I am not sure whether I am interpreting the cointegration test correctly.

This is the test result:

Johansen cointegration test

Because of the probability of the test I understand that my series are cointegrated of order 2.
Is that correct?


As I read the results, you have two cointegrating equations, or two cointegrating vectors. This translates into cointegration rank being equal to one (number of variables in the system minus the number of cointegrating vectors: $3-2=1$).

Two cointegrating vectors is not the same as cointegration order being equal to two. Your cointegration order is equal to one if each of the original variables is I(1), since there exists a linear combination (actually, two linear combinations -- due to the two cointegrating vectors) of variables that are stationary.

Side note: line 3 Included observations: 21 after adjustments suggests you have very few data points, so the results of the cointegration test may not be reliable.

  • $\begingroup$ Thank you! If i were to estimate a model i should use a VECM (instead of VAR)? $\endgroup$ – user3452075 May 27 '15 at 15:50
  • $\begingroup$ Since the Johansen test shows presence of cointegration, VECM would be the logical choice. On a side note, I am quite concerned over the low number of observations. If I am reading the table correctly, the number of observations is essentially too low for any model larger than, say, VAR(1) or VECM(1). Try to use parsimonious models as far as possible and do not trust the estimation results too much. $\endgroup$ – Richard Hardy May 27 '15 at 17:59

protected by Community Sep 2 '17 at 21:51

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

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