Hi,

I recently suffered from confusion about dynamic panel models with fixed effects for count data.

My regression model is:

Fraud_(i,t)=exp⁡ (β1*Fraud_(i,t-1)+β2*VPD_(i,t)+β3*VPD_(i,t-1)+β4*Asset_(i,t)+β5*GDP_(i,t)+u_i+ε_it )

For this, I used a dynamic GMM estimator, under the guidance of Recent Developments in Panel Models for Count Data (Pravin K. Trivedi, 2010).

My syntax in Stata 14 is given by:
gmm ( Fraud - exp({xb: L.Fraud VPD L.VPD Asset GDP}+{b0})), xtinstruments( Fraud, lags(2/2)) xtinstruments(VPD, lags(2/2)) xtinstruments(Assets GDP, lags(0/0)) winitial(xt D) onestep vce(cluster Bank)
where Fraud_(i,t-2), VPD_(i,t-2) and those strictly exogenous variables (Assets_(i,t) GDP_(i,t)) are used as instruments.

My question is:
(1) How to incorporate individual fixed effects?
(2) Are the instruments correct?

I will appreciate your help!