I'm working on a panel with N=4 and T=76, trying to run the following model:

yt = yt-1 + Xt + controls

where x is an exogenous regressor.

I have two questions:

1)I detected a unit root issue in yt, which is solved differencing. Does it still make sense to include yt-1 as a regressor?

2)I have tried hard to identify the right model to use, but I'm still confused. I have discarded Arellano/Bond-style GMM because they are designed for situations with small T, large N panels. I was considering simple fixed effects model, since stationarity has been taken care of and T is large. Do you have any suggestion?

I would really appreciate your help on this. Many thanks!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.