Working Paper: NBER ID: w14706
Authors: James J. Heckman; Sergio Urzua
Abstract: This paper compares the economic questions addressed by instrumental variables estimators with those addressed by structural approaches. We discuss Marschak's Maxim: estimators should be selected on the basis of their ability to answer well-posed economic problems with minimal assumptions. A key identifying assumption that allows structural methods to be more informative than IV can be tested with data and does not have to be imposed.
Keywords: Instrumental Variables; Structural Models; Economic Analysis; Causal Inference
JEL Codes: C31
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Structural models (E10) | Identify gains from different choices made by individuals (D91) |
| Structural models (E10) | Identify fraction of individuals induced into a state from various origin states (C24) |
| Structural methods (C10) | Reveal distributional effects of a policy change (D39) |
| IV methods (C26) | Identify mean gross gain to a program for those induced to take it (C93) |
| IV methods (C26) | Cannot specify distribution of gains across different origin states (D39) |
| IV methods (C26) | Yield aggregate estimate that masks important variations (E10) |
| IV methods (C26) | Provide biased estimates if underlying assumptions are not met (C51) |
| Heterogeneous responses among individuals (D29) | IV methods can lead to biased estimates (C36) |