Hi all,
I am doing unbalanced panel model regressions where T > N. I have first done a cross-sectional dependence test for each of my variables using xtcd, finding cross-sectional dependency in all of them. Then, I have done a unit-root test for each of my variables using pescadf. I found that my dependent variable is I(1), and my explanatory variables are a mix of I(0) and I(1). Then, I have done a conitegration test using xtwest, only including the I(1) explanatory variables, finding cointegration to be present.
What are my estimation options from here?
I initially planned to just do static fe/re panel model regressions using Driscoll-Kraay se, and I am not sure if the results achieved from this would be accurate anymore.
Any help would be greatly appreciated, thanks.
I am doing unbalanced panel model regressions where T > N. I have first done a cross-sectional dependence test for each of my variables using xtcd, finding cross-sectional dependency in all of them. Then, I have done a unit-root test for each of my variables using pescadf. I found that my dependent variable is I(1), and my explanatory variables are a mix of I(0) and I(1). Then, I have done a conitegration test using xtwest, only including the I(1) explanatory variables, finding cointegration to be present.
What are my estimation options from here?
I initially planned to just do static fe/re panel model regressions using Driscoll-Kraay se, and I am not sure if the results achieved from this would be accurate anymore.
Any help would be greatly appreciated, thanks.
