Dynamic conditional correlation multivariate GARCH

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

Abdelrazzaq
Posts: 13
Joined: Thu Mar 27, 2014 9:17 am

Re: Dynamic conditional correlation multivariate GARCH

Postby Abdelrazzaq » Wed Sep 23, 2015 5:19 am

Dear all;

I am trying to use the written code as available here, in my analysis and I am having an error.

Please see attached workfile, code and error message and advice.

best
Abdelrazzaq
Attachments
screenshot.docx
(210.09 KiB) Downloaded 733 times
code.prg
(6.41 KiB) Downloaded 713 times
workfile.WF1
(1.6 MiB) Downloaded 671 times

ANASEER
Posts: 1
Joined: Wed Oct 28, 2015 9:33 pm

Re: Dynamic conditional correlation multivariate GARCH

Postby ANASEER » Wed Oct 28, 2015 9:41 pm

i need help on DCC MGARCH in eviews 8.
how can i apply this technique on my data, comprising of 10 variables?
thank you in anticipation

trubador
Did you use forum search?
Posts: 1518
Joined: Thu Nov 20, 2008 12:04 pm

Re: Dynamic conditional correlation multivariate GARCH

Postby trubador » Fri Oct 30, 2015 5:30 am

ANASEER wrote:i need help on DCC MGARCH in eviews 8.how can i apply this technique on my data, comprising of 10 variables?thank you in anticipation

Unfortunately, there is no easy way to apply it to a system of 10 variables. You can use dccgarch11 add-in for at most 5 variables. I really do not think that it is too restrictive in practice, since expecting multivariate system to have a common correlation structure may become unrealistic as the system gets larger.

demoisellesalma
Posts: 4
Joined: Sun Nov 01, 2015 4:48 pm

Re: Dynamic conditional correlation multivariate GARCH

Postby demoisellesalma » Sun Nov 01, 2015 5:26 pm

hiii :) please I need help
I've estimated a DCC-GARCH (1,1) to residuals obtained from the adequate ARMA(p,q)applied to my series of stock . I've got results and everything seems to be ok untill checking the equation of univariate GARCH outcome from the DCC . in fact, for all my results the R_squared are negative !!! what should I do please

trubador
Did you use forum search?
Posts: 1518
Joined: Thu Nov 20, 2008 12:04 pm

Re: Dynamic conditional correlation multivariate GARCH

Postby trubador » Sun Nov 01, 2015 11:47 pm

demoisellesalma wrote:hiii :) please I need help
I've estimated a DCC-GARCH (1,1) to residuals obtained from the adequate ARMA(p,q)applied to my series of stock . I've got results and everything seems to be ok untill checking the equation of univariate GARCH outcome from the DCC . in fact, for all my results the R_squared are negative !!! what should I do please


viewtopic.php?f=18&t=8434

demoisellesalma
Posts: 4
Joined: Sun Nov 01, 2015 4:48 pm

Re: Dynamic conditional correlation multivariate GARCH

Postby demoisellesalma » Mon Nov 02, 2015 5:06 am

thank u @trubador for your quick reply

juniorgc4
Posts: 1
Joined: Tue Mar 29, 2016 11:55 am

Re: Dynamic conditional correlation multivariate GARCH

Postby juniorgc4 » Tue Mar 29, 2016 1:13 pm

Hi.
Could anybody help me wtih this question: I am using DCC GARCH, but I see on the third box it says AR lags, it means that mean equation just have autorregresive component? It does not use ARIMA models?. If I can use ARIMA models with DCC GARCH add-in how can I do it.
I really appreciate your help.

Best regards.

trubador
Did you use forum search?
Posts: 1518
Joined: Thu Nov 20, 2008 12:04 pm

Re: Dynamic conditional correlation multivariate GARCH

Postby trubador » Tue Mar 29, 2016 2:16 pm

Unfortunately, the add-in does not take into account MA specification for the mean equation. Some of the MA effects can be mitigated by the GARCH specification, but that really depends on the data. So, I suggest you to try modeling with the available options and then test the results further to see if there is any remaining MA behavior. Or, you can always build your own model from scratch.

ecko33
Posts: 3
Joined: Tue Jul 26, 2016 3:45 pm

Re: Dynamic conditional correlation multivariate GARCH

Postby ecko33 » Tue Jul 26, 2016 3:49 pm

Dear members, after i used dcc garch code as provided, my theta (t1) is negative and p-value is more than 0.05. I think there was something wrong. Could you guys guide me how to deal with it. Thank you so much.

Dimitra
Posts: 2
Joined: Mon Aug 01, 2016 8:41 am

Re: Dynamic conditional correlation multivariate GARCH

Postby Dimitra » Fri Aug 19, 2016 10:14 am

ruhee.mittal wrote:i am not able to get answers with this code.
is this code incomplete??
please help.


For me the code gave me the error that my data lose observations. Just when you write the period on the code, remember to start from a day later.
example: if my sample is 08/16/2003 to 08/16/2005, in the code I will write 08/17/2003 08/16/2005.
Please note that after you run the code and you get the graph, you should re-adjust your sample period on the eviews workfile to the "@all", because after you run the program the sample changes automatically!

deemer
Posts: 1
Joined: Sat Sep 24, 2016 4:26 am

Negative Coefficients - DCC GARCH Model

Postby deemer » Sat Sep 24, 2016 4:42 am

Hi,

I am currently doing a thesis in relation to dynamic correlation between oil returns and stock market indices returns. However, when using the DCC-GARCH model I am getting negative T(1) coefficients. For some indices T(1) is being negative with a p-value < 0.1 and for some other indices the p-value > 0.1. Although I tried different starting values, I am still getting negative coefficients.

Both the DCC code and eviews workfiles are being attached.

Would be very grateful for your help. Thank you !
Attachments
dcc.prg
(2.28 KiB) Downloaded 638 times
dcc_eurostoxx_oil.wf1
(839.77 KiB) Downloaded 508 times
dcc_nikkei_oil.wf1
(822.02 KiB) Downloaded 476 times

alphaomega
Posts: 3
Joined: Fri Nov 04, 2016 12:45 pm

Re: Dynamic conditional correlation multivariate GARCH

Postby alphaomega » Fri Nov 04, 2016 12:50 pm

Hi,

It is not abundantly clear if the DCC-GARCH add-in uses the corrected form the model (as presented in Aelli (2013) - Dynamic Conditional Correlation: On Properties and Estimation.) Can you confirm if the DCC-GARCH add-in uses the corrected cDCC-GARCH? If not, are there any plans to update the add-in to implement the consistent procedure?

Thanks!

anzh
Posts: 2
Joined: Sat Jul 02, 2016 5:29 am

Re: Dynamic conditional correlation multivariate GARCH

Postby anzh » Thu Aug 24, 2017 6:59 am

Hello

I have estimated the conditional correlation using DCC-GARCH and got insignificant DCC-alpha while DCC-beta is strongly significant. I have tried to estimate CCC instead but also got insignificant correlation. Could you please help and advise what could be the reason and possible solution?

Many thanks!

marek1
Posts: 1
Joined: Mon Nov 13, 2017 4:31 am

Re: Dynamic conditional correlation multivariate GARCH

Postby marek1 » Tue Nov 14, 2017 4:14 am

Hello, could you please advise how to put the confidence bands to the graph of dynamic conditional correlations in Eviews? The DCC model is calculated using the code. My serial number is: 90S03555. Thank you, Marek

Raqia
Posts: 4
Joined: Mon Dec 11, 2017 11:54 pm

Re: Dynamic conditional correlation multivariate GARCH

Postby Raqia » Wed Dec 20, 2017 3:52 am

I am working on correlations between stock and bond markets . For instance, I want to determine first correlation between bonds and stocks and then using that correlation as dependent variable. I want to look possible factors explaining that correlation.
However i have no idea how to extract correlation series to build my dependent variable(my data is from 1980 to 2017). Shall i go for rolling window correlation or is there any other alternative to create correlation series that will serve as dependent variable?

I would be thankful for your suggestions.
my email is raqia123@gmail.com


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 25 guests