Working Paper: NBER ID: w28364
Authors: Dmitry Arkhangelsky; Guido W. Imbens
Abstract: We study identification and estimation of causal effects in settings with panel data. Traditionally researchers follow model-based identification strategies relying on assumptions governing the relation between the potential outcomes and the unobserved confounders. We focus on a novel, complementary, approach to identification where assumptions are made about the relation between the treatment assignment and the unobserved confounders. We introduce different sets of assumptions that follow the two paths to identification, and develop a double robust approach. We propose estimation methods that build on these identification strategies.
Keywords: causal inference; panel data; double robust; identification strategy
JEL Codes: C01; C12; C23
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| traditional model-based identification strategies (C50) | design-based strategies (C90) |
| treatment assignment mechanism (C90) | ability to estimate treatment effects accurately (C51) |
| double robust estimator (C51) | valid causal inference (C20) |
| unobserved confounder (C20) | bias from unobserved confounding (C21) |
| correct model specifications (C52) | reliable causal estimates (C51) |
| identification strategy (F55) | reliable causal estimates (C51) |