[R-SIG-Finance] RUGARCH - non-negativity constraints on coefficients for exogenous variables in GARCH?

[Christian Borelli-Kjær]

Hi all,

I am using the RUGARCH package to fit a GARCH model with three exogenous
variables, but I can only get positive, insignificant parameter estimates
for the exogenous variables.

This troubles me, as I have a strong hypothesis that the exogenous
variables should somehow be "volatility-lowering" - and thus affect the
conditional volatility in a negative way.

Does the package constrain coefficients for exogenous variables to be ≥ 0,
just as alpha and beta, or does it relax this assumption, corresponding to
e.g. He & Teräsvirta (1997) ?

Please do request code and data, if needed.

Thank you for your time,

Christian Borelli, University of Copenhagen

	[[alternative HTML version deleted]]