Primary Submission Category: Instrumental Variables
Identifying Causal Effects of Nonbinary, Ordered Treatments using Multiple Instrumental Variables
Authors: Nadja van ‘t Hoff,
Presenting Author: Nadja van ‘t Hoff*
This paper addresses the challenge of identifying causal effects of nonbinary, ordered treatments with multiple binary instruments. Next to presenting novel insights into the widely-applied two-stage least squares estimand, I show that a weighted average of local average treatment effects for combined complier populations is identified under the limited monotonicity assumption. This novel causal parameter has an intuitive interpretation, offering an appealing alternative to two-stage least squares. I employ recent advances in causal machine learning for estimation. I further demonstrate how causal forests can be used to detect local violations of the underlying limited monotonicity assumption. The methodology is applied to study the impact of community nurseries on child health outcomes.