Panel Data: Persistence in endogenous variables

[Guneet Kaur]

Hi all,

In my panel data study, one of the regressors is an endogenous variable. Some estimators such as Arellano Bond and Blundell Bond (BB) are very handy to address the issue of endogeneity in a panel setting. My key question is: does Blundell Bond (BB) work when there is a serial correlation in one or more of the endogeneous variables? Can someone please explain why? BB uses difference in variables to instrument level instruments, but if the variable is persistent, a lot of delta values will be zero or close to zero in many time periods.

[Nick Cox]

Cross-posted at http://stackoverflow.com/questions/2...enous-variable

Please see the FAQ Advice for our policy on cross-posting, which is that you should tell us about it.

[Guneet Kaur]

Thank you Nick for pointing this out. I should have posted the link here.

Can I get the answer on this thread itself or I will have to repost my question?

[Nick Cox]

I see no reason to re-post this question anywhere. Unfortunately, no-one seems able or willing to answer. (I have no idea what the answer is.)