I'm performing an econometric analysis and all my variables (in daily frequency) are stationary, so I've decided to use a VAR model. After analyzing the data in Stata, it suggests all variables should have a lag of 1 day except one (let's call it Z), which should have a lag of 4 days.

When constructing the model, should I:

  1. Lag every variable up to 4 lags?
  2. Lag every variable once and variable Z until lag 4?



1 Answer 1


There is no requirement for the different variables in a VAR model to have the same number of lags. I am not sure how you decided that 1 and 4 lags were optimal, but if that was a sound procedure then go with it (option 2).


Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.