Tale of two samples : observations and couterfactuals for the evaluation of economic policy

Abstract

Applied microeconomics has, since at least 1951, concerned itself with the impact of economic policies on economic "agents". Roy's (1951) contribution provides a framework for the evaluation of these policies, based on Ricardo's (181&) principle of comparative advantage. Intuitively, participation obtains when the policy is expected to provide a net increase in benefits over non-participation. In other words, voluntary participation yields multiple samples; empirical observations can be split into participants and non-participanta. When evaluating the impact of any policy, inconsistent estimates often arise, because these samples are not directly comparable. Fortunately, a vast literature has, since, developed to provide guidance for mitigating these inconsistencies by building comparable counterfactual samples.

In this lecture, tha canonical identification (of a counterfactual) problem will be discussed, along with some of the more interesting recently developed solutions to that problem The discussion will revolve around some potential extensions of these methods, as well as applications of this method to some of the goals and objectives agreed upon within the Department of Economics. The central thesis, to be unfolded, is that policies developed to achieve a set of objectives can be evaluated through the application of these methods, and that proper evaluation requires forethought regarding the appropriate counterfacual sample.