Determination criteria in panel co-integration tests
Posted: Mon Aug 05, 2013 5:46 am
dear forum,
I have two more questions about panel co-integration test.
After doing panel unit root tests, I am conducting Pedroni's seven test statistics for identifying co-integrating relationships among non-stationary variables in their levels forms.
1) Do the variables (group object) have to pass ALL of the tests (meaning that: if prob. <0.05, there is co-integration)? or does passing the MAJORITY of the tests suffice? Let's say 4 out of 7
2) Is Pedroni test applicable to variables more than 7?
Panel co-inetgration statistics (within dimension)
Panel V
Panel P
Panel t
Panel t
Panel co-integration (between dimension)
Group rho
Group pp
Group ADF
As always, I appreciate your help on this.
best,
lp3
M.D.
I have two more questions about panel co-integration test.
After doing panel unit root tests, I am conducting Pedroni's seven test statistics for identifying co-integrating relationships among non-stationary variables in their levels forms.
1) Do the variables (group object) have to pass ALL of the tests (meaning that: if prob. <0.05, there is co-integration)? or does passing the MAJORITY of the tests suffice? Let's say 4 out of 7
2) Is Pedroni test applicable to variables more than 7?
Panel co-inetgration statistics (within dimension)
Panel V
Panel P
Panel t
Panel t
Panel co-integration (between dimension)
Group rho
Group pp
Group ADF
As always, I appreciate your help on this.
best,
lp3
M.D.