Testing GARCH specification

[vass]

Hello,

Could anyone let me know when estimating GARCH using t-distribution errors or GED whether the Log likelihood function reported is the exact likelihood function for such a distribution, i.e. all scalar quantities are included and the correct distribution is used (no quasi log likelihood fn)? Thanks. (just fyi, I need this as I am to perform log likelihood tests to check for order of GARCH model and the presence of asymmetric effects).