Dear Colleagues,
I am wondering whether it is necessity of conducting Unit root tests, cointegration, stability tests when conducting a research on banking, studying the implications of financial inclusion on banking profitability for a panel of 74 banks from 10 different countries. The methodology used here is a Dynamic Panel Data (SGMM) and not Panel cointegration analysis. With my colleagues, we have used 10 banking variables and we were not looking for the causal relationships among the variables.
Honestly, I have never seen a single High quality paper using a Dynamic panel data conducting the above-mentioned tests. But I need your input regarding this matter.
Many thanks
Dear Helmi, conducting unit root and cointegration is to test the long run relationship between variables.
Don't forget that technically if N>T then pgmm is used and you don't have to conduct the panel data time series tests...
Dear Taha Zaghdoudi ,
Many thanks for your input. Indeed, that was the purpose for conducting such prerequisite tests.. I am waiting to see if others have different opinions when it comes to Dynamic Panel data regression.
Regards,
Helmi
- Badges
- Science topic