Sensitivity of Regular Estimators

Abstract

This paper studies local asymptotic relationship between two scalar estimates. We define sensitivity of a target estimate to a control estimate to be the directional derivative of the target functional with respect to the gradient direction of the control functional. Sensitivity according to the information metric on the model manifold is the asymptotic covariance of regular efficient estimators. Sensitivity according to a general policy metric on the model manifold can be obtained from influence functions of regular efficient estimators. Policy sensitivity has a local counterfactual interpretation, where the ceteris paribus change to a counterfactual distribution is specified by the combination of a control parameter and a Riemannian metric on the model manifold.