Using a regressor with zero variance in certain cross sections of a panel
Posted: Sun Mar 10, 2019 6:35 am
Hello,
I wanted to know if there was a "smart way" of adding a regressor in panel regressions where: "if this regressor has zero variance in one cross-section, it is omitted".
For example, imagine that we regress electric demand for 40 countries and we add the variable electrification_rate in % to the list of regressors, in developped economies the rate will be 100% for all the sample and in this case, I would like to remove this regressor.
Thanks!
I wanted to know if there was a "smart way" of adding a regressor in panel regressions where: "if this regressor has zero variance in one cross-section, it is omitted".
For example, imagine that we regress electric demand for 40 countries and we add the variable electrification_rate in % to the list of regressors, in developped economies the rate will be 100% for all the sample and in this case, I would like to remove this regressor.
Thanks!