无序处理下的工具变量：来自专业领域教育回报的理论与证据

中文摘要

我们重新审视了 Kirkeboen 等（2016）的识别论证。他们展示了如何将多个无序处理的工具变量与个体对这些处理的排序信息相结合，在允许处理效应异质性的情况下实现识别。我们表明，关键的识别假设具有可检验的推论，并对这些假设被违反时可能产生的偏误提供了新的刻画。我们估计并比较了挪威和丹麦高等教育专业领域的收入回报。我们从实证上考察了专业领域的回报在两国之间是否存在差异以及为何存在差异。

Abstract

We revisit the 2016 identification argument of Kirkeboen and colleagues, who showed how one may combine instruments for multiple unordered treatments with information about individuals’ ranking of these treatments to achieve identification while allowing for heterogeneity in treatment effects. We show that the key identifying assumptions have testable implications and provide a new characterization of the bias that may arise if these assumptions are violated. We estimate and compare the earnings payoffs to postsecondary fields of study in Norway and Denmark. We empirically examine whether and why the payoffs to fields of study differ across the two countries.