Working Paper: NBER ID: w12574
Authors: James J. Heckman; Sergio Urzua; Edward J. Vytlacil
Abstract: This paper examines the properties of instrumental variables (IV) applied to models with essential heterogeneity, that is, models where responses to interventions are heterogeneous and agents adopt treatments (participate in programs) with at least partial knowledge of their idiosyncratic response. We analyze two-outcome and multiple-outcome models including ordered and unordered choice models. We allow for transition-specific and general instruments. We generalize previous analyses by developing weights for treatment effects for general instruments. We develop a simple test for the presence of essential heterogeneity. We note the asymmetry of the model of essential heterogeneity: outcomes of choices are heterogeneous in a general way; choices are not. When both choices and outcomes are permitted to be symmetrically heterogeneous, the method of IV breaks down for estimating treatment parameters.
Keywords: Instrumental Variables; Essential Heterogeneity; Causal Inference; Treatment Effects
JEL Codes: C31
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| knowledge of potential outcomes (D80) | selection bias (C24) |
| policy adoption (D78) | selection bias (C24) |
| essential heterogeneity (F12) | complications in estimating treatment effects using IV (C36) |
| instrumental variables not appropriately chosen (C36) | biased or misinterpreted estimated treatment effects (C21) |
| standard IV assumptions not holding (C36) | complications in identification of causal effects (C32) |
| treatment effect of a policy (y1 - y0) varies among countries (C21) | heterogeneous responses to interventions (C21) |
| certain conditions (C62) | standard IV can identify discrete approximation to marginal gain parameter (C36) |
| monotonicity and independence of instruments from unobserved factors (C36) | identification of causal effects (C22) |