BIVARIATE GARCH BEKK

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

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

Re: BIVARIATE GARCH BEKK

Postby trubador » Thu Sep 03, 2015 1:01 pm

It works just fine. Could please check the build date of your EViews? (go to Help->About EViews)

nenolusi
Posts: 18
Joined: Wed Jun 17, 2015 1:10 am

Re: BIVARIATE GARCH BEKK

Postby nenolusi » Thu Sep 03, 2015 6:22 pm

Dear Trubador,

Thank you for your patience and apologies for the late reply. The office closed and I did not have access to Eviews.

The build date is April 27 2015.

nenolusi
Posts: 18
Joined: Wed Jun 17, 2015 1:10 am

Re: BIVARIATE GARCH BEKK

Postby nenolusi » Thu Sep 03, 2015 10:19 pm

Dear Trubador,

Thanks for your patience and assistance. I have resolved it.

nenolusi
Posts: 18
Joined: Wed Jun 17, 2015 1:10 am

Re: BIVARIATE GARCH BEKK

Postby nenolusi » Thu Sep 03, 2015 11:10 pm

Dr Trubador,

I am having an issue with the estimation of the 10 minutes interval in the second sample period.
The error message is "Attempt to raise a negative number to a non integer power -
Missing data generated in "SERIES STRES1 = RES1 /(VAR_R1 ^ .5)"

Please how can I resolve this?

Thank you.

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

Re: BIVARIATE GARCH BEKK

Postby trubador » Fri Sep 04, 2015 12:09 am

Estimated coefficients yield a negative value for the variance at some point. Change the starting values and re-estimate the model until you have a statistically valid and feasible output.

nenolusi
Posts: 18
Joined: Wed Jun 17, 2015 1:10 am

Re: BIVARIATE GARCH BEKK

Postby nenolusi » Fri Sep 04, 2015 1:15 am

Thanks. It worked.

I am so grateful.

georgiana091296
Posts: 1
Joined: Sat Mar 10, 2018 12:19 pm

Re: BIVARIATE GARCH BEKK

Postby georgiana091296 » Mon Mar 12, 2018 2:21 am

Hello. I tried to estimate a bivariate bekk model using this code but i can't understand how to interpret the result. I need to know which is diagonal and off diagonal elements of matrix alpha and matrix beta. The result show me only 3 elements and i don't know about the fourth one. I thought maybe this matrices are restricted lower or upper triangular but the methodology says that should be unrestricted. Thanks a lot for the code and for your help.

terrytrausith
Posts: 2
Joined: Tue Dec 18, 2018 12:05 am

Re: BIVARIATE GARCH BEKK

Postby terrytrausith » Wed Dec 19, 2018 1:02 am

Hi there,
I am a beginner to Econometrics. I use the following BEKK specification for my research work. I need some help on interpreting the estimation results.
Thanks in advance,
Fist I selected OBJECT – NEW OBJECT – SYSTEM and I typed
Ashare = C(1) + C(2)*Ashare(-1)
Bshare = C(3) + C(4)*Bshare(-1)
Then I selected – ESTIMATE
Estimation method – ARCH-Conditional Hetroscadsticity
Model Type – Diagonal BEKK
Coefficient – GARCH(1)
Error Distribution- Multivariate Student t
I got the following results

System: BKK
Estimation Method: ARCH Maximum Likelihood (Marquardt)
Covariance specification: Diagonal BEKK
Date: 12/18/18 Time: 15:29
Sample: 1/02/1991 4/05/1993
Included observations: 825
Total system (balanced) observations 1650
Disturbance assumption: Student's t distribution
Presample covariance: backcast (parameter =0.7)
Convergence achieved after 38 iterations


Coefficient Std. Error z-Statistic Prob.


C(1) 0.001380 0.000432 3.197118 0.0014
C(2) -0.210809 0.028350 -7.435964 0.0000
C(3) 0.000837 0.000367 2.282075 0.0225
C(4) -0.384797 0.036372 -10.57960 0.0000


Variance Equation Coefficients


C(5) 3.32E-06 1.18E-06 2.800500 0.0051
C(6) 1.88E-06 6.48E-07 2.903840 0.0037
C(7) 2.01E-06 7.61E-07 2.635640 0.0084
C(8) 0.261430 0.025739 10.15679 0.0000
C(9) 0.281633 0.030395 9.265771 0.0000
C(10) 0.964253 0.006360 151.6200 0.0000
C(11) 0.961891 0.007597 126.6087 0.0000


t-Distribution (Degree of Freedom)


C(12) 5.025472 0.356657 14.09049 0.0000


Log likelihood 5023.331 Schwarz criterion -12.08009
Avg. log likelihood 3.044443 Hannan-Quinn criter. -12.12237
Akaike info criterion -12.14868



Equation: A = C(1) + C(2)*A(-1)
R-squared 0.011836 Mean dependent var -0.001318
Adjusted R-squared 0.010635 S.D. dependent var 0.039861
S.E. of regression 0.039649 Sum squared resid 1.293772
Durbin-Watson stat 1.272177

Equation: B = C(3) + C(4)*B(-1)
R-squared -0.033247 Mean dependent var -0.001402
Adjusted R-squared -0.034502 S.D. dependent var 0.037716
S.E. of regression 0.038361 Sum squared resid 1.211088
Durbin-Watson stat 1.065445



Covariance specification: Diagonal BEKK
GARCH = M + A1*RESID(-1)*RESID(-1)'*A1 + B1*GARCH(-1)*B1
M is an indefinite matrix
A1 is a diagonal matrix
B1 is a diagonal matrix


Transformed Variance Coefficients


Coefficient Std. Error z-Statistic Prob.


M(1,1) 3.32E-06 1.18E-06 2.800500 0.0051
M(1,2) 1.88E-06 6.48E-07 2.903840 0.0037
M(2,2) 2.01E-06 7.61E-07 2.635640 0.0084
A1(1,1) 0.261430 0.025739 10.15679 0.0000
A1(2,2) 0.281633 0.030395 9.265771 0.0000
B1(1,1) 0.964253 0.006360 151.6200 0.0000
B1(2,2) 0.961891 0.007597 126.6087 0.0000



What do the coefficients indicate and what coefficients indicate mean spillovers and volatility spillovers between A-share and B-Shares. What do negative coefficients indicate?


Thanks in advance

terrytrausith
Posts: 2
Joined: Tue Dec 18, 2018 12:05 am

Re: BIVARIATE GARCH BEKK

Postby terrytrausith » Fri Dec 21, 2018 8:03 am

Hi

Should the sum of ARCH and GARCH coefficients of Diagonal BEKK (GARCH(1)) be less than unity ?

Thanks

Yang Li
Posts: 1
Joined: Fri Jan 04, 2019 1:22 pm

Re: BIVARIATE GARCH BEKK

Postby Yang Li » Fri Jan 04, 2019 6:18 pm

Hi there,
I have the same question as you. Do you figure it out now? If yes, can you please tell me how to interpret estimation results?Thanks a lot!!!


terrytrausith wrote:Hi there,
I am a beginner to Econometrics. I use the following BEKK specification for my research work. I need some help on interpreting the estimation results.
Thanks in advance,
Fist I selected OBJECT – NEW OBJECT – SYSTEM and I typed
Ashare = C(1) + C(2)*Ashare(-1)
Bshare = C(3) + C(4)*Bshare(-1)
Then I selected – ESTIMATE
Estimation method – ARCH-Conditional Hetroscadsticity
Model Type – Diagonal BEKK
Coefficient – GARCH(1)
Error Distribution- Multivariate Student t
I got the following results

System: BKK
Estimation Method: ARCH Maximum Likelihood (Marquardt)
Covariance specification: Diagonal BEKK
Date: 12/18/18 Time: 15:29
Sample: 1/02/1991 4/05/1993
Included observations: 825
Total system (balanced) observations 1650
Disturbance assumption: Student's t distribution
Presample covariance: backcast (parameter =0.7)
Convergence achieved after 38 iterations


Coefficient Std. Error z-Statistic Prob.


C(1) 0.001380 0.000432 3.197118 0.0014
C(2) -0.210809 0.028350 -7.435964 0.0000
C(3) 0.000837 0.000367 2.282075 0.0225
C(4) -0.384797 0.036372 -10.57960 0.0000


Variance Equation Coefficients


C(5) 3.32E-06 1.18E-06 2.800500 0.0051
C(6) 1.88E-06 6.48E-07 2.903840 0.0037
C(7) 2.01E-06 7.61E-07 2.635640 0.0084
C(8) 0.261430 0.025739 10.15679 0.0000
C(9) 0.281633 0.030395 9.265771 0.0000
C(10) 0.964253 0.006360 151.6200 0.0000
C(11) 0.961891 0.007597 126.6087 0.0000


t-Distribution (Degree of Freedom)


C(12) 5.025472 0.356657 14.09049 0.0000


Log likelihood 5023.331 Schwarz criterion -12.08009
Avg. log likelihood 3.044443 Hannan-Quinn criter. -12.12237
Akaike info criterion -12.14868



Equation: A = C(1) + C(2)*A(-1)
R-squared 0.011836 Mean dependent var -0.001318
Adjusted R-squared 0.010635 S.D. dependent var 0.039861
S.E. of regression 0.039649 Sum squared resid 1.293772
Durbin-Watson stat 1.272177

Equation: B = C(3) + C(4)*B(-1)
R-squared -0.033247 Mean dependent var -0.001402
Adjusted R-squared -0.034502 S.D. dependent var 0.037716
S.E. of regression 0.038361 Sum squared resid 1.211088
Durbin-Watson stat 1.065445



Covariance specification: Diagonal BEKK
GARCH = M + A1*RESID(-1)*RESID(-1)'*A1 + B1*GARCH(-1)*B1
M is an indefinite matrix
A1 is a diagonal matrix
B1 is a diagonal matrix


Transformed Variance Coefficients


Coefficient Std. Error z-Statistic Prob.


M(1,1) 3.32E-06 1.18E-06 2.800500 0.0051
M(1,2) 1.88E-06 6.48E-07 2.903840 0.0037
M(2,2) 2.01E-06 7.61E-07 2.635640 0.0084
A1(1,1) 0.261430 0.025739 10.15679 0.0000
A1(2,2) 0.281633 0.030395 9.265771 0.0000
B1(1,1) 0.964253 0.006360 151.6200 0.0000
B1(2,2) 0.961891 0.007597 126.6087 0.0000



What do the coefficients indicate and what coefficients indicate mean spillovers and volatility spillovers between A-share and B-Shares. What do negative coefficients indicate?


Thanks in advance


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 7 guests