In the normal linear simultaneous equations model, we demonstrate a close relationship between two recently proposed methods of instrument selection by presenting a fundamental relationship between the two sets of canonical correlations upon which the methods are based.

We assess relative performance of three recently proposed instrument selection methods via a Monte Carlo study that investigates the finite sample behavior of the post-selection estimator of a simple linear IV model. Our results suggest that no one method dominates.

This paper proposes methods for both the consistent estimation of so-called long run canonical correlations (LRCCs) and also testing the null hypothesis that a subset of LRCCs are zero. Two test statistics are proposed and their limiting distribution is derived under the null hypothesis. It is...