0
$\begingroup$

I consider the model

$$y_{it} = \rho y_{it-1} + X_i\beta + \alpha_i + u_{it},$$

where $\alpha_i$ are unobservable individual effects. I am interested in estimation of $\rho$, while other coefficients are of a bit lesser interest.

  1. Is this model the right choice if I am interested primarily in $\rho$?

  2. Is Arellano-Bond estimator applicable to this model? (my $X_i$ do not vary in time as in classic dynamic panel model).

If no, what can I use instead?

$\endgroup$

1 Answer 1

0
$\begingroup$

When $X_i$ is not time-varying, you may redefine $$ \alpha_i=X_i\beta + \alpha_i $$ to obtain the baseline Arellano-Bond model $$y_{it} = \rho y_{it-1} + \alpha_i + u_{it},$$ in which you can use the corresponding estimator provided the other assumptions (e.g., no serial correlation in $u_{it}$) are met.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.