Estimating EGARCH model

[Rosa.hh]

Hi,
I'm new in financial econometrics and I am estimating an EGARCH model of the change of an interest rate (in the interbank market) using dummy variables and a set of control variables.

When running my model I found out that one of my control variables (liquidity imbalances) in the Mean equation was turning the variance equation to have statistically insignificant parameters and p-values=1 for all my dummy variables (in the variance equation). However, after excluding this problematic control variable, the equation gives results more or less similar to those present in the literature review, without having all insignificant variables in the Variance equation.

Anyone knows why does this happen? I would like to be clear of the reason of having all my parameters in the variance equation with insignificant values and dummy vars (in the variance equation) equal to 1 as a cause of one of my control variables.

As well, what's usually done in these cases when the control variable cannot be included...instrumental variables?

I enclose the workfile with equation03 (including problematic control variable, liq_imb) and equation0302 (excluding problematic control variable).

Thanks a lot!