DCCGARCH11

[soo]

Hello,

I have the version Eviews student 9, so I can't access to this add-in.
I would like to do a DCC-GARCH (1,1) model, can you provide me the code please ?

Thanks a lot,

[EViews Gareth]

If you are using the student version you cannot use code...

[soo]

Ok, so, how can I model a DCC GARCH (1,1) in the student version (Version 9)?

I have a place named "command", I was thinking it was for a code ...


I am a really new user of EViews !

Thanks a lot for your help !

[EViews Gareth]

You cannot

[ferryk21]

Hi everyone,

First, I want to say thank you to Tubrador that made this add in. This add in is awesome. However, I want to ask something about this add in. I want to use lag term in my model.But when I input the lag time into the model, the dcc parameter become NA and everything become insignificant. What's happened? Anyone can help me please?

Thank you

[trubador]

Please make sure you searched the forum or at least read the previous queries before posting your questions:

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.

[soo]

Dear mister/madam,

I would like to apply a DCC-GARCH model to find the correlation between my stock and bond returns.

Since the univariate GARCH model should be applied on the residuals of these two series, I am wondering if I have to put the variable "stock returns" and "bond returns" in the first box or their residuals ? (so "residuals stock returns" and "residuals bond returns")?

Thanks a lot for your answer,

Have a nice day !

[trubador]

Since the univariate GARCH model should be applied on the residuals of these two series, I am wondering if I have to put the variable "stock returns" and "bond returns" in the first box or their residuals ? (so "residuals stock returns" and "residuals bond returns")?

The add-in lets you build an ARIMA model for the mean part, so you can simply use actual returns as dependent variables. If you like, you can build models for the mean part at the outset and then feed residuals into the add-in: http://forums.eviews.com/viewtopic.php? ... =75#p44556

[anzh]

Dear trubator

I would like to ask about whether there are diagnostic tests on the DCC-GARCH add-in in Eviews? This is because i couldn't find these when using the add-in. Can you help me please?

Thank you!

[debbysoraya]

Hey,

So right now I am estimating bivariate dcc garch (return in 2 indices such as s&p500 and klse).
Some of the paper that I read,
define the mean equation such: r_t=γ_0 + γ_1 r_(t-1)+ γ_2 r_(t-1)^us + ε_t ( I attached the paper's method and result) , it uses ar(1) process in the mean equation

My question is, every time I put Rus and Rmy in the return series and put 1 in the ar process for mean equation box, the rho that is produced is not series instead it's only give me one number (for the first day) whereas i supposed the rho_12_01 supposed to give series right?


I wonder, if I put only Rus and Rmy ( I got the series of rho_12_01) without including ar(1) and later I modify the estimation that is saved under eq_garch_r2_01 (meaning after the dcc then i go back to the univariate garch estimation and change the command there and include the ar(1))? is this violating the dcc proposed by engle somehow?


THANKS A LOT!!!

[trubador]

Hey,

So right now I am estimating bivariate dcc garch (return in 2 indices such as s&p500 and klse).
Some of the paper that I read,
define the mean equation such: r_t=γ_0 + γ_1 r_(t-1)+ γ_2 r_(t-1)^us + ε_t ( I attached the paper's method and result) , it uses ar(1) process in the mean equation

My question is, every time I put Rus and Rmy in the return series and put 1 in the ar process for mean equation box, the rho that is produced is not series instead it's only give me one number (for the first day) whereas i supposed the rho_12_01 supposed to give series right?


I wonder, if I put only Rus and Rmy ( I got the series of rho_12_01) without including ar(1) and later I modify the estimation that is saved under eq_garch_r2_01 (meaning after the dcc then i go back to the univariate garch estimation and change the command there and include the ar(1))? is this violating the dcc proposed by engle somehow?


THANKS A LOT!!!

For the technical difficulties you experience during the estimation, please see previous posts:

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.

And note that the estimated model in the paper you cited uses the return of US index as an exogenous variable (i.e. the second box in the add-in).

[Mr. Rags]

Dear trubador,
I am new to this add-in. So I used S&P500 index and a China stock index as my practice sets. DCC GARCH worked except I don't have any value for conditional correlations. Would you help me figure out why?
Thank you,
Mr. Rags

[Mr. Rags]

Is there still a compatibility issue with eViews 9? I keep getting this result "Optimization failed: unable to evaluate objective at optimization starting values." and my rhos are all "NA."
Thanks.

[ecko33]

I have a question about the results from add-in. I'm curiuos that the results in sheet rho are dynamic correlation or conditional correlation?

[saadallah]

Hello,
I am using Eviews 9 and I keep getting this result "Optimization failed: unable to evaluate objective at optimization starting values." and my rhos are all "NA."

Could you please help?