Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Bootstrapping two stages: GMM and OLS

    Hi all,

    I want to estimate a simple OLS of the kind

    Code:
    reg Y X Z
    where Z is a variable that comes out as the predicted value of a GMM done as a first stage.
    Z has N_Z observations and X has N_X < N_Z observations.

    From my understanding, the built in bootstrap command does not work in its current form with GMM. But even overcoming this manually, what approach is recommended to perform this procedure with different sample sizes? If it matters, N_X < N_Z because the data for X is harder to collect or unobtainable for all N_Z observations.

    Thank you!
    Alex
Working...
X