Dear all,

I am encountering an issue with two variables in my fixed and random effects models. When I include both GDP and GDP² as independent variables, GDP² gets omitted due to collinearity. After reviewing the literature, I found that centering the GDP variable can help address this issue, as it reduces multicollinearity without altering the shape of the model—only the location of the reference point.

I would greatly appreciate it if anyone could share their insights or experiences regarding this approach. Additionally, if you are aware of any papers that have used this method to solve similar problems, I would be grateful if you could point me to them.

Thank you very much.