Response to the Review of ‘Public Policy in an Uncertain World’
作者回应Guido Imbens对其著作的评论，阐述部分识别概念：将识别视为集合而非二元事件，参数可能落在某个集合中，点识别时集合为单点，部分识别时集合包含多点但小于全参数空间。
I am grateful to the Economic Journal for the opportunity to respond to the comments of Guido Imbens on my recent book (Manski, 2013). Imbens characterises the book as an introduction to my work on partial identification. He evidently views this work as sufficiently familiar to the readership of the Economic Journal as not to require a definition of the term partial identification or references to the literature. However, he also writes that the conceptual issues addressed in my work have been ‘politely ignored’ by much of the profession. With this in mind, I will begin by observing that, with occasional exceptions, econometricians long thought of identification as a binary event: a parameter is either identified or it is not. Research on partial identification thinks of identification as a set concept. Combining assumptions with observation of specified features of a population, one may conclude that a parameter lies in some set. A parameter is said to be point‐identified if the set is a point. It is partially identified if the set contains multiple points but is smaller than the full parameter space.