[R-SIG-Finance] rugarch vs fgarch

alexios ghalanos alexios at 4dscape.com
Thu Jul 19 15:40:26 CEST 2012

You can do this with rugarch (and possibly fGarch) by passing the 
external regressors (factors) in the conditional mean equation. The 
documentation and vignette should provide enough information to enable 
you to figure out how to do this. You can also adopt a 2-stage approach 
in which case any GARCH package can do this (by passing the residuals 
from a first stage regression).


On 19/07/2012 12:57, Abhishek Gupta wrote:
> rugarch is a package that allows you to do Univariate GARCH.
> What package should I be using if I want to estimated a multi-factor
> univariate GARCH model.
> For example:
> sp_ret = alpha_0 + alpha1*(market_index_ret) +alpha2*(exchange_rate_return)
> + error
> sigma^2 = gamma_0 + gamm1*(error^2_{t-1}) + gamma2*(sigma^2_{t-1})
> Thanks.

More information about the R-SIG-Finance mailing list