1
$\begingroup$

I estimated a VAR model and used the Granger-causality test, to interpret the model. I used the causality function to check for possible Granger-causality, but I get conflicting results depending on using boot or not. I am confused. Which result should I take for interpretation? Variables are stationary and the residuals are all fine: no autocorrelation and homoskedasty. Number of observations used for the VAR model = 65.

causality(Var1, cause = "DBIP1ts", boot = FALSE)

--> H0 is rejected; p-value is 0.03

causality(Var1, cause = "DBIP1ts", boot = TRUE, boot.runs = 1000)

--> H0 is not rejected; p-value is 0.14

I am confused. Which result should I take for interpretation?

Improve this question
$\endgroup$

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.