2
$\begingroup$

I want to estimate the dynamic conditional correlation GARCH in stata 12. I just ran the following line:

mgarch dcc (x y =, noconstant) , arch(1) garch(1)

The x and the y variables are log first differences from two price series.

My questions are as follows:

  1. Does anybody know how I obtain the parameter estimates for the elements of Ht?
  2. Is it correct that I directly put the log first differences in this model or should I first obtain the residuals from a AR model?
$\endgroup$
1
  • $\begingroup$ Actually you need to prewhiten the series before running the model. Usually this is done by fitting a VAR model and applying the DCC GARCH model to the residuals of the VAR, or estimating both models in one-step. I'm not sure how to do it in stata, but you can do this one-step estimation with the rmgarch package in R. $\endgroup$ Feb 2, 2016 at 11:41

1 Answer 1

1
$\begingroup$
  1. Type predict H*,variance
  2. I think log first differences or other types depends on your purpose of setting model, not relevant with Dcc model.
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.