VAR & Degrees of Freedom

[meh]

I'm trying to put some tables together for a VAR estimation, but have a quick question on the t-statistics and degrees of freedom. It doesn't provide a probability level, so I am using a routine probability calculator. My issue is how to calculate the degrees of freedom in a VAR. I have 104 observations for 7 endogenous variable estimations. Including the endogenous, exogenous and constant parameters I estimate 20 different parameters for each of the 7 endogenous estimations. I assume that the df would be 84, but can't find any confirmation of this anywhere. Is this correct?