DCCGARCH11

For questions about EViews Add-ins available from the EViews Add-ins webpage. Note each add-in available on our webpage will have its own individual thread.

Moderators: EViews Gareth, EViews Moderator, EViews Esther

akash27
Posts: 7
Joined: Wed Jun 18, 2014 9:00 am

Re: DCCGARCH11

Postby akash27 » Thu Jun 19, 2014 11:41 am

Dear turbador,

Please accept my sincere apologies for the e-mail.

I am trying to plot dynamic conditional correlation between oil prices & inflation rate. In your opinion, should I filter the time series using VAR and then input the residuals in add-in or should I directly give the two time series as input to add-in? because, when I gave residuals as input the add-in gave ridiculous results

from VAR residuals
Image

without residuals
Image

I have read the add-in documentation many times, but I'm still unable to figure out what does theta (1) and theta (2) signify?

Thanks.

akash27
Posts: 7
Joined: Wed Jun 18, 2014 9:00 am

Re: DCCGARCH11

Postby akash27 » Thu Jun 19, 2014 11:49 am

I'm new to DCC-Garch model as I'm from engineering background. I have another curiosity which is not directly related to this add-in, but I would be really grateful if you could help me with those.
I am trying to find dependence of Inflation rate on monthly Oil prices, for that
1) Should I use time series for monthly oil prices or monthly returns?
2) Which time series should I provide as exogenous variables in mean & variance equations? inflation or oil?
3) How should I decide number of AR lags to be used in the model?

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

Re: DCCGARCH11

Postby trubador » Fri Jun 20, 2014 2:30 am

Add-ins are simply the tools for applying a given method, which is not built-in in EViews. In any case, you are assumed to have some background on the method before applying it. The documentation covers only the issues directly related to add-in, not the method itself. You can, however, always refer to articles cited in the reference section in order to get a better understanding of the method.

The relationship between oil prices and inflation is a widely studied topic. You can find plenty of papers on the subject that will provide you enough guidance on the methodology. I also suggest you to work on the concept of stationarity as well as simple ARMA modeling before going any further.

fa3oor89@gmail.com
Posts: 1
Joined: Thu Jun 26, 2014 7:09 am

Re: DCCGARCH11

Postby fa3oor89@gmail.com » Thu Jun 26, 2014 7:16 am

Hi,

I'm currently using the DCC add-in, which seems to work fine. I'm using it for the purpose of detecting volatility spillover effects. However, I didn't manage to find the coefficients pertaining to the direction of the spillover and all I managed to find was the thetas and the correlations. Isn't this facility provided in the add-in?

Thanks

econworker
Posts: 39
Joined: Thu Apr 24, 2014 3:51 am

Re: DCCGARCH11

Postby econworker » Mon Jun 30, 2014 6:50 am

econworker wrote:
trubador wrote:PDF document should be stored here: C:\Users\...\Documents\EViews Addins\dccgarch11\dccgarch11.pdf


Thank you very much


Hi Trabadur, it seems that its not possible to do diagnostic tests withing the add-in of DCC-GARCH, is it correct?

Tim Malone
Posts: 1
Joined: Tue Jul 01, 2014 4:41 am

Re: DCCGARCH11

Postby Tim Malone » Tue Jul 01, 2014 4:48 am

Hi

I need to find correlation between two time series. Do I insert the time series directly or I have perform some operation on them before insertion ?

Thanks

econworker
Posts: 39
Joined: Thu Apr 24, 2014 3:51 am

Re: DCCGARCH11

Postby econworker » Wed Jul 02, 2014 12:35 am

Tim Malone wrote:Hi

I need to find correlation between two time series. Do I insert the time series directly or I have perform some operation on them before insertion ?

Thanks


Hi, DCC-GARCH is a two step process, in the first step you ahould perform the univariate GARCH and obtain the residuals, cov and etc...from them and then perform the DCC-GARCH with using those results from the univariate GARCH, but this add-in does this two steps process and you dont need to do anything before, as I know.

econworker
Posts: 39
Joined: Thu Apr 24, 2014 3:51 am

Re: DCCGARCH11

Postby econworker » Mon Jul 14, 2014 2:49 am

Dear Trabadur
I have some problems with using DCC-GARCH add-in I hope you help me about it.
What I understand is that this add in is not working properly, when its a simple model without any exogenous shock to the mean or variance equations of its GARCH models it works, but when I include some variable to mean equation or if I include AR term then it can not converge. although the maximum number of iterations is 500 but it does not go more than 10 or 15 iterations and consequently it doesn't converge. Do you think there is a bug in the add in or is there any solution for that?
its good to mention that when I estimate the same model using the same data with DCC GARCH code, it works properly.
I would be very thankful if you help me about it.
Thanks

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

Re: DCCGARCH11

Postby trubador » Tue Jul 15, 2014 12:39 pm

fa3oor89@gmail.com wrote:Hi,

I'm currently using the DCC add-in, which seems to work fine. I'm using it for the purpose of detecting volatility spillover effects. However, I didn't manage to find the coefficients pertaining to the direction of the spillover and all I managed to find was the thetas and the correlations. Isn't this facility provided in the add-in?

Thanks


DCC GARCH models work on the correlation structure. You'll need unparameterized versions of multivariate GARCH models (e.g. unrestricted BEKK).

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

Re: DCCGARCH11

Postby trubador » Tue Jul 15, 2014 12:41 pm

Tim Malone wrote:Hi

I need to find correlation between two time series. Do I insert the time series directly or I have perform some operation on them before insertion ?

Thanks


DCC GARCH models represent the dynamic conditional covariances of the standardized residuals, and therefore do not actually yield dynamic conditional correlations.

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

Re: DCCGARCH11

Postby trubador » Tue Jul 15, 2014 1:02 pm

econworker wrote:Dear Trabadur
I have some problems with using DCC-GARCH add-in I hope you help me about it.
What I understand is that this add in is not working properly, when its a simple model without any exogenous shock to the mean or variance equations of its GARCH models it works, but when I include some variable to mean equation or if I include AR term then it can not converge. although the maximum number of iterations is 500 but it does not go more than 10 or 15 iterations and consequently it doesn't converge. Do you think there is a bug in the add in or is there any solution for that?
its good to mention that when I estimate the same model using the same data with DCC GARCH code, it works properly.
I would be very thankful if you help me about it.
Thanks


As usual, it is difficult to pinpoint the source of problem (if any) without seeing actual data/workfile. But;

1) Since this is a two-step model, you should check the first step (i.e. estimation of univariate GARCH models) to see if everything is OK. Try alternative GARCH models.
2) Model may become ill-defined or inconsistent after including an exogenous variable or an AR term. Try dropping them or find better RHS variables.
3) Starting values of coefficients may be too far from an optimal solution. Try different initial parameter values (i.e. theta vector).
4) Correlation targeting may be too restrictive. Try unchecking this option.
5) Sample period may not be appropriate to carry out such an analysis. Try adjusting the sample.
6) Optimization algorithm may perform poorly. Try other alternatives.
7) Algorithms may fall out the domain of estimation parameters. Try optimizing the squared coefficients.

Such models are nonlinear in nature and therefore there is no guarantee that they will always converge and yield proper estimation results. It really needs "your" time and effort to get it work.

bvguizar
Posts: 40
Joined: Sat Aug 28, 2010 6:56 am

Re: DCCGARCH11

Postby bvguizar » Fri Dec 19, 2014 5:08 am

I can't find the documentation on the

c:/... or the manage add in. Can someone help me please?

Thank you

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

Re: DCCGARCH11

Postby trubador » Fri Dec 19, 2014 6:36 am

bvguizar wrote:I can't find the documentation

Look into the add-ins folder: C:\Users\...\Documents\EViews Addins\dccgarch11

dnguyen1
Posts: 3
Joined: Mon Dec 29, 2014 2:35 pm

Re: DCCGARCH11

Postby dnguyen1 » Mon Dec 29, 2014 2:56 pm

Hi,

I'm using dccgarch11 add-in for my research. I just got some negative values for dcc. Could you please help me explain why the results for dcc could be negative? Or did I do anything wrong when using the add-in?
The eviews file attached below shows an example. The dcc between china and eu (us also) has some negative value.

Thank you very much!
Attachments
weekly sectoral indices.wf1
(227.42 KiB) Downloaded 835 times

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

Re: DCCGARCH11

Postby trubador » Tue Dec 30, 2014 1:16 am

Your results seem fine. If you are referring to rho_* outputs, then it is normal to observe some negative values as they represent the correlations not the variances.


Return to “Add-in Support”

Who is online

Users browsing this forum: No registered users and 26 guests