Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Information for data used.

Subject   Re: st: Information for data used.
Date   Tue, 15 Jan 2013 05:12:47 +0000

Dear Patrick,
Thanks so much, I appreciate your effort. I will go over it.
Sent from my BlackBerry wireless device from MTN

-----Original Message-----
From: annoporci <>
Date: Thu, 17 Jan 2013 06:18:13 
To: <>
Reply-To: statalist@hsphsun2.harvard.eduSubject: Re: st: Information for data used.

Dear Sola,

I don't know if the dataset is available online, but you can find another
illustration of Stata's mgarch commands here:

where the dataset is available from the Stata website, via the command:

      webuse stocks

Using that dataset instead, you would be doing something like:

      clear all
      webuse stocks
      tsset t
      mgarch dcc (toyota nissan honda =, noconstant) , arch(1) garch(1)
      tsappend, add(50)
      tsset date
      predict H*, variance dynamic(td(31dec2010))
      tsline H_nissan_toyota H_honda_toyota H_honda_nissan, legend(rows(3))

In the above I used tsset t before mgarch to ensure that lags were
computed wrt to adjacent trading/business dates rather than calendar dates.
After tsappend I then tsset date in order to use tsline and graph wrt to
calendar dates.

In fact, I couldn't get tline(td(31dec2010)) to work, except like this:

tsline [...] tline(`=td(31dec2010)')

or like this:

scalar fdate = td(31dec2010)
tsline [...] tline(`=fdate')

tline(td(31dec2010)) is the syntax suggested by Rafal Raciborski, so there  
must be a way to make it work. Note that in the Stata help page cited  
above, tline is used with the value of t in the dataset, tline(2015) or  
tline(2016), depending on whether you include the last sample date or not;  
but for that to work you need to tsset wrt to t, otherwise if you tsset  
with the calendar date, you need to use tline(18627) (18627 is the  
distance to some time in the 1960s), which you can find from:

di td(31dec2010)

I think it's natural to plot wrt to calendar date, rather than wrt to
business date, no?

Now I'm very new to this and I'm not sure at all that's the right way to
do it, so if someone is listening it would be nice to hear if there's an


Patrick Toche.

On Thu, 17 Jan 2013 05:19:42 +0800, olorunfemi sola <>

> Hi All,
> Sorry i did not include subject in the previous request i posted.
> I read through The Stata News- vol 27 no 3, just sent to me.
>  The topic on Multivariate GARCH (MGARCH) written by Rafal Raciborski is
>  quite interesting and topical. Can anyboby please direct me to the
> address where I can get the data the author used so that I can replicate
>  his result.
> I appreciate you in advance.
> Olorunfemi, Sola.
> ***********************************************************************
> official e-mail:
> TEL NO +234 803 581 0893
> **********************************************************************
> *
> *   For searches and help try:
> *
> *
> *

Patrick Toche.
*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index