Asymmetric Dynamic Conditional Correlation in Matlab
2 vues (au cours des 30 derniers jours)
Afficher commentaires plus anciens
I need to interpret the results of the estimation of the A-DCC model by Cappiello, Engle and Sheppard, 2006, Journal of Financial Econometrics. I used the Matlab routines of the MFE Toolboox http://www.kevinsheppard.com/MFE_Toolbox in this form:
parameters = dcc(data,[],1,1,1,[],[],[],[],'2-stage')
where data is a 100x2 matrix, hence the estimation is performed on two variables observed 100 times. The output "parameters" is a vector of 10 components. Unfortunately I am not able to derive a precise expression of the model and hence to interpret the parameters. The help of the routine says:
OUTPUTS:
PARAMETERS - Estimated parameters. Output depends on METHOD.
3-stage: [VOL(1) ... VOL(K) corr_vech(R)' vech(N)' alpha gamma beta]
2-stage: [VOL(1) ... VOL(K) corr_vech(R)' alpha gamma beta]
where VOL(j) is a (1+P(i)+O(i)+Q(i)) vector containing the parameters from volatility model i.
Which is the volatility model i? Unfortunately the documentation on the A-DCC model in the MFE toolbox page is missing.
Something more is said in the end of the help:
2-stage:
Q(t) = R.*scale + a(1)*e(t-1)'*e(t-1) + ... + a(m)*e(t-m)'*e(t-m)
+ g(1)*v(t-1)'*v(t-1) + ... + g(l)*v(t-l)*v(t-l) + b(1)*Q(t-1) + ... + b(n)*Q(t-1)
where v(t,:) = e(t,;).*(e(t,:)<0) and s = sqrt((1-sum(a)-sum(b)-gScale*sum(g))) and scale = s*s'
Q(t) should be a 2x2 matrix and hence a(1), b(1) and g(1) should be 2x2 matrices. My problem is to find how the element of the three matrices are linked to the entries in the vector PARAMETERS.
2 commentaires
Lorenzo Orlando
le 14 Juin 2017
did you solve the problem? I have the same problem here, but I only got 9 parameters
Obaidur Rehman
le 10 Sep 2017
Hi Lorenzo,
Did you manage to figure it out? I think I have managed to figure out on how to interpret the results, however I am struggling to follow the difference between the 2-stage estimation versus 3-stage estimation.
Réponses (0)
Voir également
Catégories
En savoir plus sur Preprocessing Data dans Help Center et File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!