After cross-section dependence test?

[Sen Bulduk]

Hello, I have panel data between 1999-2022.
I did a cross-sectional dependency test on my variables, and all my variables were cross-sectionally dependent. After cross-section dependence, I performed the CIPS unit root test, one of the 2nd Generation Unit Root tests. In the CIPS test results, I concluded that some variables were stationary and some were not. Which test should I do now? Can I create a first difference series of my variables? Then can ı perform fixed, random or pooled regression?

[Eshaan Gupta]

Hi Sen, I am also in the same position as you found yourself. I wanted to ask whether you found a solution to your problem. Thanks.

[George Ford]

Look into Common Correlated Effects (xtmg) and see if that fits. I believe you can mix I(1) and I(0) and it takes into account the cross-sectional dependence.

first differencing on the mix is not ideal.

[Eshaan Gupta]

Hi George Ford, thanks for the response. I have looked into CCE, and it worked so thanks for that. However, when I try include time-invariant dummy variables, in the results table it says that they have been omitted. Is there a way to solve this, or would i need a different method other than CCE. Thanks.

[Eshaan Gupta]

Also, is CCE a long-run analysis tool or simultaneously short-run and long-run. Thanks.