Hello community,
I have some questions about panel unit root test:
1- My dataset contains 3000+ units, and the time length is 14. (N=3000 and T=14). Can I ignore unitroot test because of T<15.
2- If I have to check stationary, is Harris-Tzavalis the most suitable test for my datasets? As far as I know, Harris-Tzavalis is proper if T is fixed and N --> ∞. I have seen this from stata help file and that refers to Baltagi.
3- Lastly, when should I include time trends. I have no clear idea about this. Variables don't show a clear trend when I use the xtline command. In the light of this, should I include time trend?
I have some questions about panel unit root test:
1- My dataset contains 3000+ units, and the time length is 14. (N=3000 and T=14). Can I ignore unitroot test because of T<15.
2- If I have to check stationary, is Harris-Tzavalis the most suitable test for my datasets? As far as I know, Harris-Tzavalis is proper if T is fixed and N --> ∞. I have seen this from stata help file and that refers to Baltagi.
3- Lastly, when should I include time trends. I have no clear idea about this. Variables don't show a clear trend when I use the xtline command. In the light of this, should I include time trend?
