I just ran a Vector auto-regressive model using 6 endogenous variables. i have attached an image of the result. Can some one tell me what should i be wary of before interpreting the results? i know we should see if there are any auto correlation (what test should i use?). can you guys tell me more robustness checks. Also does it make sense making casual interpretation from coefficients or is it better to run the impulse response function and interpret them instead .
should the time series variables be differences before running the VAR model? (the literature seems to be in splits about this issue, what do practitioners do? pls advice)
i am currently using e-views.