Hi,

This is a follow-up question for this topic, but because the question is different, I thought may be it can be more useful in the future if someone searches this question.
Older topic: https://www.statalist.org/forums/for...are-endogenous

As before, here is the characteristics of my data:
"My data has three waves, but with gaps (unbalanced), and the N is approximately 2500. All the variables below are statistically significant in the FE/RE estimations."

I have estimated the following equation (the original variable names are changed for ease of reading):

Code:
xtabond2 DV L.DV X1 X2 X3 X4 wave2 wave3, gmm(L.DV) gmm(X1, lag(2 .)) gmm(X2 lag(2 .)) gmm(X3, lag(1 .)) gmm(X4, lag(1 .)) iv(wave2 wave3, eq(level)) robust small twostep
All Hansen tests, including the incremental ones, are above 0.25 threshold (Kiviat 2019, Roodman 2009).

My question is as follows:
- In the pooled OLS regression, B for X1 is 0.66, and SE is 0.05.
- In the FE regression, B for X1 is 0.22, and SE is 0.05.
- In the Lagged FE (biased) regression, B for X1 is 0.19, and SE is 0.07.

BUT, in the dynamic panel specified above:
- B for X1 is 4.87, and SE is. 2.26.

Other variables (X2, X3, X4) are not like this, they are within the bounds of pooled OLS and FE regression.

What could be the reason for this?