Methods in Labor Economics

Ryan incentives work getting teachers to come to

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ol,” American Economic Review, 102 (June 2012): 1241-1278. Card, David, and Dean Hyslop. "Estimating the Effects of a Time-Limited Earnings Subsidy For Welfare-Leavers." Econometrica Vol. 73 (Nov. 2005) 1723-1770. Fortin – Econ 560 Lecture 0 Causal Inference in Economics Goal is to determine the effects of particular interventions on policies, or to estimate features of the behavioral relationships suggested by economic theory Examples: What is the effect of education on earnings? What is the impact of unionization of wages? What is the effect of class size on student achievement? What is the effect of gifted education on the student achievement? Card, D. and L. Giuliano “ Does Gifted Education Work? For Whom?” Paper presented at the NBER SI2013. There a few challenges in experimental or quasi-experiment research Sampling uncertainty, the fact that inferences are drawn from a particular data sample rather than the entire population Presumes that a clear counterfactual outcome can be agreed upon (counterexample: which non-unionized wage is appropriate?) A more important difficulty is that there is no data on counterfactual states of the world Fortin – Econ 560 Lecture 0 The method of random assignment invented by R. A. Fisher (1925) to study agricultural interventions when soil conditions are hard to measure and held fixed, break the fundamental log-jam in the quest for counterfactual outcomes. Randomized trials form the conceptual benchmark for assessing the success or failure of studies based on quasi-natural experiments and much of the language of randomized trial has now been adopted in empirical economics. When an intervention (called “treatment”) is randomly assigned, differences between groups with different exposure levels are due to the exposure itself and not to other uncontrolled factors. More formally, let (1) denote an outcome of interest if the observation unit i is treated. If unit i is not treated, we cannot observe this counterfactual or potential outcome. Let (0) denote the outcome if the observatio...
View Full Document

This document was uploaded on 02/26/2014 for the course ECON 560 at UBC.

Ask a homework question - tutors are online