) not recognized?


Dear all
sincere thanks for your efforts in developing Gretl.


Here my question. I was unable to pick a hint looking at the past discussions.


It seems to me - sometimes - yahoo_get.gfn does not recognize the "right" ticker
[tested only on Linux]


this is the gretl output

--------------------------------------------------------------------------
gretl versione 2016a
Sessione corrente: 2016-05-02 19:02

# Apro un set di 100 giornate con
? nulldata 100
Periodicità: 1, oss. max.: 100
Intervallo delle osservazioni: 1-100

# 5 rilevazioni settimanali che iniziano il 1 gennaio 2015 e indico che è una serie temporale
? setobs 5 2016-01-01 --time-series
Campione completo dei dati: 2016-01-01 - 2016-05-19 (n = 100)

# chiama la funzione yahoo_get
? include yahoo_get.gfn
 /home/guido/.gretl/functions/yahoo_get.gfn

# scarico i titoli
? loop foreach i FP.PA UL.PA FR.PA
> A=yahoo_full("$i")
> endloop

? A=yahoo_full("FP.PA")
FP.PA imported ok (full)

? A=yahoo_full("UL.PA")
Ticker UL.PA not found [/home/guido/.gretl/yg_0785.csv]

? A=yahoo_full("FR.PA")
FR.PA imported ok (full)

Numero di iterazioni: 3

---------------------------------------------------------

As you can see UL.PA is a valid ticker

It seems a similar behavior happened also with other stocks (some ticker of the LSE).


Could somebody explain me if i made something wrong?

Many thanks

Guido




<div>
<div><br></div>
<div>Dear all</div>
<div>sincere thanks for your efforts in developing Gretl.</div>
<div><br></div>
<div><br></div>
<div>Here my question. I was unable to pick a hint looking at the past discussions.</div>
<div><br></div>
<div><br></div>
<div>It seems to me - sometimes - yahoo_get.gfn does not recognize the "right" ticker</div>
<div>[tested only on Linux]</div>
<div><br></div>
<div><br></div>
<div>this is the gretl output</div>
<div><br></div>
<div>--------------------------------------------------------------------------</div>
<div>gretl versione 2016a</div>
<div>Sessione corrente: 2016-05-02 19:02</div>
<div><br></div>
<div># Apro un set di 100 giornate con</div>
<div>? nulldata 100</div>
<div>Periodicit&agrave;: 1, oss. max.: 100</div>
<div>Intervallo delle osservazioni: 1-100</div>
<div><br></div>
<div># 5 rilevazioni settimanali che iniziano il 1 gennaio 2015 e indico che &egrave; una serie temporale</div>
<div>? setobs 5 2016-01-01 --time-series</div>
<div>Campione completo dei dati: 2016-01-01 - 2016-05-19 (n = 100)</div>
<div><br></div>
<div># chiama la funzione yahoo_get</div>
<div>? include yahoo_get.gfn</div>
<div>&nbsp;/home/guido/.gretl/functions/yahoo_get.gfn</div>
<div><br></div>
<div># scarico i titoli</div>
<div>? loop foreach i FP.PA UL.PA FR.PA</div>
<div>&gt; A=yahoo_full("$i")</div>
<div>&gt; endloop</div>
<div><br></div>
<div>? A=yahoo_full("FP.PA")</div>
<div>FP.PA imported ok (full)</div>
<div><br></div>
<div>? A=yahoo_full("UL.PA")</div>
<div>Ticker UL.PA not found [/home/guido/.gretl/yg_0785.csv]</div>
<div><br></div>
<div>? A=yahoo_full("FR.PA")</div>
<div>FR.PA imported ok (full)</div>
<div><br></div>
<div>Numero di iterazioni: 3</div>
<div><br></div>
<div>---------------------------------------------------------</div>
<div><br></div>
<div>As you can see UL.PA is a valid ticker</div>
<div><a href="http://finance.yahoo.com/q/hp?s=UL.PA+Historical+Prices">http://finance.yahoo.com/q/hp?s=UL.PA+Historical+Prices</a></div>
<div><br></div>
<div>It seems a similar behavior happened also with other stocks (some ticker of the LSE).</div>
<div><br></div>
<div><br></div>
<div>Could somebody explain me if i made something wrong?</div>
<div><br></div>
<div>Many thanks</div>
<div><br></div>
<div>Guido</div>
<div><br></div>
<div><br></div>
<div><br></div>
<div><br></div>
<div></div>
</div>
Fábio Aloísio | 2 May 18:52 2016
Picon

ARMA (1,1)

Good afternoon,

Sorry for any inconvenience. I'm trying to forecast the ARMA (1,1) process out of sample (X t+1), but it shows a message that this cannot be done (only in sample fcast). I tried everything, including some commands - fcast,  --out-of-sample, dataset addobs 1, --dynamic. I also tried to ad new observations via the graphical interface, to generate forecast, but, without success either.

There is something that I'm not doing right. Can anyone please give a thought, what goes wrong? Maybe my commands are inadequate.

Tanks in advance, for any valuable help.

My best,

Fabio

ps. I know I can use the coeficients in spreadsheet to generate the fcast, but Im trying to figure out whats is happening. 
<div><div dir="ltr">
<div>
<div>
<div>
<div>
<div>Good afternoon,<br><br>Sorry for any inconvenience. I'm trying to forecast the ARMA (1,1) process out of sample (X t+1), but it shows a message that this cannot be done (only in sample fcast). I tried everything, including some commands - fcast,&nbsp; --out-of-sample, dataset addobs 1, --dynamic. I also tried to ad new observations via the graphical interface, to generate forecast, but, without success either.<br><br>
</div>There is something that I'm not doing right. Can anyone please give a thought, what goes wrong? Maybe my commands are inadequate. <br><br>
</div>Tanks in advance, for any valuable help.<br><br>
</div>My best,<br><br>
</div>Fabio <br><br>
</div>ps. I know I can use the coeficients in spreadsheet to generate the fcast, but Im trying to figure out whats is happening.&nbsp; <br>
</div></div>
Schaff, Frederik | 2 May 11:51 2016
Picon

feature request: (gnu)plot -- options that cannot be controlled (yet)

Hello, 

currently (the snapshot) one can modify almost anything of a plot by streaming it as literals (including
manipulation of the linetypes), which is very nice. But there are some things which I for once always need
to post-process manually:

1) The way in which the lines are plotted (i.e. dots, impulses, ...) 
2) and the titles for the legend

For 1) it is at least possible to specify some options, using i.e. "--with-lines=series1", but not all options.

Perhaps it is possible, in future, to provide an additional command to allow for literals that are
substituted into the 
---------
plot \
'-' using 1:($2) XYZ 
---------
part directly for XYZ? This would further reduce the necessity to post-process files.

- Frederik
------------
Frederik Schaff, Dipl.-Volkswirt

University of Hagen
Department of Business and Economics
Chair of Economic Theory (Prof. A. Endres)
Universitätsstraße 11 (TGZ)
58097  Hagen
 
Phone: +49 (0) 2331 987-4454
E-Mail: Frederik.Schaff@...
http://www.fernuni-hagen.de/wirtschaftstheorie/en/team/frederik.schaff.shtml 

cociuba mihai | 1 May 15:02 2016
Picon

can't install IVpanels addon (Ubuntu 16.04, gretl source build)

I've just build gretl from sources (with ./configure --enable-build-doc --enable-openmp --enable-build-addons) on a Ubuntu 16.04 and I noticed 2 things:
- when trying to install the IVPANEL addon gretl shows and error: 
<html>
 <head>
  <title>404 Not Found</title>
 </head>
 <body>
  <h1>404 Not Found</h1>
  The resource could not be found.<br /
from the terminal it shows that 
"Couldn't find ivpanel for gretl 2016.-1252336200.32767"
the same for SVAR package
"Couldn't find SVAR for gretl 2016.-1252336200.32767."

- also in the terminal the following message appears: 
"Disabling OpenBLAS multi-threading (OpenMP/pthreads collision)"
I had some message regarding multi-threading but I clicked Don't show anymore.

Any ideas on what may be wrong.

Also I would like to know if Guilherme test scripts will be merge in gretl git because it would be a great opportunity to the study gretl behavior on different machines and OS. I've noticed that there is a NIST test suite and the two of them can be merged?


Mihai 

<div><div dir="ltr">
<div class="gmail_extra">I've just build gretl from sources (with ./configure --enable-build-doc --enable-openmp --enable-build-addons) on a Ubuntu 16.04 and I noticed 2 things:</div>
<div class="gmail_extra">- when trying to install the IVPANEL addon gretl shows and error:&nbsp;</div>
<div class="gmail_extra">
<div class="gmail_extra">&lt;html&gt;</div>
<div class="gmail_extra">&nbsp;&lt;head&gt;</div>
<div class="gmail_extra">&nbsp; &lt;title&gt;404 Not Found&lt;/title&gt;</div>
<div class="gmail_extra">&nbsp;&lt;/head&gt;</div>
<div class="gmail_extra">&nbsp;&lt;body&gt;</div>
<div class="gmail_extra">&nbsp; &lt;h1&gt;404 Not Found&lt;/h1&gt;</div>
<div class="gmail_extra">&nbsp; The resource could not be found.&lt;br /</div>
<div class="gmail_extra">from the terminal it shows that&nbsp;</div>
<div class="gmail_extra">"Couldn't find ivpanel for gretl 2016.-1252336200.32767"<br>
</div>
<div class="gmail_extra">the same for SVAR package</div>
<div class="gmail_extra">"Couldn't find SVAR for gretl 2016.-1252336200.32767."<br>
</div>
<div class="gmail_extra"><br></div>
<div class="gmail_extra">- also in the terminal the following message appears:&nbsp;</div>
<div class="gmail_extra">
<div class="gmail_extra">"Disabling OpenBLAS multi-threading (OpenMP/pthreads collision)"</div>
<div class="gmail_extra">I had some message regarding multi-threading but I clicked Don't show anymore.</div>
<div class="gmail_extra"><br></div>
<div class="gmail_extra">Any ideas on what may be wrong.</div>
<div class="gmail_extra"><br></div>
<div class="gmail_extra">Also I would like to know if Guilherme test scripts will be merge in gretl git because it would be a great opportunity to the study gretl behavior on different machines and OS. I've noticed that there is a NIST test suite and the two of them can be merged?</div>
<div class="gmail_extra"><br></div>
<div class="gmail_extra"><br></div>
<div class="gmail_extra">Mihai&nbsp;</div>
<div><br></div>
</div>
</div>
</div></div>
Wingenroth, Thorsten | 1 May 10:14 2016
Picon

plot command not working

Hi,

 

I am using gretl 2016b-git on Windows (x86_64) with build date 2016-02-02.

 

I ran the following script:

 

scalar anzahl = 1000

nulldata anzahl

setobs 1000 1:0001 --special-time-series

x = normal(0,1)

# gnuplot x --time-series --output=display

plot x --time-series --output=display

 

Nothing is displayed. If I run the gnuplot command (shown as comment) instead, I get a proper picture. What is my mistake using “plot”?

 

Thank you and kind regards,

 

Thorsten

 

<div>
<div class="WordSection1">
<p class="MsoNormal"><span lang="EN-US">Hi,<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">I am using gretl 2016b-git on Windows (x86_64) with build date 2016-02-02.<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">I ran the following script:<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">scalar anzahl = 1000<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">nulldata anzahl<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">setobs 1000 1:0001 --special-time-series<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">x = normal(0,1)<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"># gnuplot x --time-series --output=display<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">plot x --time-series --output=display<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">Nothing is displayed. If I run the gnuplot command (shown as comment) instead, I get a proper picture. What is my mistake using &ldquo;plot&rdquo;?<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">Thank you and kind regards,<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">Thorsten<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
</div>
</div>
Wingenroth, Thorsten | 1 May 10:09 2016
Picon

setobs using variables

Hi,

 

I am using gretl 2016b-git on Windows (x86_64) with build date 2016-02-02.

 

I ran the following script:

 

scalar anzahl = 1000

nulldata anzahl

setobs anzahl 1:0001 --special-time-series

 

The log script follows:

 

? scalar anzahl = 1000

Generated scalar anzahl = 1000

? nulldata anzahl

periodicity: 1, maxobs: 1000

observations range: 1 to 1000

? setobs anzahl 1:0001 --special-time-series

'anzahl': not a scalar

starting obs '1.0001' is incompatible with frequency

 

Error executing script: halting

> setobs anzahl 1:0001 --special-time-series

 

 

Why is anzahl not a scalar? I explicitly declared it to be so.

 

How can I make the first observation 1:0001 depending on anzahl? Should I build a string?

 

Thank you and kind regards,

 

Thorsten

 

<div>
<div class="WordSection1">
<p class="MsoNormal">Hi,<p></p></p>
<p class="MsoNormal"><p>&nbsp;</p></p>
<p class="MsoNormal"><span lang="EN-US">I am using gretl 2016b-git on Windows (x86_64) with build date 2016-02-02.<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">I ran the following script:<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">scalar anzahl = 1000<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">nulldata anzahl<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">setobs anzahl 1:0001 --special-time-series<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">The log script follows:<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">? scalar anzahl = 1000<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">Generated scalar anzahl = 1000<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">? nulldata anzahl<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">periodicity: 1, maxobs: 1000<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">observations range: 1 to 1000<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">? setobs anzahl 1:0001 --special-time-series<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">'anzahl': not a scalar<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">starting obs '1.0001' is incompatible with frequency<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">Error executing script: halting<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&gt; setobs anzahl 1:0001 --special-time-series<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">Why is anzahl not a scalar? I explicitly declared it to be so.<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">How can I make the first observation 1:0001 depending on anzahl? Should I build a string?<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">Thank you and kind regards,<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">Thorsten<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
</div>
</div>
Hélio Guilherme | 28 Apr 18:35 2016
Picon

About (Unstable regression results)

As a work in progress we have Regression Tests (Software Development Domain Language ) being set up at https://github.com/HelioGuilherme66/gretl/tree/master/gretl-tests/test-gretl/.

There are some missing gretl files (which I hope Allin can provide), there are no missing applications detection, and structure is designed only for Linux.

From these 900 gretl test scripts I have only Failing 34 (some due to missing file and missing Stata). There are some minor numeric differences, when comparing the outputs with Allin's provided tests results.

I would like to have more people running the tests, but first would like to have fixed the missing files, and creating a common testing structure to Windows, Mac OS X and Linux.

In Linux you will have to do some minor edits:
- correct paths in bin/sitevars
- edit BASE2 variable in bin/refactor_data

I am posting to the user list, because I suspect that my emails are being blocked to devel.

Best Regards,
Hélio



<div><div dir="ltr">
<div>
<div>
<div>
<div>
<div>As a work in progress we have Regression Tests (Software Development Domain Language ) being set up at <a href="https://github.com/HelioGuilherme66/gretl/tree/master/gretl-tests/test-gretl/">https://github.com/HelioGuilherme66/gretl/tree/master/gretl-tests/test-gretl/</a>.<br><br>
</div>There are some missing gretl files (which I hope Allin can provide), there are no missing applications detection, and structure is designed only for Linux.<br><br>From these 900 gretl test scripts I have only Failing 34 (some due to missing file and missing Stata). There are some minor numeric differences, when comparing the outputs with Allin's provided tests results.<br><br>
</div>I would like to have more people running the tests, but first would like to have fixed the missing files, and creating a common testing structure to Windows, Mac OS X and Linux.<br><br>
</div>
<div>In Linux you will have to do some minor edits:<br>
</div>
<div>- correct paths in bin/sitevars<br>
</div>
<div>- edit BASE2 variable in bin/refactor_data<br>
</div>
<div><br></div>I am posting to the user list, because I suspect that my emails are being blocked to devel.<br><br>
</div>Best Regards,<br>
</div>H&eacute;lio<br><div><div><div>
<br><br><div><div><div><br></div></div></div>
</div></div></div>
</div></div>
Mikael Postila | 27 Apr 15:53 2016

question - Unstable regression results

Hi

 

We’ve been experiencing a problem where exactly same data yields different regression results depending on A) which computer is used and B) when regression is being run on same computer.

This is the third time the problem arises (first time was in February, next was in March and third now in April).

 

The issue arises typically in location specific variable(s) and is in the last digit of coefficient.

An example below

Here variable dLAT1km gets a value 2.68357. Using same data but on the next day coefficient is 2.68356. Everything else is the same.

This particular error has now been replicated on two computers running regressions today and yesterday!

My colleague is doing a test run later this evening and trying to see if the problem arises between two runs during the same day.

 

                   coefficient   std. error    t-ratio    p-value

  ----------------------------------------------------------------

  const             9.61714      0.0383124    251.0      0.0000    ***

  size             -0.0360853    0.00170093   -21.22     1.07e-095 ***

  size2             0.000328265  2.48074e-05   13.23     2.61e-039 ***

  size3            -1.04849e-06  1.14382e-07   -9.167    6.99e-020 ***

  age              -0.00931640   0.00120263    -7.747    1.14e-014 ***

  age2             -0.000252806  4.43133e-05   -5.705    1.23e-08  ***

  age3              4.79820e-06  4.62571e-07   10.37     5.91e-025 ***

  D_c_erinom        0.0248313    0.00777210     3.195    0.0014    ***

  D_c_tyyd         -0.0711856    0.00418330   -17.02     3.68e-063 ***

  D_c_huonot       -0.147710     0.0139738    -10.57     7.66e-026 ***

  apartment_sauna   0.0611526    0.00529751    11.54     1.94e-030 ***

  Q_7              -0.00308946   0.00694088    -0.4451   0.6563  

  Q_6              -0.000603071  0.00697798    -0.08642  0.9311  

  Q_5              -0.0102530    0.00680023    -1.508    0.1317  

  Q_4               0.00770398   0.00652315     1.181    0.2377  

  Q_3               0.00443441   0.00680937     0.6512   0.5149  

  Q_2              -0.00412167   0.00661880    -0.6227   0.5335  

  Q_1               0.00559893   0.00676218     0.8280   0.4077  

  dLAT1km           2.68357      1.30054        2.063    0.0391    **

  dLON1km          -6.45504      3.12653       -2.065    0.0390    **

  DP_84            -0.119379     0.0860273     -1.388    0.1653  

  lot_ownership    -0.0894216    0.00785090   -11.39     1.09e-029 ***

  d1200            -0.300451     0.00921419   -32.61     1.54e-211 ***

 

The problem

 

We run the regression monthly with two different computers [we verify our valuation by duplicating the process].

In both computers the Gretl version is the same 64-bit, and also the Windows 7 (pro) version is same in both computers. Now we’ve tested the runs on following days, and same problem appears even when using same computer. Each run consists of 480 regressions for different properties (this exercise is done for property valuation) and the problem occurs in c. 1-6 regressions, which are not the same on each time – i.e. the problem is rather small. On portfolio level it was this time 7,246e-7 % - but on property level error can be even 0,2%.

 

It is not the monetary value, but the real problem is that we can’t repeat the estimations.

 

In principle all these regressions ought to be solvable in closed form. Just wondering if one of the following could be the reason:

-        some algorithm is used in order to make calculations faster

-        somewhere in the Gretl code a random number generator is used

-        some rounding rule applies computer internal clock (odd/even date)

 

Kind Regards

 

Mikael Postila, MRICS

Head of Analysis

Orava Funds plc

 

t. +358 (0)50 347 2373

e. Mikael.Postila-z/r6kTLFilUoM3ANB2zruQ@public.gmane.org

a. Fabianinkatu 14B, FI-00100 Helsinki, FINLAND

 

<div><div class="WordSection1">
<p class="MsoNormal">Hi<p></p></p>
<p class="MsoNormal"><p>&nbsp;</p></p>
<p class="MsoNormal"><span lang="EN-US">We&rsquo;ve been experiencing a problem where exactly same data yields different regression results depending on A) which computer is used and B) when regression is being run on same computer.<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">This is the third time the problem arises (first time was in February, next was in March and third now in April).<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">The issue arises typically in location specific variable(s) and is in the last digit of coefficient.<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">An example below<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">Here variable dLAT1km gets a value 2.68357. Using same data but on the next day coefficient is 2.68356. Everything else is the same.<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">This particular error has now been replicated on two computers running regressions today and yesterday!<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">My colleague is doing a test run later this evening and trying to see if the problem arises between two runs during the same day.<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; coefficient&nbsp;&nbsp; std. error&nbsp;&nbsp;&nbsp; t-ratio&nbsp;&nbsp;&nbsp; p-value <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;----------------------------------------------------------------<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; const&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 9.61714&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0.0383124&nbsp;&nbsp;&nbsp; 251.0&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0.0000&nbsp;&nbsp;&nbsp; ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; size&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.0360853&nbsp;&nbsp;&nbsp; 0.00170093&nbsp;&nbsp; -21.22&nbsp;&nbsp;&nbsp;&nbsp; 1.07e-095 ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; size2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0.000328265&nbsp; 2.48074e-05&nbsp;&nbsp; 13.23&nbsp;&nbsp;&nbsp;&nbsp; 2.61e-039 ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; size3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -1.04849e-06&nbsp; 1.14382e-07&nbsp;&nbsp; -9.167&nbsp;&nbsp;&nbsp; 6.99e-020 ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; age&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.00931640&nbsp;&nbsp; 0.00120263&nbsp;&nbsp;&nbsp; -7.747&nbsp;&nbsp;&nbsp; 1.14e-014 ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; age2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.000252806&nbsp; 4.43133e-05&nbsp;&nbsp; -5.705&nbsp;&nbsp;&nbsp; 1.23e-08&nbsp; ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; age3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 4.79820e-06&nbsp; 4.62571e-07&nbsp;&nbsp; 10.37&nbsp;&nbsp;&nbsp;&nbsp; 5.91e-025 ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; </span>D_c_erinom&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0.0248313&nbsp;&nbsp;&nbsp; 0.00777210&nbsp;&nbsp;&nbsp;&nbsp; 3.195&nbsp;&nbsp;&nbsp; 0.0014&nbsp;&nbsp;&nbsp; ***<p></p></p>
<p class="MsoNormal">&nbsp; D_c_tyyd&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.0711856&nbsp;&nbsp;&nbsp; 0.00418330&nbsp;&nbsp; -17.02&nbsp;&nbsp;&nbsp;&nbsp; 3.68e-063 ***<p></p></p>
<p class="MsoNormal">&nbsp; <span lang="EN-US">D_c_huonot&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.147710&nbsp;&nbsp;&nbsp;&nbsp; 0.0139738&nbsp;&nbsp;&nbsp; -10.57&nbsp;&nbsp;&nbsp;&nbsp; 7.66e-026 ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; apartment_sauna&nbsp;&nbsp; 0.0611526&nbsp;&nbsp;&nbsp; 0.00529751&nbsp;&nbsp;&nbsp; 11.54&nbsp;&nbsp;&nbsp;&nbsp; 1.94e-030 ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; Q_7&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.00308946&nbsp;&nbsp; 0.00694088&nbsp;&nbsp;&nbsp; -0.4451&nbsp;&nbsp; 0.6563&nbsp;&nbsp; <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;Q_6&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.000603071&nbsp; 0.00697798&nbsp;&nbsp;&nbsp; -0.08642&nbsp; 0.9311&nbsp;&nbsp; <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;Q_5&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.0102530&nbsp;&nbsp;&nbsp; 0.00680023&nbsp;&nbsp;&nbsp; -1.508&nbsp;&nbsp;&nbsp; 0.1317&nbsp;&nbsp; <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;Q_4&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &nbsp;0.00770398&nbsp;&nbsp; 0.00652315&nbsp;&nbsp;&nbsp;&nbsp; 1.181&nbsp;&nbsp;&nbsp; 0.2377&nbsp;&nbsp; <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;Q_3&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0.00443441&nbsp;&nbsp; 0.00680937&nbsp;&nbsp;&nbsp;&nbsp; 0.6512&nbsp;&nbsp; 0.5149&nbsp;&nbsp; <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;Q_2&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.00412167&nbsp;&nbsp; 0.00661880&nbsp;&nbsp;&nbsp; -0.6227&nbsp;&nbsp; 0.5335&nbsp;&nbsp; <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;Q_1&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 0.00559893&nbsp;&nbsp; 0.00676218&nbsp;&nbsp;&nbsp;&nbsp; 0.8280&nbsp;&nbsp; 0.4077&nbsp;&nbsp; <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;dLAT1km&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 2.68357&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 1.30054&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 2.063&nbsp;&nbsp;&nbsp; 0.0391&nbsp;&nbsp;&nbsp; **≤p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; dLON1km&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -6.45504&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 3.12653&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -2.065&nbsp;&nbsp;&nbsp; 0.0390&nbsp;&nbsp;&nbsp; **≤p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; DP_84&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.119379&nbsp;&nbsp;&nbsp;&nbsp; 0.0860273&nbsp;&nbsp;&nbsp;&nbsp; -1.388&nbsp;&nbsp;&nbsp; 0.1653&nbsp;&nbsp; <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp;&nbsp;lot_ownership&nbsp;&nbsp;&nbsp; -0.0894216&nbsp;&nbsp;&nbsp; 0.00785090&nbsp;&nbsp; -11.39&nbsp;&nbsp;&nbsp; &nbsp;1.09e-029 ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">&nbsp; d1200&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; -0.300451&nbsp;&nbsp;&nbsp;&nbsp; 0.00921419&nbsp;&nbsp; -32.61&nbsp;&nbsp;&nbsp;&nbsp; 1.54e-211 ***<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">The problem <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">We run the regression monthly with two different computers [we verify our valuation by duplicating the process]. <p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">In both computers the Gretl version is the same 64-bit, and also the Windows 7 (pro) version is same in both computers. Now we&rsquo;ve tested the runs on following days, and same problem appears even when using same computer. Each run consists of 480 regressions for different properties (this exercise is done for property valuation) and the problem occurs in c. 1-6 regressions, which are not the same on each time &ndash; i.e. the problem is rather small. On portfolio level it was this time 7,246e-7 % - but on property level error can be even 0,2%.<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">It is not the monetary value, but the real problem is that we can&rsquo;t repeat the estimations.<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">In principle all these regressions ought to be solvable in closed form. Just wondering if one of the following could be the reason:<p></p></span></p>
<p class="MsoListParagraph"><span lang="EN-US"><span>-<span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </span></span></span><span lang="EN-US">some algorithm is used in order to make calculations faster<p></p></span></p>
<p class="MsoListParagraph"><span lang="EN-US"><span>-<span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </span></span></span><span lang="EN-US">somewhere in the Gretl code a random number generator is used<p></p></span></p>
<p class="MsoListParagraph"><span lang="EN-US"><span>-<span>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; </span></span></span><span lang="EN-US">some rounding rule applies computer internal clock (odd/even date)<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">Kind Regards<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">Mikael Postila, MRICS<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">Head of Analysis<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">Orava Funds plc<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US"><p>&nbsp;</p></span></p>
<p class="MsoNormal"><span lang="EN-US">t. +358 (0)50&nbsp;347 2373<p></p></span></p>
<p class="MsoNormal"><span lang="EN-US">e. <a href="mailto:Mikael.Postila@..."><span lang="FI">Mikael.Postila@...</span></a></span><p></p></p>
<p class="MsoNormal">a. Fabianinkatu 14B, FI-00100 Helsinki, FINLAND<p></p></p>
<p class="MsoNormal"><p>&nbsp;</p></p>
</div></div>
Andreas Zervas | 25 Apr 22:47 2016
Picon

Features request

(Sven:)
Right, I should have thought of that. Still I would be interested to see a direct comparison between the AB approach there and a competing C model doing the same thing. (It's a bit off-topic here, but if you have literature pointers they would be welcome.)


I think Caldara - Kamps (ECB WP, 857 I think) is what you would like to see for the US.


Andreas
<div><div>(Sven:)<br>Right, I should have thought of that. Still I would be interested to see 
a direct comparison between the AB approach there and a competing C 
model doing the same thing. (It's a bit off-topic here, but if you have 
literature pointers they would be welcome.)<br><br><br>I think Caldara - Kamps (ECB WP, 857 I think) is what you would like to see for the US.<br><br><br>Andreas<br>
</div></div>
Andreas Zervas | 25 Apr 20:41 2016
Picon

Feature request

Am 25.04.2016 um 12:44 schrieb Andreas Zervas: > coding, while adding identities in systems seems to me that it only > extends a capability already present in the FIML estimator. But off > course the programmers know better. But why do you want to have identities for those estimators where they are irrelevant, given a situation where the simulation apparatus is not yet here?

Answer : For 2 reasons. Because for example one can simulate a basic macroeconomic model and do the transformations needed to impose the accounting identities, e.g. the GDP identity. A second reason is that I would like to be able to add debt in a fiscal VAR as exogenous (see Favero - Giavazzi AEJEP 2012), and simulate forecasts where debt follows its path dictated by the identity governing its evolution; I can do it on my own for a couple of models, but a general solution is preferable.

 (Sven:) > Personally I'm not a big fan of AB models, because I never saw a > convincing case where you absolutely had to distribute your > contemporaneous restrictions over two matrices. > (Andreas:) > Answer: I work in fiscal policy issues - there the workhorse is the > Blanchard - Perotti (2002) identification restrictions, that > correspond to an AB model. Right, I should have thought of that. Still I would be interested to see a direct comparison between the AB approach there and a competing C model doing the same thing. (It's a bit off-topic here, but if you have literature pointers they would be welcome.)

Answer: As the most common C models are Cholesky restrictions, placing government spending first as is customary does not make a big difference with AB models for fiscal policy. But it matters a lot for taxes.

 cheers, sven

All best,

Andreas
<div><div>Am 25.04.2016 um 12:44 schrieb Andreas Zervas:

&gt; coding, while adding identities in systems seems to me that it only
&gt; extends a capability already present in the FIML estimator. But off
&gt; course the programmers know better.

But why do you want to have identities for those estimators where they 
are irrelevant, given a situation where the simulation apparatus is not 
yet here?<br><br>Answer : For 2 reasons. Because for example one can simulate a basic macroeconomic model and do the transformations needed to impose the accounting identities, e.g. the GDP identity. A second reason is that I would like to be able to add debt in a fiscal VAR as exogenous (see Favero - Giavazzi AEJEP 2012), and simulate forecasts where debt follows its path dictated by the identity governing its evolution; I can do it on my own for a couple of models, but a general solution is preferable.<br><br>&nbsp;(Sven:)
&gt;  Personally I'm not a big fan of AB models, because I never saw a
&gt; convincing case where you absolutely had to distribute your
&gt; contemporaneous restrictions over two matrices.
&gt;
(Andreas:)
&gt; Answer: I work in fiscal policy issues - there the workhorse is the
&gt; Blanchard - Perotti (2002) identification restrictions, that
&gt; correspond to an AB model.

Right, I should have thought of that. Still I would be interested to see 
a direct comparison between the AB approach there and a competing C 
model doing the same thing. (It's a bit off-topic here, but if you have 
literature pointers they would be welcome.)<br><br>Answer: As the most common C models are Cholesky restrictions, placing government spending first as is customary does not make a big difference with AB models for fiscal policy. But it matters a lot for taxes.<br><br>&nbsp;cheers,
sven<br><br>All best, <br><br>Andreas<br>
</div></div>
Andreas Zervas | 25 Apr 12:44 2016
Picon

Features request

Am 24.04.2016 um 16:12 schrieb Andreas Zervas: > > I was wondering whether it would be possible to add some features in > Gretl that some of us will find extremely useful (provided I guess it is > not such a big burden to code them). That's a lot of different features, maybe it would help if you named some priorities. Apart from that, feel free to create items/tickets on the Sourceforge feature request tracker for gretl so that these things are not forgotten. > > The first is related to systems: I was wondering whether it would be > possible to add the ability to include identities in systems estimated > with estimators other than FILM; it would be very useful to have such an > ability for forecasting / simulation purposes, even if estimation is not > fully efficient; in any case OLS / IV estimation of systems should give > more robust results. This is actually directly connected to your next point, because as you say identities for OLS/IV etc. are only needed for forecasting or simulation, not for estimation.

Answer: Truly, they are connected. But I think mimicking model structure in EViews should be much more demanding with respect to coding, while adding identities in systems seems to me that it only extends a capability already present in the FIML estimator. But off course the programmers know better.

> > A related extension would be to develop a framework to imitate the > EViews model functionality: add single (estimated or not) equations or > systems of equations to a bigger model and simulate this structure. I > understand that this is much more complex from a computational and > programming point of view than the previous suggestion. Yes I suggested something like this to Allin and Jack myself some time ago. There has been some underlying work in gretl's foundations to make this feasible at some point at the level of Hansl programming/scripting. (For example making more use of the 'bundle' collection datatype.) However, there is only so much that Allin and Jack can do. Note that if forecasting or simulation is done in an extra function package then the specification of identities would not be necessary in gretl's 'system' estimation command. Instead you would provide the information about the identitites in the system to the extended forecasting package, and that package (not yet existing...) might internally call gretl's system estimation. Of course, it's all a question about the user interface of that package / add-on. > > Perhaps you could start with systems having recursive structure so as to > initially avoid the problem of finding the solution satisfying all > equations in all periods. Maybe. > > A third very useful feature in my opinion relates to graphs: it is very > useful to have subplots in the same graph, as in matlab / octave. > graphpg is a substitute with limitations, not least in the number the > way you put the subplots. You mean whether it's 2-by-4 or 4-by-2 for example?


Answer: or 3-by-3 etc... The 2 by 4 limit seems to me like a logical, yet arbitrary limit that assumes a A4 page with portrait orientation. Gretl itself allows more subplots (4-by-4 I think) in the case of its native SVAR IRF plotting implementation using the GUI.


 > Secondary graph request: is it possible to allow for other linestyles > like the usual dashes, dots etc in Gretl without editing gnuplot files? This feature is approaching, there has been some discussion already (on the devel list I think). It depends on the underlying gnuplot version.


 Answer: Very nice - eager to see it.

> > Finally (to Riccardo only) : please, when it is possible add some > remaining functionalities to SVAR package. First, there is no reason > (and should not be very difficult) to have AB models for VECMs (as > eventually these are also VARs). Happy to hear that there's demand for the SVAR package. Your request is fair enough, but at least this variant is not very common I think. Personally I'm not a big fan of AB models, because I never saw a convincing case where you absolutely had to distribute your contemporaneous restrictions over two matrices.


Answer: I work in fiscal policy issues - there the workhorse is the Blanchard - Perotti (2002) identification restrictions, that correspond to an AB model. You can do it with IV estimation, but it is useful to have the structural factorization. Nevertheless, in general you are right.


 Secondly, please try to add long run > restrictions on permanent shocks. Essentially, would it be possible to > allow more general identification restrictions like in Jmulti or Warne's > SVAR? Assuming we're talking about the permanent shocks in a cointegrated SVEC model, yes there were some issues to be resolved about long-run restrictions if I remember correctly. I think Jack had a preliminary paper about something related. This is work-in-progress in principle, but I will openly admit that progress isn't very quick there.



Answer: Ok, yet it will be great help when it is available.



 > A related thing is that it would be nice to have a function to simply > estimate the structural factorization. I think that the necessary > ingredients to do that is the variance covariance matrix form the VAR / > VECM and a pattern matrix holding restrictions and free elements; could > you have a public function to do just that, or how is it possible to do > that in the current state of the SVAR package? I tried to do it myself > that but the code is so complicated that I gave it up. Here I don't quite understand what you mean. Once you estimate the SVAR, you always get the estimated A or B or C. What else do you mean?


Answer: Think of a Bayesian VAR with a natural conjugate prior - IRFs etc come from simulation. Starting from a draw of the variance covariance matrix of residuals, the only feasible SVAR is one using a Cholesky decomposition. But having a public function to estimate the contemporaneous matrices without needing the full VAR estimation would make it possible to do this kind of analysis.



 > > Thank you for all the good stuff you have given us so far. I hope these > are not too much to ask. > It all depends on your time-frame and patience ;-) cheers, sven


Hope I made myself clearer.
Thank you all,

Andreas
<div><div>Am 24.04.2016 um 16:12 schrieb Andreas Zervas:
&gt;
&gt; I was wondering whether it would be possible to add some features in
&gt; Gretl that some of us will find extremely useful (provided I guess it is
&gt; not such a big burden to code them).

That's a lot of different features, maybe it would help if you named 
some priorities. Apart from that, feel free to create items/tickets on 
the Sourceforge feature request tracker for gretl so that these things 
are not forgotten.

&gt;
&gt; The first is related to systems: I was wondering whether it would be
&gt; possible to add the ability to include identities in systems estimated
&gt; with estimators other than FILM; it would be very useful to have such an
&gt; ability for forecasting / simulation purposes, even if estimation is not
&gt; fully efficient; in any case OLS / IV estimation of systems should give
&gt; more robust results.

This is actually directly connected to your next point, because as you 
say identities for OLS/IV etc. are only needed for forecasting or 
simulation, not for estimation.
<br><br>Answer: Truly, they are connected. But I think mimicking model structure in EViews should be much more demanding with respect to coding, while adding identities in systems seems to me that it only extends a capability already present in the FIML estimator. But off course the programmers know better.<br><br>
&gt;
&gt; A related extension would be to develop a framework to imitate the
&gt; EViews model functionality: add single (estimated or not) equations or
&gt; systems of equations to a bigger model and simulate this structure. I
&gt; understand that this is much more complex from a computational and
&gt; programming point of view than the previous suggestion.

Yes I suggested something like this to Allin and Jack myself some time 
ago. There has been some underlying work in gretl's foundations to make 
this feasible at some point at the level of Hansl programming/scripting. 
(For example making more use of the 'bundle' collection datatype.) 
However, there is only so much that Allin and Jack can do.

Note that if forecasting or simulation is done in an extra function 
package then the specification of identities would not be necessary in 
gretl's 'system' estimation command. Instead you would provide the 
information about the identitites in the system to the extended 
forecasting package, and that package (not yet existing...) might 
internally call gretl's system estimation. Of course, it's all a 
question about the user interface of that package / add-on.

&gt;
&gt; Perhaps you could start with systems having recursive structure so as to
&gt; initially avoid the problem of finding the solution satisfying all
&gt; equations in all periods.

Maybe.

&gt;
&gt; A third very useful feature in my opinion relates to graphs: it is very
&gt; useful to have subplots in the same graph, as in matlab / octave.
&gt; graphpg is a substitute with limitations, not least in the number the
&gt; way you put the subplots.

You mean whether it's 2-by-4 or 4-by-2 for example?<br><br><br>Answer: or 3-by-3 etc... The 2 by 4 limit seems to me like a logical, yet arbitrary limit that assumes a A4 page with portrait orientation. Gretl itself allows more subplots (4-by-4 I think) in the case of its native SVAR IRF plotting implementation using the GUI.<br><br><br>&nbsp;&gt; Secondary graph request: is it possible to allow for other linestyles
&gt; like the usual dashes, dots etc in Gretl without editing gnuplot files?

This feature is approaching, there has been some discussion already (on 
the devel list I think). It depends on the underlying gnuplot version.<br><br><br>&nbsp;Answer: Very nice - eager to see it.
<br><br>&gt;
&gt; Finally (to Riccardo only) : please, when it is possible add some
&gt; remaining functionalities to SVAR package. First, there is no reason
&gt; (and should not be very difficult) to have AB models for VECMs (as
&gt; eventually these are also VARs).

Happy to hear that there's demand for the SVAR package. Your request is 
fair enough, but at least this variant is not very common I think. 
Personally I'm not a big fan of AB models, because I never saw a 
convincing case where you absolutely had to distribute your 
contemporaneous restrictions over two matrices.<br><br><br>Answer: I work in fiscal policy issues - there the workhorse is the Blanchard - Perotti (2002) identification restrictions, that correspond to an AB model. You can do it with IV estimation, but it is useful to have the structural factorization. Nevertheless, in general you are right.<br><br><br>&nbsp;Secondly, please try to add long run
&gt; restrictions on permanent shocks. Essentially, would it be possible to
&gt; allow more general identification restrictions like in Jmulti or Warne's
&gt; SVAR?

Assuming we're talking about the permanent shocks in a cointegrated SVEC 
model, yes there were some issues to be resolved about long-run 
restrictions if I remember correctly. I think Jack had a preliminary 
paper about something related. This is work-in-progress in principle, 
but I will openly admit that progress isn't very quick there.<br><br><br><br>Answer: Ok, yet it will be great help when it is available.<br><br><br><br>&nbsp;&gt; A related thing is that it would be nice to have a function to simply
&gt; estimate the structural factorization. I think that the necessary
&gt; ingredients to do that is the variance covariance matrix form the VAR /
&gt; VECM and a pattern matrix holding restrictions and free elements; could
&gt; you have a public function to do just that, or how is it possible to do
&gt; that in the current state of the SVAR package? I tried to do it myself
&gt; that but the code is so complicated that I gave it up.

Here I don't quite understand what you mean. Once you estimate the SVAR, 
you always get the estimated A or B or C. What else do you mean?<br><br><br>Answer: Think of a Bayesian VAR with a natural conjugate prior - IRFs etc come from simulation. Starting from a draw of the variance covariance matrix of residuals, the only feasible SVAR is one using a Cholesky decomposition. But having a public function to estimate the contemporaneous matrices without needing the full VAR estimation would make it possible to do this kind of analysis.<br><br><br><br>&nbsp;&gt;
&gt; Thank you for all the good stuff you have given us so far. I hope these
&gt; are not too much to ask.
&gt;

It all depends on your time-frame and patience ;-)

cheers,
sven<br><br><br>Hope I made myself clearer.<br>Thank you all,<br><br>Andreas<br>
</div></div>

Gmane