【1129期】 4月19日计量经济学学术讲座：Doubly Robust Identification and Estimation of the LATE model with a Continuous Treatment（Dong Yingying董颖颖，University of California Irvine，Associate Professor）

【主题】Doubly Robust Identification and Estimation of the LATE model with a Continuous Treatment

【报告人】Dong Yingying董颖颖，University of California Irvine，Associate Professor

【时间】2023年4月19日周三10:00-11:30

【地点】腾讯会议室：959-504-571

【语言】英文

【摘要】We consider identification and estimation of the LATE model with a continuous treatment. We discuss two alternative restrictions on the first-stage instrument effect heterogeneity that allow for causal identification - monotonicity and treatment rank similarity. The former is popular in the LATE literature while a slightly stronger version of the latter is exploited in the non-separable IV model literature. Neither assumption implies the other. Both assumptions can at best be partially tested. We propose causal estimands that have doubly robust properties in that they are valid under either of these two alternative restrictions. We further propose semiparametric estimators and derive the asymptotic properties of these estimators. When monotonicity holds, our primary estimand reduces to the standard LATE estimand; otherwise, when treatment rank similarity holds, our approach allows for identifying treatment effect heterogeneity at different (conditional) treatment quantiles. We apply the proposed estimators to evaluate the impacts of neighborhood poverty rate (a continuous treatment variable) on adults' labor market outcomes using the Moving to Opportunity (MTO) social experiment.