Dear all,
I'm currently working on my thesis about the relationship between financial development, trade openness, and economic growth with N=5 and T=30 (strongly balance). my data are heterogeneous and contain cross-section dependency, so I wish to perform a causality test with xtgcause command.
My unit root test indicates that my 1 variable is stationary at the level when the other two are in the first difference. then I perform a cointegration test by Westerlund with xtwest command in the level.
Code:
xtwest lngdp lnfd_100 lnto, lags(1 3) bootstrap(400)
The result shows that my data can't reject H0 so there is no cointegration.
Code:
Bootstrapping critical values under H0..........
Calculating Westerlund ECM panel cointegration tests.....

Results for H0: no cointegration
With 5 series and 2 covariates
Average AIC selected lag length: 2.2
Average AIC selected lead length: 0

----------------------------------------------------------------+
 Statistic |   Value   |  Z-value  |  P-value  | Robust P-value |
-----------+-----------+-----------+-----------+----------------|
     Gt    |    0.010  |    2.970  |   0.999   |      0.883     | 
     Ga    |   -0.034  |    2.368  |   0.991   |      0.913     | 
     Pt    |    0.629  |    2.050  |   0.980   |      0.873     | 
     Pa    |    0.425  |    1.334  |   0.909   |      0.870     | 
----------------------------------------------------------------+
my question is, can I simply perform xtgcause command when my data are not cointegrated? or I have to do something about this issue?