Cross-sections dropped in Pedroni resid. cointeg. test
Posted: Fri Jan 10, 2014 6:40 pm
I don't understand why my panels have been dropped and no estimation given (output is 'NA') in the 2 ADF tests (both Panel and Group) for the Pedroni residual conitegration tests.
The read-out from Eviews 8.2 is as follows: "Cross-sections included: 6 in non-parametric (PP) test; 0 (6 dropped) parametric (ADF) test."
The other tests all seem to have been carried out correctly. The data-set I am using is bootstrap-generated. I will mention that one of the original series, GDP, is the same for each of the 6 individual sectors in the panel -- both in the original data, and in its bootstrapped extension. But the other series are different for each sector, and so are their bootstrapped extensions.
Any suggestions as to what could be causing this problem? Could it be the standard assumption of independence between sectors (obviously violated here)? But if so, why is it only causing problems for the ADF test?
The read-out from Eviews 8.2 is as follows: "Cross-sections included: 6 in non-parametric (PP) test; 0 (6 dropped) parametric (ADF) test."
The other tests all seem to have been carried out correctly. The data-set I am using is bootstrap-generated. I will mention that one of the original series, GDP, is the same for each of the 6 individual sectors in the panel -- both in the original data, and in its bootstrapped extension. But the other series are different for each sector, and so are their bootstrapped extensions.
Any suggestions as to what could be causing this problem? Could it be the standard assumption of independence between sectors (obviously violated here)? But if so, why is it only causing problems for the ADF test?