White test with and without cross terms

[titzaaa]

Hi everyone,
when estimating my VAR, the White test with and without cross terms give contradictory results.
Here are my tables:

Including Cross terms:
VAR Residual Heteroskedasticity Tests (Includes Cross Terms)
Date: 09/19/19 Time: 11:55
Sample: 2006M03 2019M03
Included observations: 144


Joint test:

Chi-sq df Prob.

426.3940 405 0.2230


Individual components:

Dependent R-squared F(135,8) Prob. Chi-sq(135) Prob.

res1*res1 0.987047 4.515857 0.0134 142.1348 0.3202
res2*res2 0.981139 3.082676 0.0445 141.2841 0.3383
res2*res1 0.994861 11.47309 0.0005 143.2601 0.2970


No cross terms:
VAR Residual Heteroskedasticity Tests (Levels and Squares)
Date: 09/19/19 Time: 12:03
Sample: 2006M03 2019M03
Included observations: 144


Joint test:

Chi-sq df Prob.

199.1152 90 0.0000


Individual components:

Dependent R-squared F(30,113) Prob. Chi-sq(30) Prob.

res1*res1 0.485044 3.547876 0.0000 69.84636 0.0001
res2*res2 0.558858 4.771787 0.0000 80.47562 0.0000
res2*res1 0.377138 2.280689 0.0010 54.30791 0.0042

Can anybody help with interpretation or how to proceed? Can I assume I have no heteroskedasticity and continue?

Best,
Franziska

[titzaaa]

Also, when exchanging some variables, I get the error message with cross terms saying: "Positive or non-negative argument to function expected". Can somebody tell me what that means?

Thanks a lot!