Note on "Location Properties of Point Estimators in Linear Instrumental Variables and Related Models"
In this paper, Example 1 is not covered by our Theorem 2.5 because the sign restriction on the mean implies that Assumption 2.4 does not hold. We thank Isaiah Andrews and Timothy Armstrong for pointing this out to us. Their paper "Unbiased Instrumental Variables Estimation under Known First-Stage Sign" shows that in the case of our Example 1, and in the linear instrumental variables model with sign restrictions, it is possible to construct a mean-unbiased estimator.