1
$\begingroup$

I am estimating a var model with four variables: GDP, investemnt, inflation and unemployment. (all in growth rates). Now I estimated it once with a constant and one without. The model is not changing much but when I calculate the irf functions of the gdp after a shock in investment it is only significant when I dont add a constant to the model. Now my question is if I should include a constant in the model and live with the insignificant results or if I should continue with the model without constant.

Can someone help me ? :)

enter image description here

Improve this question
$\endgroup$
2
  • $\begingroup$ Looks like you have growth rates or first differences here. Accordingly, adding a constant may imply a linear trend in logarithms or levels. You may want to think whether that is reasonable. $\endgroup$
    – Richard Hardy
    Commented Jun 22, 2021 at 18:02
  • $\begingroup$ Yes sorry I didnt specify enough. I have all variables in growth rates and yes in levels I have in nearly all variables a linear trend $\endgroup$
    – Joe94
    Commented Jun 22, 2021 at 18:15

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.