GMM: constant as a part of the instrument list
Posted: Sun May 16, 2021 6:30 am
Hi,
when using the GMM estimator, one may specify that a constant is not to be included in the set of instruments using the option: nocinst, see:
http://www.eviews.com/help/helpintro.html#page/content/equationcmd-gmm.html
However, in a panel workfile, the constant is automatically added when estimating using panel GMM.
Why is that? is there a way to not include a constant term? Or does it not matter?
When trying in a time-series setting, where I can drop the constant term using the option: nocinst, the instrument rank is the number of non-constant variables, i.e. not taking into account the constant term. The instrument rank do not change when I include the constant term. In a panel setting, however, where the constant term is automatically added, the instrument rank is the number of variables plus 1 (i.e. the constant term). is there a reason for this?
t
when using the GMM estimator, one may specify that a constant is not to be included in the set of instruments using the option: nocinst, see:
http://www.eviews.com/help/helpintro.html#page/content/equationcmd-gmm.html
However, in a panel workfile, the constant is automatically added when estimating using panel GMM.
Why is that? is there a way to not include a constant term? Or does it not matter?
When trying in a time-series setting, where I can drop the constant term using the option: nocinst, the instrument rank is the number of non-constant variables, i.e. not taking into account the constant term. The instrument rank do not change when I include the constant term. In a panel setting, however, where the constant term is automatically added, the instrument rank is the number of variables plus 1 (i.e. the constant term). is there a reason for this?
t