Dear Members,
I am checking imapct of financial development on economic growth by using System GMM.
Is it right to use option of eq(diff) with gmm style(L.GDP private credit,eq(diff)) and eq(level) with iv style(time dummies, control variables(eq(level)) in System GMM estimation?
Regards.
Dear Gitz Negi
In my opinion, I suggest that you have to try and check wether your GMM results are consistence and good through the test of Sargan, Hansen and the AR(2). Beside that you have to consider the number of instruments that must be less then the cross - section group in your sample.
In addition, I suggest that you must try to see in the theory what your variables are endogenous so that you can put it into the GMM style and try to do with difference lag in the GMM option
Best
Canh
Hi
If diff EQ specified, levels of the variables are used to form instruments, similarly differences of variables are used to form instruments if level EQ specified. if standard instruments specified, then differences of variables are used as instruments. Instead of standard instruments, level can be specified and then instruments are based on level variables.
Thanks
- Badges
- Science topic