Dear Professor Joao,
Thank you very much for your reply. It is truly appreciated. The point you raised is indeed puzzling me. I originally kept GNI in values, instead of logging it, in order to get a more understandable interpretation once I look a the marginal effects. For instance, by keeping GNI in value, I would be able to assess the marginal effect of RD for exporters with a GNI per capita of 10.000$ or 80.000$.
Since I am interested in GNI only as a moderator (the inclusion of exporter-year FE take into account its main effects), do you think there is any shortcoming in including it in values and interpret the coefficient I listed below (as well as the marginal effects) as normal elasticities?
Thank you again for the valuable time and the help you provide every day on this forum.
Best regards
Thank you very much for your reply. It is truly appreciated. The point you raised is indeed puzzling me. I originally kept GNI in values, instead of logging it, in order to get a more understandable interpretation once I look a the marginal effects. For instance, by keeping GNI in value, I would be able to assess the marginal effect of RD for exporters with a GNI per capita of 10.000$ or 80.000$.
Since I am interested in GNI only as a moderator (the inclusion of exporter-year FE take into account its main effects), do you think there is any shortcoming in including it in values and interpret the coefficient I listed below (as well as the marginal effects) as normal elasticities?
Code:
-----------------------------------------------------------------------------------
| Robust
gvc_total | Coef. Std. Err. z P>|z| [95% Conf. Interval]
------------------+----------------------------------------------------------------
lag1_lnRD | -.2286747 .0245165 -9.33 0.000 -.2767261 -.1806232
GNIcap_exp | 0 (omitted)
|
|
c.lag1_lnRD#|
c.GNIcap_exp | .0046135 .0005504 8.38 0.000 .0035347 .0056922
Best regards
Comment