[R-SIG-Finance] External regressor bounds in rmgarch

[Josh Segal]

Hi all,

Is it possible to change the bounds for external regressor coefficients
when estimating a *multivariate* volatility model (e.g. dccspec) in rmgarch?

I know how to do this for a univariate spec in rugarch using
'setbounds<-'.  However, when I define a multivariate spec from such
univariate specs, the revised bounds seem to be overridden.

I'm doing something like this:

uspec = ugarchspec(variance.model = list(external.regressors =
external.regressors))
setbounds(uspec)<-list(vxreg1=c(-1,1))
spec = dccspec(uspec = multispec(replicate(n.dim, uspec)))
fit = dccfit(spec, data = data)

I care because I want to allow for the possibility of negative coefficients.

Thanks for any help!

	[[alternative HTML version deleted]]