From ???@??? Wed Mar 08 09:29:36 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id CAA12796
	for <robt@localhost>; Wed, 8 Mar 2000 02:30:11 +1100
Message-Id: <200003071530.CAA12796@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Wed, 08 Mar 2000 02:30:11 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 8 Mar 100 02:19:30 GMT+1000
From: Vivek Ghosal <Vivek.Ghosal@usdoj.gov>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Panel data
Date: Tue, 7 Mar 2000 10:15:00 -0500
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
X-Mailer: Worldtalk (NetJunction 4.6-p7)/MIME (via Mercury MTS (Bindery) v1.40)

Hello-

I have a panel data set with (cross-section) N=332 and (years) T=37.
The cross-section units have an ID with non-consecutive values from
1011 to 4500. I want to estimate a regression for each cross-section unit
(as opposed to panel estimation).

How do I write the *DO* or *DOFOR* command given that the cross
-section ID values are non-consecutive? (Being a new RATS user, I*m
unsure.)

Thanks a lot.

Vivek

From ???@??? Wed Mar 08 09:30:25 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id EAA08650
	for <robt@localhost>; Wed, 8 Mar 2000 04:00:04 +1100
Message-Id: <200003071700.EAA08650@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Wed, 08 Mar 2000 04:00:04 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 8 Mar 100 03:46:25 GMT+1000
From: "Estima" <tomm@estima.com>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: Panel data
Date: Tue, 7 Mar 2000 10:34:05 -0600
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Organization: Estima
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
X-mailer: Pegasus Mail for Win32 (v3.11) (via Mercury MTS (Bindery) v1.40)

> I have a panel data set with (cross-section) N=332 and (years) T=37.
> The cross-section units have an ID with non-consecutive values from
> 1011 to 4500. I want to estimate a regression for each cross-section unit
> (as opposed to panel estimation).
> 
> How do I write the *DO* or *DOFOR* command given that the cross
> -section ID values are non-consecutive? (Being a new RATS user, I*m
> unsure.)

First, when you say "regression for each cross-section", I assume you 
mean that you want to do a regression using the 37 data points for 
one individual (e.g., ID # 1011), then a separate regression using 
the 37 data points for another individual (e.g., ID # 4500), and so 
on?

If so, and if you don't care about the order in which the regressions 
are done (that is, you want to do a regression for each ID #, but 
don't care about the order in which the regressions are done), just 
use DO to walk through the whole data set:

compute numind = 332
compute numper = 37
do start=1,numper*numind,37
 linreg depvar start start+36
 # regressors
 ...
end do

If you do want to do the regressions in ID # order, there are 
probably a lot of ways to go about that. The ideal situation would be 
if you have a series containing the ID #'s, such that each 
observation for a given individual contains the ID # for that 
individual. If so, just sort on the ID #'s and do your estimation.

If you don't have lags in the model (and thus it doesn't matter where 
the time periods fall within each inidividual), just sort on the ID 
#'s, being sure to sort your dependent variable and all regressors 
allong with the data. You should also define an "index" series so you 
can "unsort" the data later:

set index = t
order idnums / y x1 x2 x3 index
do start=1,numper*numind,37
 linreg y start start+36
 # x1 x2 x3
 ...
end do
order index / idnums y x1 x2 x3

If you do have lags, and thus need to preserve the ordering of the 
time periods, you need to create a sort key that includes botht he ID 
info and the period index. You can do that by multiplying the ID 
series by a scale factor and adding a period index:

set period = %period(t)
set sortkey = idnums*100 + period
print / sortkey idnums period


 ENTRY       SORTKEY          IDNUMS          PERIOD
  1//01  101101.00000000 1011.0000000000  1.000000000000
  1//02  101102.00000000 1011.0000000000  2.000000000000
  1//03  101103.00000000 1011.0000000000  3.000000000000
  1//04  101104.00000000 1011.0000000000  4.000000000000
  1//05  101105.00000000 1011.0000000000  5.000000000000
  1//06  101106.00000000 1011.0000000000  6.000000000000
  1//07  101107.00000000 1011.0000000000  7.000000000000

etc.

Then do:

order sortkey / idnums period index y x1 x2 x3
do start=1,numper*numind,37
 linreg y start start+36
 # x1 x2 x3
 ...
end do
order index / sortkey idnums period y x1 x2 x3

Sincerely,
Tom Maycock

--
------------------------------------------------------------
|   Estima                    |  Sales:   (800) 822-8038   |
|   P.O. Box 1818             |           (847) 864-8772   |
|   Evanston, IL 60204-1818   |  Support: (847) 864-1910   |
|   USA                       |  Fax:     (847) 864-6221   |
|   http://www.estima.com     |  estima@estima.com         |
------------------------------------------------------------

From ???@??? Wed Mar 08 09:30:43 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id GAA17102
	for <robt@localhost>; Wed, 8 Mar 2000 06:50:05 +1100
Message-Id: <200003071950.GAA17102@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Wed, 08 Mar 2000 06:50:05 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 8 Mar 100 06:46:01 GMT+1000
From: Alejandro Arrieta Herrera <aarrieta@sbs.gob.pe>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: Panel data
Date: Tue, 07 Mar 2000 14:46:25 -0500
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
X-Mailer: Mozilla 4.5 [es] (WinNT; I) (via Mercury MTS (Bindery) v1.40)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

I have a Panel Data and I am trying to test cointegration. Do someone know how
to test it?

Alejandro Arrieta H.
Superintendencia de Banca y Seguros




From ???@??? Wed Mar 08 09:30:47 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id HAA14664
	for <robt@localhost>; Wed, 8 Mar 2000 07:15:04 +1100
Message-Id: <200003072015.HAA14664@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Wed, 08 Mar 2000 07:15:04 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 8 Mar 100 07:12:35 GMT+1000
From: "Christopher F. Baum" <baum@bc.edu>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: Panel data
Date: Tue, 07 Mar 2000 15:12:35 -0500
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
X-Mailer: Mulberry (MacOS) [1.4.5, s/n P020-300786-009] (via Mercury MTS (Bindery) v1.40)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Peter Pedroni @ Indiana U. <ppedroni@indiana.edu> has written RATS
procedures to accomplish these tests.

Kit Baum
Boston College Economics

--On Tue, Mar 7, 2000 14:46 -0500 Alejandro Arrieta Herrera
<aarrieta@sbs.gob.pe> wrote:

> I have a Panel Data and I am trying to test cointegration. Do someone
> know how to test it?
> 
> Alejandro Arrieta H.
> Superintendencia de Banca y Seguros
> 
> 
> 



-----------------------------------------------------------------
Kit Baum    baum@bc.edu    http://fmwww.bc.edu/ec-v/baum.fac.html

From ???@??? Fri Mar 10 09:28:15 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id JAA10242
	for <robt@localhost>; Thu, 10 Mar 1994 09:10:11 +1100
Message-Id: <199403092210.JAA10242@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Thu, 10 Mar 1994 09:10:11 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 10 Mar 100 06:22:23 GMT+1000
From: "Bob Shain" <bshain@mtppi.org>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: RE: Equivalent of SAS 'by' statement in SAS procs
Date: Thu, 9 Mar 2000 14:22:33 -0500
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
Content-Type: text/plain;
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0 (via Mercury MTS (Bindery) v1.40)

I'm doing a panel data problem with rats and I need to do a regression
separately for each member fo the panel.

If I were doing this in SAS, I would do a 'proc reg' by panel and the output
would come out in a file indexed by panel.

How can I do this in RATS.  Do I have to use 'do' instructions or is there a
way more or less equivalent to the SAS language way?


From ???@??? Fri Mar 10 09:28:45 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id JAA10336
	for <robt@localhost>; Thu, 10 Mar 1994 09:10:20 +1100
Message-Id: <199403092210.JAA10336@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Thu, 10 Mar 1994 09:10:21 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 10 Mar 100 08:25:42 GMT+1000
From: "Estima" <tomm@estima.com>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: RE: Equivalent of SAS 'by' statement in SAS procs
Date: Thu, 9 Mar 2000 15:18:56 -0600
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Organization: Estima
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
X-mailer: Pegasus Mail for Win32 (v3.11) (via Mercury MTS (Bindery) v1.40)

> I'm doing a panel data problem with rats and I need to do a regression
> separately for each member fo the panel.
> 
> If I were doing this in SAS, I would do a 'proc reg' by panel and the output
> would come out in a file indexed by panel.
> 
> How can I do this in RATS.  Do I have to use 'do' instructions or is there a
> way more or less equivalent to the SAS language way?
> 

Somebody asked almost exactly the same question a couple of days ago 
(March 7), and I provided a fairly lengthy answer. Check your mail 
from that date, or check the on-line archives. If you can't find it, 
let me know and I'll dig out my answer again.

Tom Maycock
Estima

--
------------------------------------------------------------
|   Estima                    |  Sales:   (800) 822-8038   |
|   P.O. Box 1818             |           (847) 864-8772   |
|   Evanston, IL 60204-1818   |  Support: (847) 864-1910   |
|   USA                       |  Fax:     (847) 864-6221   |
|   http://www.estima.com     |  estima@estima.com         |
------------------------------------------------------------

From ???@??? Fri Mar 10 09:59:21 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id JAA10738
	for <robt@localhost>; Thu, 10 Mar 1994 09:50:04 +1100
Message-Id: <199403092250.JAA10738@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Thu, 10 Mar 1994 09:50:04 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 10 Mar 100 09:48:14 GMT+1000
From: Justin Maharaj <recsearch1@excite.com>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Multiplicative Seasonality
Date: Thu, 9 Mar 2000 14:44:42 -0800 (PST)
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Excite Inbox (via Mercury MTS (Bindery) v1.40)

Is it possible, using RATS, to model multiplicative seasonal coefficients of
the form:

(1-aL)y = (1+B1L)(1+B4L^4)et

where:
B1 and B4 reads as B subscript 1 and 4 resp. and et as e subscript t

(See Enders, W., 1995, Applied Econometric Time Series, pg. 111-113.)

Justin Maharaj





_______________________________________________________
Get 100% FREE Internet Access powered by Excite
Visit http://freeworld.excite.com


From ???@??? Sat Mar 11 09:46:31 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id UAA11492
	for <robt@localhost>; Thu, 10 Mar 1994 20:55:05 +1100
Message-Id: <199403100955.UAA11492@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Thu, 10 Mar 1994 20:55:05 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 10 Mar 100 20:50:39 GMT+1000
From: George Semertzakis <georgios.semertzakis@durham.ac.uk>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: diagnostic tests
Date: Fri, 10 Mar 2000 09:49:04 +0000
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Organization: Deptartment of Econ. & Finance
X-Mailer: Mozilla 4.7 [en] (Win98; I) (via Mercury MTS (Bindery) v1.40)
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-7
Content-Transfer-Encoding: 7bit

<x-charset iso-8859-7>Does RATS has any command to give you  the diagnostic tests? I would
like to have the Heteroscedasticity and the Serial Correlation, as well
as the normality etc..  for my time series. In mean something like the
Microfit that gives you these information at the bottom of the print
out.

George Semertzakis

--
-----------------------------------------------------
George Semertzakis          | Tel: +44(0)191-3741821
Dept.of Economics & Finance |
University of Durham        | Fax: +44(0)191-3747289
23/26 Old Elvet             |
Durham                      |
P.C: DH1 3HY                |
UK                          |http://www.thehungersite.com/
------------------------------------------------------



</x-charset>
From ???@??? Sat Mar 11 09:47:01 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id XAA08176
	for <robt@localhost>; Thu, 10 Mar 1994 23:30:04 +1100
Message-Id: <199403101230.XAA08176@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Thu, 10 Mar 1994 23:30:04 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 10 Mar 100 23:16:05 GMT+1000
From: Andreas Faust <andreasfaust@yahoo.com>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: Multiplicative Seasonality
Date: Fri, 10 Mar 2000 04:16:16 -0800 (PST)
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit            
X-Mailer: Mercury MTS (Bindery) v1.40

<x-charset iso-8859-1>
yes, look for example the HEGY.SRC which is a test for
seasonal unit roots (available on www.estima.com). You
can calculate this form of differences with the SET
command for series.

Andreas


--- Justin Maharaj <recsearch1@excite.com> schrieb:
> Is it possible, using RATS, to model multiplicative
> seasonal coefficients of
> the form:
> 
> (1-aL)y = (1+B1L)(1+B4L^4)et
> 
> where:
> B1 and B4 reads as B subscript 1 and 4 resp. and et
> as e subscript t
> 
> (See Enders, W., 1995, Applied Econometric Time
> Series, pg. 111-113.)
> 
> Justin Maharaj
> 
> 
> 
> 
> 
>
_______________________________________________________
> Get 100% FREE Internet Access powered by Excite
> Visit http://freeworld.excite.com
> 
> 

=====
Andreas Faust
Av. Carúpano N° 23
Qta. El Rosedal
Las Palmas, Caracas
Venezuela
Tel.: +58  2 7818621
andreasfaust@yahoo.com
__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com

</x-charset>
From ???@??? Sat Mar 11 09:47:42 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id DAA12284
	for <robt@localhost>; Fri, 11 Mar 1994 03:00:13 +1100
Message-Id: <199403101600.DAA12284@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Fri, 11 Mar 1994 03:00:13 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 11 Mar 100 02:48:23 GMT+1000
From: "Estima" <tomm@estima.com>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: Multiplicative Seasonality
Date: Fri, 10 Mar 2000 09:51:10 -0600
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Organization: Estima
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
X-mailer: Pegasus Mail for Win32 (v3.11) (via Mercury MTS (Bindery) v1.40)

> Is it possible, using RATS, to model multiplicative seasonal coefficients of
> the form:
> 
> (1-aL)y = (1+B1L)(1+B4L^4)et
> 
> where:
> B1 and B4 reads as B subscript 1 and 4 resp. and et as e subscript t

That is in fact exactly how the SAR and SMA parameters work on the 
BOXJENK isntruction--see the formulation shown on page 14-12 in the 
manual. 

Sincerely,
Tom Maycock
Estima

--
------------------------------------------------------------
|   Estima                    |  Sales:   (800) 822-8038   |
|   P.O. Box 1818             |           (847) 864-8772   |
|   Evanston, IL 60204-1818   |  Support: (847) 864-1910   |
|   USA                       |  Fax:     (847) 864-6221   |
|   http://www.estima.com     |  estima@estima.com         |
------------------------------------------------------------

From ???@??? Sat Mar 11 09:47:56 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id DAA12866
	for <robt@localhost>; Sat, 11 Mar 2000 03:30:07 +1100
Message-Id: <200003101630.DAA12866@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Sat, 11 Mar 2000 03:30:07 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 11 Mar 100 03:07:46 GMT+1000
From: "Vasquez Escobar Diego M." <dvasques@banrep.gov.co>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: RE: diagnostic tests
Date: Fri, 10 Mar 2000 11:04:06 -0500
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21) (via Mercury MTS (Bindery) v1.40)
Content-Type: text/plain;

You can use RATS to do a set of tests for Heteroscedasticity,
autocorrelation and normality, take a glance at the RATS User's manual.
However these are some examples of BREUSCH-PAGAN, WHITE, ARCH and
BREUSCH-GOODFREY tests for Heteroscedasticity and autocorrelation:
For Normality take a glance at www.estima.com There is a set of procedures
that implement normality test such as KOLMTEST.SRC AND NORMTEST. SRC.

***************************************************
* BREUSCH-PAGAN test for Heteroscedasticity       *
***************************************************
SET LDEP = LOG(DEPVAR)
LINREG DEP / RESIDS
# CONSTANT INDEP
COMPUTE CHISTAT=%NOBS*%RSQUARED
CDF CHISQR CHISTAT 1
SET RESSQR = RESIDS**2
LINREG(NOPRINT) RESSQR
# CONSTANT INDEP
COMPUTE CHISTAT=%NOBS*%RSQUARED
CDF CHISQR CHISTAT 1
********************************************
* WHITE TEST FOR Heteroscedasticity        *
********************************************
LINREG DEPVAR / RESIDS
# CONSTANT IND1 IND2
SET IND3 = IND1*IND1
SET IND4 = IND2*IND2
SET MIX = IND1*IND2
SET RESSQR = RESIDS**2
LINREG(NOPRINT) RESSQR
# CONSTANT IND1 IND2 IND3 IND4 MIX
COMPUTE CHISTAT=%NOBS*%RSQUARED
CDF CHISQR CHISTAT 5
********************************************
* Test ARCH para HETEROSCEDASTICIDAD       *
********************************************
LINREG DEPVAR / RESIDS
# CONSTANT IND1 IND2
SET RESSQR = RESIDS**2
LINREG(NOPRINT) RESSQR
# CONSTANT RESSQR{1}
COMPUTE CHISTAT=%NOBS*%RSQUARED
CDF CHISQR CHISTAT 1


TO TEST FOR AUTOCORRELATION you can implement BREUSCH y GOODFREY and
Ljung-Box tests. This is an example.


***************************************************
* BREUSCH-GOODFREY test for AUTOCORRELATION       *
***************************************************
LINREG DEPVAR / RESIDS
# CONSTANT IND1 IND2
LINREG(NOPRINT) RESIDS
# CONSTANT PIB INF RESIDS{1 TO 4}
CDF CHISQR %TRSQ 4
Exclude
# resids{1 to 4}

Good Luck





From ???@??? Sat Mar 11 09:48:19 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id DAA12950
	for <robt@localhost>; Sat, 11 Mar 2000 03:30:15 +1100
Message-Id: <200003101630.DAA12950@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Sat, 11 Mar 2000 03:30:15 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 11 Mar 100 03:23:52 GMT+1000
From: Simon van Norden <Simon.van-Norden@hec.ca>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: diagnostic tests
Date: Fri, 10 Mar 2000 11:27:40 -0500
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Organization: Ecole des Hautes Etudes Commerciales
X-Mailer: Mozilla 4.61 [en] (Win98; U) (via Mercury MTS (Bindery) v1.40)
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit            

<x-charset iso-8859-1>There is no good built-in function in RATS for this stuff.
A friend and I programmed some residual-based diagnostic tests many years ago.
You can still get the code (with some others) by following the links from my
homepage. I've attached the relevant proc listing below.

Hope this helps.
Regards,

SvN

-- 
Simon van Norden, Prof. agrégé, www.hec.ca/pages/simon.van-norden
Service de l'enseignement de la finance, École des H.E.C.
3000 Cote-Sainte-Catherine, Montreal QC, CANADA   H3T 2A7
simon.van-norden@hec.ca or (514)340-6781 or fax:(514)340-5632
--
env noecho
*
* RESDIAG.SRC       by Robert Amano and Simon van Norden    02-07-91
* Copyright 1993, 1991 by Bank of Canada
* (modified for RATS v4 August 1993 by Jeff Gable)
*
* This procedure calculates a number of residual diagnostic tests
*
* The syntax for the procedure is
* @RESDIAG(options) residuals
* # list of series
* # dependent_variable fitted_values   
*  (the last card is used with the RESET option only)
*
* where
*   residuals       = the name of the residual series
*   list of series  = the variables used in your regression
*
* We assume that the list of regressors supplied by the user includes a CONSTANT
* as the first series in the list.
*
*  NOTE: that on the supplementary card you cannot use the usual RATS notation
*        for leads and lags (eg. X{0 to 2}).  If you use leads or lags in your 
*        model, you have two choices:
*        
*        1. Turn the LM and Breusch-Pagan test off (ie NOLMP, NOBPHET) and 
*           leave the reset test off.  The procedure will still calculate all 
*           other test statistics.
*        OR
*        2. Code each one into a new unlagged variable, then list the new
*           variable on the supplementary card.  Eg. set X1 = X(t-1)
*                                                    etc.
*                                                    @resdiag resid
*                                                    # constant X X1 ...
* options:
*    [PRINT]/NOPRINT        Prints the series autocorrelations and partial-
*                           autocorrelations up to LBQDOF (see below)
*    LBQTEST=[season freq.] Does Ljung-Box Q-tests for serial correlation 
*                           up to the LBQTESTth order
*    LBQDOF=[0]             Degrees of freedom correction for above test.  If
*                           the residuals are from an estimated ARMA(p,q), use
*                           QDOF=p+q
*    [LMP]/NOLMP            LM tests for serial correlation
*    LMTEST=[season freq.]  sets the LM test to the seasonal frequency of
*                           the data
*    [LMLOOP]/NOLMLOOP      tests up to the LMTESTth order of serial
*                           correlation
*    ARTEST=[season freq.]  sets the simple and restricted AR test to the
*                           seasonal frequency of the data
*    [ARLOOP]/NOARLOOP      tests up to the ARth order of simple serial
*                           correlation
*    [WALLIS]/NOWALLIS      performs the Wallis test for fourth order serial
*                           correlation
*    TESTARCH=[1]           sets the order of the ARCH test
*    [LOOPARCH]/NOLOOPARCH  tests up to the TESTARCHth order of ARCH
*    [BPHET]/NOBPHET        tests the modified Breusch-Pagan
*    RESET/[NORESET]        To do the RESET test, users must specify
*                           the RESET option and supply a second supplementary 
*                           card listing the dependent variable and the model's
*                           fitted values(RESET is appropriate only for models
*                           estimated with OLS).
*
*
procedure resdiag  residual  first last
type series[real] residual
type integer first last
option switch print      1
option integer lbqtest   0
option integer lbqdof    0
option integer lmtest    0
option switch  lmloop    1
option switch  lmp       1
option integer artest    0
option switch  arloop    1
option switch  wallis    1
option integer testarch  0
option switch  looparch  1
option switch  reset     0
option switch  bphet     1
local  integer  fobs lobs nregs lm arch base k j p ar pos neg run zero lbq
local  vector[integer]  reglist rsetlist
local  series[real]  resid2 pfxsq pfxcub pfxqua resnum resden prsd
local  vector[real]  residmom
local  rectangular[real]  newspace
local  real  chistat d4 Zrun Zsign lbqstat
dis; dis
*
inquire(series=residual) fobs lobs
if (first .and. last)
    {;  compute fobs=first;   compute lobs=last;    }
enter(varying,entries=nregs) reglist
dis 'Analyzing residuals in series ' %label(residual) $
        ' from ' %datelabel(fobs) ' to ' %datelabel(lobs)
dis '    using the following series:'
dis(hold) '    '
do j=1,nregs
   dis(hold) %label([series]reglist(j))  '  '
end do j
dis;dis;dis
*
* Durbin-Watson Statistic
*
linreg(noprint) residual fobs lobs
# constant
* Results displayed below under Ljung-Box Q-statistic 
* 
* Ljung-Box Q-Statistic
*
if lbqtest == 0
   compute lbq = %cal(%year(1)+1,1) - %cal(%year(1),1) 
else
   compute lbq = lbqtest
correlate(partial=prsd, print=print) residual fobs lobs resid2 
dis 'Durbin-Watson Statistic:' @30 #.#### %durbin
dis
compute lbqstat = 0.0
do k=lbqdof+1,lbq
  compute lbqstat = lbqstat+%nobs*(%nobs+2)*resid2(k+1)**2/(%nobs-k)
  dis 'Ljung-Box Q-statistic using ' ## k ' autocorrelations:'
  cdf chisqr lbqstat k-lbqdof
end do k
dis
*
*  LM TEST AR(k) or MA(k)
*
* First, decide on the order of the test.
* It defaults to the seasonal frequency
if lmp
 {
if lmtest == 0
   compute lm = %cal(%year(1)+1,1) - %cal(%year(1),1)
else
   compute lm = lmtest
* Test over all intermediate values if desired
if lmloop
  do k=1,lm
    linreg(noprint) residual fobs+k lobs
    # reglist residual{1 to k}
    dis 'LM test for ' k 'th order serial correlation' 
    cdf chisqr %trsq k
  end do k
else 
  {
  linreg(noprint) residual fobs+lm lobs
  # reglist residual{1 to lm}
  dis 'LM test for ' lm 'th order serial correlation' 
  cdf chisqr %trsq lm
 }
 }
dis
*
* Simple AR(k) process tests
*
* default to seasonal frequencies
*
if artest == 0
   compute ar = %cal(%year(1)+1,1) - %cal(%year(1),1)
else
   compute ar = artest
*
* Test over intermediate values if desired
*
if arloop
 do k=1,ar
   linreg(noprint) residual fobs+k lobs
   # residual{k}
   dis 'Testing for a simple AR(' k ') order process'
   cdf ttest %beta(1)/sqrt(%xx(1,1)*%seesq) %nobs-1+k
 end do k
else
 {
  linreg(noprint) residual fobs+ar lobs
  # residual{ar}
  dis 'Testing for a simple AR(' ar ') order process'
  cdf ttest %beta(1)/sqrt(%xx(1,1)*%seesq) %nobs-1+k
  }
dis
*
* Restricted AR(1,k) test
*
if (ar .ne. 1)
 linreg(noprint) residual fobs+ar lobs
 # residual{1} residual{ar}
 exclude(noprint)
 # residual{1} residual{ar}
 dis 'Testing for a restricted AR(1 and' ar ')process'
 dis 'Test statistic =' %cdstat 'with significance' %signif
 dis
*
* Wallis Test
*
if wallis
  {
  set resnum fobs+5 lobs = (residual-residual{4})**2
  acc resnum fobs+5 lobs
  set resden fobs   lobs = residual**2
  acc resden fobs   lobs
  compute d4 = resnum(lobs)/resden(lobs)
  dis 'WALLIS Test for fourth-order serial correlation:' d4
  }
dis
*
* Non-Parametric Tests
*
compute pos = 0
compute run = 1
compute neg = 0
do p = fobs,lobs
  if (residual(p) > 0)
    compute pos = pos+1 
  if (residual(p) < 0)
    compute neg = neg+1
  if p <> fobs
     if (residual(p)*residual(p-1) < 0)
        compute run = run+1
end do p
compute zero= lobs-fobs+1-pos-neg
compute rpos = float(pos)
compute rneg = float(neg)
*
* Runs Test
*
compute zrun = (run-(2*rneg*rpos/(rneg+rpos))+0.5)/sqrt(2*rneg*rpos* $
           (2*rneg*rpos-rneg-rpos)/((rneg+rpos)**2*(rneg+rpos-1)))
*dis 'Number of: Positives =' pos 'Negatives =' neg
*dis '           Runs=      ' run '      Zeros=' zero       
dis 'Runs Test for Randomness'
cdf normal zrun 
dis;dis
*
*  ARCH(k) TEST
*
clear resid2
multiply residual residual fobs lobs resid2
if testarch == 0
  compute arch = 1
else
  compute arch = testarch
* Test over all intermediate values if desired
if looparch
  do k=1,arch
    linreg(noprint) resid2 fobs+k lobs
    # constant resid2{1 to k}
    dis 'Engle(1981) test for ' k 'th order ARCH' 
    cdf chisqr %nobs*%rsquared k
  end do k
else 
  {
  linreg(noprint) resid2 fobs+arch lobs
  # constant resid2{1 to arch}
  dis 'Engle(1981) test for ' arch 'th order ARCH' 
  cdf chisqr %nobs*%rsquared arch
  }
dis
*
* Breusch-Pagan (Koenker actually)
*
if bphet
  {
 linreg(noprint) resid2 fobs lobs
 # reglist 
 dis 'Modified Breusch-Pagan test for conditional heteroscedasticity'
 dis ' using only the original variables:' 
 cdf chisqr %nobs*%rsquared nregs-1
 scratch(matrix=newspace) (nregs-1)  fobs lobs  base
 do k=2,nregs
    multiply reglist(k) reglist(k)  fobs lobs  base+k-1
 end do k
 linreg(noprint) resid2 fobs lobs
 # reglist base+1 to base+nregs-1
 dis ' using the original variables and their squares:' 
 cdf chisqr %nobs*%rsquared 2*(nregs-1)
  }
dis;dis
*
*  RATS Statistics with higher option
*
stat residual fobs lobs
*
* Sign Test
*
compute zsign = (2*pos-(pos+neg)-1)/sqrt(pos+neg)
*dis 'Number of usable observations:' lobs-fobs+1-zero
dis 'Sign Test'
cdf normal zsign
dis
*
* Jarque-Bera test
*
dimension residmom(4)
do k=2,4
  set resid2 fobs lobs = residual**k
  acc resid2 fobs lobs
  compute residmom(k) = resid2(lobs)/%nobs
end do k
compute chistat = (%nobs/6.0)*((residmom(3)/(sqrt(residmom(2)**(3))))**2) + $
              (%nobs/24.0)*((residmom(4)/(residmom(2)**2) - 3)**2)
dis 'Jarque-Bera Test for normality:'
cdf chisqr chistat 2
dis;dis
*
*  RESET Test
*
if reset
  {
  enter(varying) rsetlist
  set pfxsq  fobs lobs = [series]rsetlist(2)**2
  set pfxcub fobs lobs = [series]rsetlist(2)**3
  set pfxqua fobs lobs = [series]rsetlist(2)**4
  dis 'Ramsey RESET test using the following powers of fitted values:'
  cmoment fobs lobs
  # rsetlist(1) reglist pfxsq pfxcub pfxqua
  linreg(cmom,noprint) rsetlist(1) fobs lobs
  # pfxsq reglist
  exclude(noprint)
  # pfxsq 
  dis 'Powers = 2, test statistic = ' %cdstat ' with significance ' %signif
  linreg(cmom,noprint) rsetlist(1) fobs lobs
  # pfxsq pfxcub reglist
  exclude(noprint)
  # pfxsq pfxcub 
  dis 'Powers=2,3, test statistic = ' %cdstat ' with significance ' %signif
  linreg(cmom,noprint) rsetlist(1) fobs lobs
  # pfxsq pfxcub pfxqua reglist
  exclude(noprint)
  # pfxsq pfxcub pfxqua
  dis 'Powers 2-4, test statistic = ' %cdstat ' with significance ' %signif
  }
dis;dis 
*
release newspace
end proc
env echo

 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
 *  RESDIAG.SRC
 *  Copyright 1993 by the Bank of Canada                                 *
 *  Programmed for RATS by Robert Amano, Jeff Gable and Simon van Norden *
 *  In exchange for access to these procedures, users are requested to   *
 *  acknowledge the use of the Bank of Canada RATS procedures in         *
 *  all published work.                                                  *
 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

</x-charset>
From ???@??? Mon Mar 13 15:10:09 2000
Received: from SpoolDir by EFS01 (Mercury 1.40); 11 Mar 100 16:20:49 GMT+1000
From: Azali Ong Mari <azali@econ.upm.edu.my>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: 2 s.e. upper & lower bounds Impulse response functions for  Berna
Date: Sat, 11 Mar 2000 13:19:12 +0000
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.6 (32) (via Mercury MTS (Bindery) v1.40)
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"

Dear All,

Does anyone can help me how to write a programme to calculate 2 s.e. upper
& lower bounds impulse response functions for Bernanke.src.

I would be grateful for any help, thanks in advance.

Regards,



Azali Ong Mari
Department of Economics
Faculty of Economics and Management
University Putra Malaysia
43400 SERDANG Selangor
Malaysia


Tel No (o) :   + 603 - 9486101  xtn 1671
email      :   azali@econ.upm.edu.my

From ???@??? Tue Mar 14 21:02:24 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id UAA07230
	for <robt@localhost>; Tue, 14 Mar 2000 20:55:05 +1100
Message-Id: <200003140955.UAA07230@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Tue, 14 Mar 2000 20:55:05 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 14 Mar 100 20:49:53 GMT+1000
From: "VIJAYA SARATHI N." <SARATHI@FPMLAN.iimahd.ernet.in>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: ged distribution
Date: Tue, 14 Mar 2000 14:28:25 GMT+0530
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Organization:  IIM, Ahmedabad
X-mailer: Pegasus Mail v3.31 (via Mercury MTS (Bindery) v1.40)

hello colleagues,

How can I generate a GED distribution through RATS?

Thanks in advance for your help,

Regards,

sarathi


From ???@??? Wed Mar 15 08:01:01 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id BAA12730
	for <robt@localhost>; Wed, 15 Mar 2000 01:00:04 +1100
Message-Id: <200003141400.BAA12730@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Wed, 15 Mar 2000 01:00:04 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 15 Mar 100 00:37:17 GMT+1000
From: "Christopher F. Baum" <baum@bc.edu>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: ged distribution
Date: Tue, 14 Mar 2000 08:37:46 -0500
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
X-Mailer: Mulberry (MacOS) [1.4.5, s/n P020-300786-009] (via Mercury MTS (Bindery) v1.40)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Take a look at Norman Morin's GARCH.SRC on the Estima web site. It allows
for GED errors, and you can see how it is done in the code (search for
DIST.eq.3).

Kit Baum
Boston College Economics

--On Tue, Mar 14, 2000 14:28 +0000 "VIJAYA SARATHI N."
<SARATHI@FPMLAN.iimahd.ernet.in> wrote:

> hello colleagues,
> 
> How can I generate a GED distribution through RATS?
> 
> Thanks in advance for your help,
> 
> Regards,
> 
> sarathi
> 



-----------------------------------------------------------------
Kit Baum    baum@bc.edu    http://fmwww.bc.edu/ec-v/baum.fac.html

From ???@??? Wed Mar 15 08:02:41 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id GAA10954
	for <robt@localhost>; Wed, 15 Mar 2000 06:35:04 +1100
Message-Id: <200003141935.GAA10954@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Wed, 15 Mar 2000 06:35:04 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 15 Mar 100 06:30:44 GMT+1000
From: "Estima" <tomm@estima.com>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: ged distribution
Date: Tue, 14 Mar 2000 12:53:02 -0600
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Organization: Estima
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
X-mailer: Pegasus Mail for Win32 (v3.11) (via Mercury MTS (Bindery) v1.40)

> hello colleagues,
> 
> How can I generate a GED distribution through RATS?
> 
> Thanks in advance for your help,
> 
> Regards,
> 
> sarathi

If by "generate" you mean creating draws from, the procedure below 
will do it.

Tom Doan
Estima


proc ged x n
type vect *x
type integer n
*
*  Draws for a generalized error distribution by acceptance-
*  rejection method.
*
*  Written March, 2000 by Tom Doan, Estima.
*
*  Syntax: @GED(nu=shape parameter) x n
*    x = vector of reals (output)
*    n = desired dimension of n
*    nu = shape parameter. Should be >=1.
*
option real nu 2.0
local integer i
local real e r
local real scalef
if nu<1.0 {
   disp '**** NU = ' nu 'is illegal. It should be >=1 ****'
   return
}
*
*  For convenience, we're using a simple one-sided exponential
*  variate as the comparison. The kernel of the target distribution
*  is exp(-x**nu). This needs to be symmetrized and scaled to
*  a variance of 1.0 at the end. The following computes the
*  needed scale factor.
*
compute scalef=exp(.5*(%lngamma(1.0/nu)-%lngamma(3.0/nu)))
dim x(n)
do i=1,n
   loop
      compute e = -log(%uniform(0.0,1.0))
      compute r = exp(-(e**nu))/exp(-e)
      if r>%uniform(0.0,1.0)
         break
   end loop
   compute x(i)=%if(%uniform(0.0,1.0)>.5,e,-e)*scalef
end do i
end



--
------------------------------------------------------------
|   Estima                    |  Sales:   (800) 822-8038   |
|   P.O. Box 1818             |           (847) 864-8772   |
|   Evanston, IL 60204-1818   |  Support: (847) 864-1910   |
|   USA                       |  Fax:     (847) 864-6221   |
|   http://www.estima.com     |  estima@estima.com         |
------------------------------------------------------------

From ???@??? Sun Mar 19 17:52:00 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id MAA14032
	for <robt@localhost>; Sat, 18 Mar 2000 12:55:04 +1100
Message-Id: <200003180155.MAA14032@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Sat, 18 Mar 2000 12:55:04 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 18 Mar 100 12:53:58 GMT+1000
From: "Yuan Sun" <ysun@bsa3.kent.edu>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: kalman filter
Date: Fri, 17 Mar 2000 21:04:10 -0800
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
Content-Type: text/plain;
Content-Transfer-Encoding: 7bit
X-Mailer: Microsoft Outlook Express 5.00.2615.200 (via Mercury MTS (Bindery) v1.40)


  I have a measurement equation: y = a + b*x + e, and the state vector   bt
= p*b t-1 + Vt, where var(Vt) = Mt. I want to use the kalman filter to
estimate the two equations together and get the estimate of b and Mt. How
can I do it in Rats?


  Tiffany Sun


From ???@??? Fri Mar 24 09:46:28 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id GAA08396
	for <robt@localhost>; Fri, 24 Mar 2000 06:00:04 +1100
Message-Id: <200003231900.GAA08396@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Fri, 24 Mar 2000 06:00:04 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 24 Mar 100 05:39:10 GMT+1000
From: Ying-Man Chung <y_mchung@yahoo.com>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: RE: 250-days moving average of stock prices
Date: Thu, 23 Mar 2000 10:39:17 -0800 (PST)
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mercury MTS (Bindery) v1.40

Dear Rats users,

I want to draw a line of 250-days moving average of
the stock price every day. Anybody can show my the
program to calculate the moving average, please?

regards,

Sarah,

HKSYC

__________________________________________________
Do You Yahoo!?
Talk to your friends online with Yahoo! Messenger.
http://im.yahoo.com

From ???@??? Fri Mar 24 09:46:42 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id GAA11578
	for <robt@localhost>; Fri, 24 Mar 2000 06:10:05 +1100
Message-Id: <200003231910.GAA11578@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Fri, 24 Mar 2000 06:10:05 +1100 (EETDT)
Received: from SpoolDir by EFS01 (Mercury 1.40); 24 Mar 100 06:08:25 GMT+1000
From: "Iacoviello,M  (pg)" <M.Iacoviello@lse.ac.uk>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: RE: 250-days moving average of stock prices
Date: Thu, 23 Mar 2000 19:07:30 -0000
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21) (via Mercury MTS (Bindery) v1.40)
Content-Type: text/plain

Use the following command

mvstats(span=250,means=avstockprice) stockprice

Replace avstockprice with the name that you want to assign to the averaged
variable, and stockprice with the original name of your variable, of course.

Regards

Matteo

> ----------
> From: 	Ying-Man Chung[SMTP:y_mchung@yahoo.com]
> Reply To: 	RATS Discussion List
> Sent: 	23 March 2000 18:39
> To: 	RATS Discussion List
> Subject: 	RE: 250-days moving average of stock prices
> 
> Dear Rats users,
> 
> I want to draw a line of 250-days moving average of
> the stock price every day. Anybody can show my the
> program to calculate the moving average, please?
> 
> regards,
> 
> Sarah,
> 
> HKSYC
> 
> __________________________________________________
> Do You Yahoo!?
> Talk to your friends online with Yahoo! Messenger.
> http://im.yahoo.com
> 

From ???@??? Mon Mar 27 07:45:54 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id FAA15674
	for <robt@localhost>; Mon, 27 Mar 2000 05:30:04 +1000
Message-Id: <200003261930.FAA15674@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Mon, 27 Mar 2000 05:30:04 +1000 (EET)
Received: from SpoolDir by EFS01 (Mercury 1.40); 27 Mar 100 05:02:41 GMT+1000
From: Xi Li <Xi.Li@owen.vanderbilt.edu>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: constraints on GARCH
Date: Sun, 26 Mar 2000 13:03:02 -0600
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21) (via Mercury MTS (Bindery) v1.40)
Content-Type: text/plain;

Dear friends,

I wonder if any one know how to put positivity constraints on estimated
coefficients of GARCH process and if we need to adjust the stadard error,
etc. for the estimated coefficient. And any one know how to impose
restriction that the sum of the ARCH process coefficients is less than 1?

Thanks a lot.

Xi



From ???@??? Mon Mar 27 07:47:05 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id GAA16624
	for <robt@localhost>; Mon, 27 Mar 2000 06:50:04 +1000
Message-Id: <200003262050.GAA16624@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Mon, 27 Mar 2000 06:50:04 +1000 (EET)
Received: from SpoolDir by EFS01 (Mercury 1.40); 27 Mar 100 06:48:03 GMT+1000
From: Xi Li <Xi.Li@owen.vanderbilt.edu>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Date: Sun, 26 Mar 2000 14:48:31 -0600
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21) (via Mercury MTS (Bindery) v1.40)
Content-Type: text/plain;

Dear Rob,

Sorry to send you the second message. I found that you have written
multivariate GARCH procedure in Estima's procedure page. I wonder if your
version of procedure include the positivity constraints on coefficients in
variance processes?

Thanks a lot and hope you have a nice day.

Best regards,

Xi

From ???@??? Mon Mar 27 07:47:03 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id GAA16620
	for <robt@localhost>; Mon, 27 Mar 2000 06:50:04 +1000
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Mon, 27 Mar 2000 06:50:04 +1000 (EET)
Received: from SpoolDir by EFS01 (Mercury 1.40); 27 Mar 100 06:48:02 GMT+1000
Received: from baldrick.ocs.mq.edu.au (137.111.1.12) by efs01.efs.mq.edu.au (Mercury 1.40) with ESMTP;
    27 Mar 100 06:48:00 GMT+1000
Received: from mailserver.vanderbilt.edu (owen.Vanderbilt.Edu [129.59.210.168])
	by baldrick.ocs.mq.edu.au (8.9.2/8.9.2) with ESMTP id GAA09089;
	Mon, 27 Mar 2000 06:48:32 +1000 (EST)
Received: by owen.Vanderbilt.Edu with Internet Mail Service (5.5.2650.21)
	id <HM36TCWS>; Sun, 26 Mar 2000 14:48:32 -0600
Message-ID: <8A33E15E48FFD311BD6C0008C756CD1B59B9D6@owen.Vanderbilt.Edu>
From: Xi Li <Xi.Li@owen.vanderbilt.edu>
To: "'RATS-L@EFS.MQ.EDU.AU'" <RATS-L@EFS.MQ.EDU.AU>
Cc: "'RATS-L-OWNER@EFS.MQ.EDU.AU'" <RATS-L-OWNER@EFS.MQ.EDU.AU>
Date: Sun, 26 Mar 2000 14:48:31 -0600
MIME-Version: 1.0
X-Mailer: Internet Mail Service (5.5.2650.21)
Content-Type: text/plain;
	charset="iso-8859-1"

<x-charset iso-8859-1>Dear Rob,

Sorry to send you the second message. I found that you have written
multivariate GARCH procedure in Estima's procedure page. I wonder if your
version of procedure include the positivity constraints on coefficients in
variance processes?

Thanks a lot and hope you have a nice day.

Best regards,

Xi

</x-charset>
From ???@??? Tue Mar 28 10:09:48 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id HAA08434
	for <robt@localhost>; Tue, 28 Mar 2000 07:45:05 +1000
Message-Id: <200003272145.HAA08434@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Tue, 28 Mar 2000 07:45:05 +1000 (EET)
Received: from SpoolDir by EFS01 (Mercury 1.40); 28 Mar 100 07:43:42 GMT+1000
From: SsnaithIPR@aol.com
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: ERRORS command
Date: Mon, 27 Mar 2000 16:43:36 EST
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
X-Mailer: AOL 4.0 for Windows 95 sub 61 (via Mercury MTS (Bindery) v1.40)

s there any way to get to a specific value saved with the save option in the 
Errors command.  I want to access values for a specific equation, step and 
variable from a decomposition of forecast variance

Or perhaps more generally, how does one access a single element when you have 
a vector of arrays.


Regards,

Sean

From ???@??? Wed Mar 29 09:28:55 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id JAA18340
	for <robt@localhost>; Wed, 29 Mar 2000 09:25:04 +1000
Message-Id: <200003282325.JAA18340@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Wed, 29 Mar 2000 09:25:04 +1000 (EET)
Received: from SpoolDir by EFS01 (Mercury 1.40); 29 Mar 100 09:21:33 GMT+1000
From: "Estima" <tomm@estima.com>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: ERRORS command
Date: Tue, 28 Mar 2000 10:38:11 -0600
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Organization: Estima
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
X-mailer: Pegasus Mail for Win32 (v3.11) (via Mercury MTS (Bindery) v1.40)

> s there any way to get to a specific value saved with the save option in the 
> Errors command.  I want to access values for a specific equation, step and 
> variable from a decomposition of forecast variance
> 
> Or perhaps more generally, how does one access a single element when you have 
> a vector of arrays.
> 

As (sort of) mentioned on page 14-75, you can get the fraction of the 
variance for:

equation i,
at step k,
that is explained by variable j,

from:

decomp(i)(k,j)

For example, for equation 1, step 2, explained by variable 3:

compute d = decomp(1)(2,3)

Note that there should be no spaces anywhere in "decomp(i)(k,j)".

What you have is a VECTOR of RECTANGULAR arrays, so the first set of 
parens contains a one-dimensional index number--indicating which 
element of the VECTOR you are referring to. The second set of parens 
indicates which element of that RECTANGULAR array you are referring 
to.

If you instead had a RECTANGUALR array of VECTORS, for example, a 
reference to a single number would take the form:

   RV(x,y)(z)

Sincerely,
Tom Maycock
Estima

--
------------------------------------------------------------
|   Estima                    |  Sales:   (800) 822-8038   |
|   P.O. Box 1818             |           (847) 864-8772   |
|   Evanston, IL 60204-1818   |  Support: (847) 864-1910   |
|   USA                       |  Fax:     (847) 864-6221   |
|   http://www.estima.com     |  estima@estima.com         |
------------------------------------------------------------

From ???@??? Wed Mar 29 22:12:30 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id WAA11024
	for <robt@localhost>; Wed, 29 Mar 2000 22:10:03 +1000
Message-Id: <200003291210.WAA11024@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Wed, 29 Mar 2000 22:10:04 +1000 (EET)
Received: from SpoolDir by EFS01 (Mercury 1.40); 29 Mar 100 22:05:51 GMT+1000
From: "Stephan Kohns" <kohns@iiw.uni-bonn.de>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Use of INPUT in a procedure
Date: Wed, 29 Mar 2000 14:05:41 +0100
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
MIME-Version: 1.0
Content-Type: multipart/alternative;
X-Mailer: Microsoft Outlook Express 5.00.2919.6600 (via Mercury MTS (Bindery) v1.40)

------=_NextPart_000_0045_01BF9987.DE347E00
Content-Type: text/plain;
	charset="iso-8859-1"

<x-charset iso-8859-1>Dear RATS users,
 I want to initialize a rectangular matrix within a procedure using the INPUT-command. Consider the following sample program and procedure (what I actually want to do is to write a procedure that computes certain critical values, and the matrix would contain parameters needed for the calculation):

Main program:
dec real r
dec int i j
source(echo) d:\winrats\inputtst.src
com r = 0.0
com i = 2, j = 3
@inputtest i j r
dis r

Procedure:
procedure inputtest i j wert
type integer i j
type real *wert
local rect[real] testm
dim testm(1,3)
dis 'before input'
input testm
 1.1 1.2 1.3
 2.1 2.2 2.3
 3.1 3.2 3.3
 4.1 4.2 4.3
dis 'after input'
com wert = testm(i,j)
end

When running the program, I get the following output:
procedure inputtest i j wert
(01.0016) type integer i j
(01.0016) type real *wert
(01.0016) local rect[real] testm
(01.0030) dim testm(1,3)
(01.0049) dis 'before input'
(01.0064) input testm
(01.0077)  1.1 1.2 1.3
(01.0104)  2.1 2.2 2.3
(01.0131)  3.1 3.2 3.3
(01.0158)  4.1 4.2 4.3
(01.0185) dis 'after input'
(01.0199) com wert = testm(i,j)
(01.0225) end

before input
## IO40. Expected real number. Got "dis".
The Error Occurred At Location 0076 of

 Does anybody have an idea what is going wrong here? The input-section works smoothly in the main program, but somehow the procedure seems to get stuck after the input command (note that the second 'dis' is not executed). The error message appears when executing the display-command 'dis r' in the main program, i.e. after the procedure has been finished.

 Thanks a lot for your help
 Stephan Kohns

------extPart_000_0045_01BF9987.DE347E00
Content-Type: text/html;
	charsetso-8859-1"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META content="text/html; charset=iso-8859-1" http-equiv=Content-Type>
<META content="MSHTML 5.00.2919.6307" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face=Arial size=2>Dear RATS users,</FONT></DIV>
<DIV><FONT face=Arial size=2>&nbsp;I want to initialize a rectangular matrix 
within a procedure using the INPUT-command. Consider the following sample 
program and procedure (what I actually want to do is to write a procedure that 
computes certain critical values, and the matrix&nbsp;would contain parameters 
needed for the calculation):</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT face=Arial size=2>Main program:</FONT></DIV>
<DIV><FONT face=Arial size=2>dec real r<BR>dec int i j<BR>source(echo) 
d:\winrats\inputtst.src<BR>com r = 0.0<BR>com i = 2, j = 3<BR>@inputtest i j 
r<BR>dis r</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT face=Arial size=2>Procedure:</FONT></DIV>
<DIV><FONT face=Arial size=2>procedure inputtest i j wert<BR>type integer i 
j<BR>type real *wert<BR>local rect[real] testm<BR>dim testm(1,3)<BR>dis 'before 
input'<BR>input testm<BR>&nbsp;1.1 1.2 1.3<BR>&nbsp;2.1 2.2 2.3<BR>&nbsp;3.1 3.2 
3.3<BR>&nbsp;4.1 4.2 4.3<BR>dis 'after input'<BR>com wert = 
testm(i,j)<BR>end</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT face=Arial size=2>When running the program, I get the following 
output:</FONT></DIV>
<DIV><FONT face=Arial size=2>procedure inputtest i j wert<BR>(01.0016) type 
integer i j<BR>(01.0016) type real *wert<BR>(01.0016) local rect[real] 
testm<BR>(01.0030) dim testm(1,3)<BR>(01.0049) dis 'before input'<BR>(01.0064) 
input testm<BR>(01.0077)&nbsp; 1.1 1.2 1.3<BR>(01.0104)&nbsp; 2.1 2.2 
2.3<BR>(01.0131)&nbsp; 3.1 3.2 3.3<BR>(01.0158)&nbsp; 4.1 4.2 4.3<BR>(01.0185) 
dis 'after input'<BR>(01.0199) com wert = testm(i,j)<BR>(01.0225) 
end</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT face=Arial size=2>before input<BR>## IO40. Expected real number. Got 
"dis".<BR>The Error Occurred At Location 0076 of</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT face=Arial size=2>&nbsp;Does anybody have an idea what is going wrong 
here? The input-section works smoothly in the main program, but somehow the 
procedure seems to get stuck after the input command (note that the second 'dis' 
is not executed). The error message appears when executing the display-command 
'dis r' in the main program, i.e. after the procedure has been 
finished.</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT face=Arial size=2>&nbsp;Thanks a lot for your help</FONT></DIV>
<DIV><FONT face=Arial size=2>&nbsp;Stephan Kohns</FONT></DIV></BODY></HTML>

------extPart_000_0045_01BF9987.DE347E00--


</x-charset>
From ???@??? Thu Mar 30 08:40:43 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id IAA07282
	for <robt@localhost>; Thu, 30 Mar 2000 08:40:04 +1000
Message-Id: <200003292240.IAA07282@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Thu, 30 Mar 2000 08:40:04 +1000 (EET)
Received: from SpoolDir by EFS01 (Mercury 1.40); 30 Mar 100 08:35:48 GMT+1000
From: "Estima" <tomm@estima.com>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re: Use of INPUT in a procedure
Date: Wed, 29 Mar 2000 16:18:28 -0600
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Organization: Estima
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
X-mailer: Pegasus Mail for Win32 (v3.11) (via Mercury MTS (Bindery) v1.40)

> 
> Main program:
> dec real r
> dec int i j
> source(echo) d:\winrats\inputtst.src
> com r =3D 0.0
> com i =3D 2, j =3D 3
> @inputtest i j r
> dis r
> 
> Procedure:
> procedure inputtest i j wert
> type integer i j
> type real *wert
> local rect[real] testm
> dim testm(1,3)
> dis 'before input'
> input testm
>  1.1 1.2 1.3
>  2.1 2.2 2.3
>  3.1 3.2 3.3
>  4.1 4.2 4.3
> dis 'after input'
> com wert =3D testm(i,j)
> end
> 

The INPUT and READ commands don't work the way we would really like 
them to in a compiled section. Instead of handling the above 
situations properly, they will look for the input data after the 
procedure call.

So, the solution would be to do one of the following:

1) Move the data to the main program following the procedure call. 
For example, you would delete the data block from the procedure and 
(assuming you change the dimensions of TESTM to (4,3), do the 
following in your main program:

@inputtest i j r
1.1 1.2 1.3
2.1 2.2 2.3
3.1 3.2 3.3
4.1 4.2 4.3

2) Input the data in your procedure file, but do it before the 
PROCEDURE command. This means you'll be putting it into a global 
variable. For example:

declare rect testm
dim testm(4,3)
input testm
1.1 1.2 1.3
2.1 2.2 2.3
3.1 3.2 3.3
4.1 4.2 4.3
procedure inputtest i j wert
type integer i j
type real *wert
com wert =testm(i,j)
end

3) Use a COMPUTE command to set up the matrix variable.

procedure inputtest i j wert
type integer i j
type real *wert
local rect[real] testm
compute testm = || 1.1, 1.2, 1.3 | $
                   2.1, 2.2, 2.3 | $
                   3.1, 3.2, 3.3 | $
                   4.1, 4.2, 4.3 ||
etc.

Sincerely,
Tom Maycock
Estima


--
------------------------------------------------------------
|   Estima                    |  Sales:   (800) 822-8038   |
|   P.O. Box 1818             |           (847) 864-8772   |
|   Evanston, IL 60204-1818   |  Support: (847) 864-1910   |
|   USA                       |  Fax:     (847) 864-6221   |
|   http://www.estima.com     |  estima@estima.com         |
------------------------------------------------------------

From ???@??? Thu Mar 30 11:15:36 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id LAA13828
	for <robt@localhost>; Thu, 30 Mar 2000 11:15:05 +1000
Message-Id: <200003300115.LAA13828@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Thu, 30 Mar 2000 11:15:05 +1000 (EET)
Received: from SpoolDir by EFS01 (Mercury 1.40); 30 Mar 100 11:13:15 GMT+1000
From: Rob Trevor <robt@mafc.mq.edu.au>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Re:
Date: Thu, 30 Mar 2000 11:13:27 +1000
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii" ; format="flowed"
X-Mailer: Mercury MTS (Bindery) v1.40

<x-flowed>There was a discussion you will find in the archives re different 
ways of forcing a parameter to be non-negative. From memory:

	abs(a) or sqrt(a)*sqrt(a) or sqrt(a*a)

will work and give correct asymptotic se's - some a bit better than 
others. (I've never had trouble with sqrt(a)*sqrt(a).)

 From memory the more recent garch procedure on the Estima web site 
incorporates non-negativity constraints. (The one written by Normin 
Morin.) If it doesn't, it should be easy to add. (If you want to 
ensure that the parameters sum to less than 1, just define one of the 
parameters as the residual and force that to be non-negative.)

Cheers

Rob


At 2:48 PM -0600 26/3/00, Xi Li wrote:
>Dear Rob,
>
>Sorry to send you the second message. I found that you have written
>multivariate GARCH procedure in Estima's procedure page. I wonder if your
>version of procedure include the positivity constraints on coefficients in
>variance processes?
>
>Thanks a lot and hope you have a nice day.
>
>Best regards,
>
>Xi

</x-flowed>
From ???@??? Fri Mar 31 09:25:43 2000
Received: from localhost (localhost [127.0.0.1])
	by amethyst.efs.mq.edu.au (8.9.1/8.9.1) with ESMTP id HAA08150
	for <robt@localhost>; Fri, 31 Mar 2000 07:55:04 +1000
Message-Id: <200003302155.HAA08150@amethyst.efs.mq.edu.au>
Received: from efs01.efs.mq.edu.au
	by localhost with POP3 (fetchmail-5.0.0)
	for robt@localhost (single-drop); Fri, 31 Mar 2000 07:55:05 +1000 (EET)
Received: from SpoolDir by EFS01 (Mercury 1.40); 31 Mar 100 07:51:25 GMT+1000
From: Vivek Ghosal <Vivek.Ghosal@usdoj.gov>
To: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Subject: Panel data
Date: Thu, 30 Mar 2000 10:04:00 -0500
Errors-to: <rats-l-owner@efs.mq.edu.au>
Reply-to: "RATS Discussion List" <RATS-L@efs.mq.edu.au>
Sender: Maiser@efs01.efs.mq.edu.au
X-listname: <RATS-L@efs01.efs.mq.edu.au>
X-Mailer: Worldtalk (NetJunction 4.6-p7)/MIME (via Mercury MTS (Bindery) v1.40)

Hello - 

I have a panel with N (=350) cross-section units and T (=40) time series.
The data are already sorted by N and then T. I will run a regression for
each of the N units; i.e., first for N=1, then N=2,....

Question: Is there a way to output all the estimates (and standard errors)
in a convenient format. That is, I do not want to look through each of the
350 regressions, but want the parameter est's and s.e.'s in some format
like...(my regression has multiple beta's):

N      beta                s.e. (if possible, here or separately)
1      beta(1,1)         s.e.(1,1)
2      beta(2,1)
3
..
..
350   beta(350,1)    s.e.(350,1)

(Also, if possible, output the regression R-square or any summary
statistic.)

My ultimate objective is to use these beta's in a second-stage
cross-section regression.

I can do some of this in SAS fairly easily, but not sure how to do it
in RATS.

Many thanks in advance.

Vivek