Omitted variables bias in GMM

[beth cooper]

is it possible to have omitted variables bias in system GMM? When I don't include a regressor in my regression, all the other coefficients turn insignificant, whilst they're significant before.

[Sebastian Kripfganz]

Yes, you can have omitted-variables bias in the usual way when estimating a model by GMM.

[Carlo Lazzaro]

Beth:
as an aside to Sebastian's helpful reply, you should also consider whether the omission has any bearing on a fair and true representation of the data generating process you're investigating, as this issue is much more relevant than any statistical significance.

[beth cooper]

Beth:
as an aside to Sebastian's helpful reply, you should also consider whether the omission has any bearing on a fair and true representation of the data generating process you're investigating, as this issue is much more relevant than any statistical significance.

I'm not sure I understand what you're saying. I think the Omitted variable is causing downwards bias on the explanatory variables making their effect insignificant.

[Carlo Lazzaro]

Beth:
in my previous reply I was wondering whether it makes sense to omit the regressor you mentioned in the light of the data generating process you're investigating.
But probably your question was different, and you simply tested what happens to your coefficients when you omit that predictor.

[Nariman Sayed]

Hi carlo,

can I know why one or two of my independent variables are omitted after running GMM. I have t = 30 and n = 11. the variables are only sig when I add i.time, however, it omit twoof my explanatory variables and the lag of the dependent variables. I use collapse to minimize the instruments in the two-step sys GMM command. but sargan test is 1 indicating a problem of instruments profilaration . can you help me please ?