Panel Unit Root & Panel Cointegration

[Shin]

Dear friends,

I strongly need your helps on this matter.
I have total 372 data with 54 years and 7 countries, total 5 variables namely Y, L, F, I, S where S is dummy variable and Y is independent variables.
I have tested unit root test for these 5 variables, and found except S is stationary at level others is stationary at first difference.
Thus, I proceed to panel cointegration test (Pedroni, Kao and Fisher/Johansen test). Below are conclusion can be drawn from results display from Eviews (as attached):

Pedroni Test: No cointegration among variables
Kao Test: cointegration among variables
Fisher Test: At most 1 cointegration among variables (significance of none cointegration is very high compare to at most 1 cointegration among variables)

Can I conclude that these variables are not cointegrated? As I am going to proceed to PMG test, thus cointegration results seems important to identify long run relationship did exist before we can proceed further.

Besides, I do not find ARDL Bound test for panel which is more suitable to test I(0) and I(1) data.

Thanks for your helps & advises in advance.

[Santosh]

Hello,
Can we apply panel time series techniques (like panel unit root tests, panel cointegration tests, Panel ECM, and panel causality test) to "unbalanced panel data" structure?

IS IT NECESSARY THAT DATA SHOULD BE IN BALANCED PANEL STRUCTURE?

A clarification would be greatly appreciated.
Thanking you in advance.

[EViews Glenn]

In general, EViews allows you to perform the analysis with unbalanced data. There may be a couple of exceptions where this is not the case.

[Santosh]

Okay. Thank.