Including whitend Standard Error in VAR Model

[Alaska21]

Hi,

Because I had heteroskedasticity and non-normal distribution of the residuals I whitend the residuals and included them in my estimation as exogenous Variables, next to the constant and dummys. I now yield good results in the normality and heteroskedasticity tests (2 lags are autocorrelated according to LM test) and my R^2 is 0.97. This seems to good to be true. Was is wrong to include the whitend residuals in the Exogenous Variable Box?

Thanks!