I am doing a study using countries having different characteristics and therefore i was wonderiong if i can add country fixed effects in the estimation. Due to the issue of endogeneity, my estimation technique is static GMM and i couldn't find a way to add country fixed effects in the estimation. I tried incorporating "i.country" in the model but the stata returns "couldn't evaluate equation 1". So any suggestions about adding country effects in static GMM model in stata? The command which i am writing currently is:
gmm (log_L-{b0}-{b1}*log_W-{b2}*KOF_Trade-{b3}*log_Y-{b4}*log_K-{b5}* log_rer-{b6}*log_ILO-{b7}*logt), twostep instruments(l2.log_exports l.Tariff_rate l2.log_Y l.log_industry l.log_K l2.log_agri l.log_manufacturing l2.log_rer)
The doesn't work in the command the way you may want it to.
Depending on how many values of country you have, you could try one of two things:
a). Add to equation 1, where X is a single level you want to test against (can also be a string).
b). Try adding all countries as follows:
* begin stata code
local i = 7
levelsof country, local(cntry)
foreach c in `cntry' {
if `c' > 1 {
local estimators `estimators' + "- {b`i'}*(country==`c') "
}
local i = `i'+1
}
gmm (log_L-{b0}-{b1}*log_W-{b2}*KOF_Trade-{b3}*log_Y-{b4}*log_K-{b5}* log_rer-{b6}*log_ILO-{b7}*logt `estimators'), twostep instruments(l2.log_exports l.Tariff_rate l2.log_Y l.log_industry l.log_K l2.log_agri l.log_manufacturing l2.log_rer)
* end stata code
But I believe you will need to make sure you do not have more parameters than instruments.
have a look at https://www.stata.com/meeting/uk19/slides/uk19_kripfganz.pdf
https://www.researchgate.net/post/DPD_with_GMM_can_I_estimate_it_using_STATA
https://arxiv.org/pdf/1902.10318.pdf
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation