1. Whether variables are cointegrated or not? Zed Taha posted the following figure.
Sayed Hossain commented> Out of 11 statistics, only 2 statistics are significant while the rest 9 statistics have failed to reject null hypothesis meaning that all variables LY, LEN, LF and LK are not cointegrated. Meaning all these variables do not have long run relationship. So we can not run panel VECM model.
2. I have two panel data variable such as GDP and CPI for five countries and I have run Fisher panel cointegrating test (Johansen sense) and got following result. I assume here 10 percent level of signficance. Are these two variables cointegrated? Sayed Hossain posted the following figure below.
Sayed Hossain commented> At "None" we fail to reject null Fisher trace statistics and also Fisher max-eigen statistics meaning that we accept null. That is two variables GDP and CPI are not cointegrated meaning they do not move together in the long run. We are unable to run panel VECM model. It is Johansen type multivariate of panel cointegration test meaning that if we change the side of GDP and CPI setting while running Johansen Fisher panel cointegration test, decision would be same.