Fitting a DCC-GARCH with more than one external regressor per single process using rmgarch
If I understand correctly, the DCC estimation fails when including 2 external regressors per univariate variance equation (but not 1), and this does not happen when estimating them separately? A quick solution is to supply the estimated "multifit" object to the DCC estimation (via the 'fit' option). ##################################################################### args(dccfit) dccfit(spec, data, out.sample = 0, solver = "solnp", solver.control = list(), fit.control = list(eval.se = TRUE, stationarity = TRUE, scale = FALSE), cluster = NULL, fit = NULL, VAR.fit = NULL, ...) #################################################################### i.e. pass the multispec object to the multifit method and then the resulting object to dccfit. The dccfit does the same thing, but you have more control over the estimation and fine-tuning of solver/estimation setup in the multifit case. Otherwise convergence problems, particularly in the presence of external regressors in the variance equation, are discussed in the FAQs of the vignette and elsewhere in the archives of this mailing list. You may also find the following paper of interest: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.158.1593&rep=rep1&type=pdf) -Alexios
On 03/09/2013 16:52, leopoldo.catania wrote:
Hi,
I'm trying to fit a dcc model using the rmgarch package, but when I put in
the variance equation more than one external regressor the dccfit fails.
Estimating one by one the single process (is an eGARCH(1,1)) there are no
problems using more than one external regressor.
I tryied also to use the "fit" parameter in the dccfit() command, but
nothing change, the error still remain the same.
Estimating the DCC insering only one external regressor in each single
process works great.
the dccfit() command reports:
Error in `rownames<-`(`*tmp*`, value = c("mu", "ar1", "ma", "arfima", :
length of 'dimnames' [1] not equal to array extent
Best regards,
Leopoldo Catania.
Below the complete dcc garch specification
spec
[[1]] *---------------------------------* * GARCH Model Spec * *---------------------------------* Conditional Variance Dynamics ------------------------------------ GARCH Model : eGARCH(1,1) Variance Targeting : FALSE Exogenous Regressor Dimension: 2 Conditional Mean Dynamics ------------------------------------ Mean Model : ARFIMA(1,0,0) Include Mean : FALSE GARCH-in-Mean : FALSE Conditional Distribution ------------------------------------ Distribution : std Includes Skew : FALSE Includes Shape : TRUE Includes Lambda : FALSE [[2]] *---------------------------------* * GARCH Model Spec * *---------------------------------* Conditional Variance Dynamics ------------------------------------ GARCH Model : eGARCH(1,1) Variance Targeting : FALSE Exogenous Regressor Dimension: 2 Conditional Mean Dynamics ------------------------------------ Mean Model : ARFIMA(1,0,0) Include Mean : FALSE GARCH-in-Mean : FALSE Conditional Distribution ------------------------------------ Distribution : std Includes Skew : FALSE Includes Shape : TRUE Includes Lambda : FALSE
mspec = multispec( spec ) mspec
*-----------------------------* * GARCH Multi-Spec * *-----------------------------* Multiple Specifications : 2 Multi-Spec Type : equal
multispec=dccspec(mspec,model="DCC") multispec
*------------------------------* * DCC GARCH Spec * *------------------------------* Model : DCC(1,1) Estimation : 2-step Distribution : mvnorm No. Parameters : 19 No. Series : 2
a=as.xts(industry.tf.list[[tf]][,1:2]) dccfit.industry.tf[[q]]=dccfit(multispec,a)
Error in `rownames<-`(`*tmp*`, value = c("mu", "ar1", "ma", "arfima", :
length of 'dimnames' [1] not equal to array extent
--
View this message in context: http://r.789695.n4.nabble.com/Fitting-a-DCC-GARCH-with-more-than-one-external-regressor-per-single-process-using-rmgarch-tp4675275.html
Sent from the Rmetrics mailing list archive at Nabble.com.
_______________________________________________ R-SIG-Finance at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-sig-finance -- Subscriber-posting only. If you want to post, subscribe first. -- Also note that this is not the r-help list where general R questions should go.