标题： 因果效应的识别与聚集设计 显示英文标题

标题： Identification of Causal Effects with a Bunching Design

摘要： 我们证明，当连续处理变量的分布存在聚集时，可以识别因果效应，而无需施加任何参数假设。 这产生了一种新的非参数方法，用于在没有工具变量、面板数据或其他流行的因果推断研究设计的情况下克服选择偏差。 该方法利用积分理论中的变量变换定理，将选择偏差与处理密度和与混杂因素变化的结果部分的密度比相关联。 在聚集点，处理水平是恒定的，因此结果的变化完全由不可观测因素引起，使我们能够识别分母。 我们的主要结果确定了那些边际选择进入聚集点的个体的平均因果反应。 我们进一步表明，在对选择偏差施加额外光滑性假设的情况下，聚集点以外的处理效果也可能被识别。 我们提出了基于标准软件包的估计量，并将该方法应用于估计孕期母亲吸烟对出生体重的影响。 显示英文摘要

摘要： We show that causal effects can be identified when there is bunching in the distribution of a continuous treatment variable, without imposing any parametric assumptions. This yields a new nonparametric method for overcoming selection bias in the absence of instrumental variables, panel data, or other popular research designs for causal inference. The method leverages the change of variables theorem from integration theory, relating the selection bias to the ratio of the density of the treatment and the density of the part of the outcome that varies with confounders. At the bunching point, the treatment level is constant, so the variation in the outcomes is due entirely to unobservables, allowing us to identify the denominator. Our main result identifies the average causal response to the treatment among individuals who marginally select into the bunching point. We further show that under additional smoothness assumptions on the selection bias, treatment effects away from the bunching point may also be identified. We propose estimators based on standard software packages and apply the method to estimate the effect of maternal smoking during pregnancy on birth weight.