Replicate Table with xtabond2

[Thomas Haoi]

Hello everybody,

I have some paneldata that I first regress with FE and clustered at the region lvl. The Code is:

global controls "Var5 Var6 Var 7"

eststo mod1: quietly xtreg Var1 Var2 Var3 yr*, fe cluster(rg)
eststo mod2: quietly xtreg Var1 Var2 Var3 Var4 yr*, fe cluster(rg)
eststo mod3: quietly xtreg Var1 Var2 Var3 Var4 $controls yr*, fe cluster(rg)

Now I want to reproduce these using the GMM estimation with a lagged dependent variable, but am lost on what to do.
I currently have:

xtabond2 L(0/1) Var1 Var2 Var3 yr*, ///
gmm (L.(Var1 Var2 Var3 yr*))

xtabond2 L(0/1) Var1 Var2 Var3 Var4 yr*, ///
gmm (L.(Var1 Var2 Var3 Var4 yr*))

xtabond2 L(0/1) Var1 Var2 Var3 Var4 $controls yr*, ///
gmm (L.(Var1 Var2 Var3 Var4 $controls yr*))

But to be honest, I am not sure what es happening in generalized method of moments.
Tried reading the paper by David Roodman on xtabond 2, but got lost pretty quickly.

Would appreciate any help.

Greetings, Thomas