Hello,

I am working on a panel data of six Emerging economies with quarterly data from 2000-2020. Some of the variables are cross sectionally invariant. for example: US interest rate which has same data across time for all the countries in the panel. I would like to control for cross sectional dependency in the data (wherein a variable across economies tend to be driven by common external disturbance) for which demeaning the data (subtracting by time means) is the most suggested solution. I am doing FMOLS estimation in the paper.

However, because of cross sectional invariant nature of few variables, the data cannot be demeaned. I would really appreciate if anyone can suggest a possible solution to this? Is it not possible to include cross sectionally invariant variables in a panel setting? if possible, then how to control for cross sectional dependency.

I hope my query is clear.

Thank you!