Hello all,

I would like to see the proportional marginal effects (i.e.: ey/ex) after a regression with fixed effects. There is always an error message which says:

"could not calculate numerical derivatives -- discontinuous region with missing values
encountered r(459);"

I am able to attain dy/ex and dy/dx but because of the variable measurements, I need to see ey/ex to be consistent. ey/ex is possible with random effects, however, they are not appropriate in my model.

Is there a way to calculate ey/ex?

Thank you,
Karen.